Digitized by tine Internet Arciiive in 2008 witii funding from IVIicrosoft Corporation http://www.archive.org/details/cellaetrichoraeoOOfresrich CELL.E TRICHOE^E AXD OTHER GHKTSTIAX AXriQlTITIE>S IN THE BYZxVNTINE PROVINCES OF SICILY WITH CALABRIA AND NORTH AFRICA INCLUDING SARDINIA. Vol. I. ILL 'U'S'-T^UA'T'EB'. EDWIN HANSON FEESHFIELD, M.A., F.S.A. I'lUXXiiJ) pKnATElA, 1913. F7S- A few Copies of Hiis Book may he obtained by Students from the Author tkrouqh the Printers and Publishers, IRijon S, BvtiolD, XonDon, JE.C. TO THE MEMOKV OF MY MOTHER, TO MY FATHER, AND TO MY WIFE WHO ACCOMPANIED ME IN MY TRAVELS, THESE PAGES ARE AFFECTIONATELY DEDICATED. iv,5i28aJi5 Sicily. 1898, 1899, 1905, 1906, 1907, 1908 and 1913. Caiabma. 1899 and 1905. Sardinia. 1901 and 1909. AixiiERS and Tunis. 1910 and 1911. CONTENTS. Introduction Illustrations, Plans and Maps. . Books of Reference Additions and Corrections Six Ancient Churches in Eastern Sicily Notes upon these Churches . . Western Sicily Sardinia Calabria PAGE V. to xviii. xix. to xxi. . . xxii. & xxiii. xxiv. 1 15 28 50 77 Tunis . . Cella) Trichoras Henchir Maatria, and Sidi Mohanuned El Gebioui 103 Tunis. .Byzantine Churches at El Kef and Haidra ,. 120 Notes on Coins . . . . . . . . . . . . . . 128 Index INTRODUCTION. The purpose of this book is to record for English students the existence and present condition of some little chapels in 8ieily and Tunis built on a trefoil plan known to French antiquaries as cellsB trichorae, and of some churclies and other remains of the period of the IJyzantine occupation of Sicily, Calabria, Sardinia, and Tunis. This period commences in the middle of the Gth century in the reign of Justinian wlien Belisarius broke up the dominion of the Vandals in Africa and afterwards that of the Goths in Sicily and Calabria. In Africa the occupation lasted till the Saracens took Carthage in 099, in Sicily and Calabria till the Norman conquest in the 11th century, and in Sardinia till the Byzantine governors, after acquiring a quasi independent position as Judges, came under the dominion of the Republic of Pisa about 1073. Gellrp Trichora. A suggestion that these chapels in Sicily were built by the Greeks to suit the ritual requirements of the Eastern church first led me to enquire whether there was any connection between the triple apse built trefoilwise and the three apses, the sanctuary, prothesis, and diaconicon, essential to the due performance of the Byzantine liturgy as elaborated about Justinian's time. The only way to determine this question seemed to be to ascertain when the Greek liturgy was introduced into Sicily and Calabria, then to make a systematic search for these trefoil chapels and then to try and find a Byzantine church with a presbytery arranged in the trefoil plan. I need scarcely say that this search has already involved me in many journeys to a number of out of the way places that few archaologists are likely to trouble themselves to visit, liut as the buildings described are almost the only evidence we now have of a Byzantine occupation of Sicily, Calabria, and Sardinia, that lasted for nearly Wye hundred years, they possess some historical interest and so I venture to record what I have seen and been able to ascertain concerning them. vi. Introduction. After two or threo fruitless errands into Calabria and Sardinia, fruitless at any rate so far as this partienlar question is coneerned, my attention was directed to two trefoil chapels in the central plain of Northern Tunis. A visit to these two buildings satisfied me that there is nothing Byzantine about them either in origin, in plan, or in purpose, and that they belong to the same early period in Thristian history after the Peace of the Church as the monastery churches at .Soliag, the chapel of the Trinity at S. Honorat, and the t\No chapels in the cemetery of S. Callixtus at Eome. Why this trefoil plan was favoured by the early Christians is an open question. Upon the analogy of the round churches built by the Crusaders in imitation of the so-called church of the Holy Sepulchre they may have obtained the idea from some building in Palestine then existing but now destroyed, traditionally connected with our Lord's Ministry or with the Apostles. Or perhaps the trefoil plan was adopted at a crucial period in Christian history to emphasise the doctrine and worship of the Holy Trinity : or from a common form of early sepulchres : or merely by an accidental development of architecture. Be that as it may, it seems certain that the trefoil plan is not a Christian invention, for a small trefoil chamber in the monastery of S. Menas, in Egypt, is fitted up as a vapour bath or laconicum, and the heating apparatus under the floor still exists. And the same trefoil vapour baths are found in the therma^ at Lambessa and Thelepta, in Tunis, built abo^it the period of the Emperors Septimius Severus and Caracalla. Upon this ground some archaeologists have suggested that the trefoil plan was first used by the Christians for their Baptisteries. CJmrchrs. A recent discovery in Sicily itself justifies the suggestion that the chapel at Malvagna may have been built, or at any late used, for the Byzantine liturgy. The style of architecture; shows that this chapel, and also that at Maccari, belong to a much later period than the third Sicilian chapel at S. Theresa, near Syracuse. They are covered with domes supported on the little angle arches or squinches common to the so-called Byzantine -Norman buildings in Sicily, and to Arab buildings of a much earlier date in North Africa and Egypt. The student is aware from tlie text books Introduction. vu. that almost without exception the Byzantine-Norman churches in Sicily are provided with the three parallel apses, the sanctuary, prothesis, and the diaconicon, but as in Constantinople, so in .Sicily, there seemed to be no example of a Byzantine church with a presbytery arranged in the trefoil way. The discovery of a Byzantine church^ at Castiglione, in Sicily, supplies a piece of evidence that was missing. The important facts about it are that it is situated at the foot of Etna, quite close to the Malvagna cliapel, that it is of about the same date as Malvagna, and that the presbytery is arranged in th(^ trefoil way. That is to say that two little niches do duty for the prothesis and diaconicon, and are placed facing one another in the Xorth and South walls, and not on either side of the apse in the East wall, where there is ample room to receive them. Mches instead of apses placed on each side of the central apse are not unknown, or indeed uncommon ; they occur at S. Thechi, in Constantinople, and at the Favara, or Mare Dolce chapel, at Palermo. The Byzantines, like their i)redecessors the Eomans, have left much evidence of their occupation of Africa. Beside many churches there are a large number of walled cities, forts and castles l)uilt by Justinian's generals after the Vandal war. But beyond the modest buildings described in these pages they have left no monuments in Sicily, Calabria, or Sardinia. I say modest because they have no pretence to artistic merit and, architectural features apart, they are only interesting and w^ortliy of record as pieces of historical evidence of the Greek community. The absence of many or important buildings of this period in Sicily and Calabria is due to various causes ; the depopulation and poverty after the Gothic and Vandal w^ars ; the concentration of Justinian's revival in the West to the more important provinces, like Eavenna and Africa, of which Sicily and Calabria did not form part ; the Saracen raids and ecclesiastical controversies in the 7th century ; and the Norman conquest in the 11th century. There is more evidence of the 1. T use the term Byzantine advisedly, for there is not a scrap of Norman Lombard or Arab work either in i)lan, elevation, design, or decoration about it. As this church at Castiglione was not discovered till this iirst instalment of my notes was in print, the description and illustration of it will ha given in the next volume. viii. Tnrrofhicfion. Byzantines in Sardinia which formed part of the exarchate of Africa, for unlike the other Western provinces of the Empire it escaped a Saracen occupation. Immediately after their arrival in these countries the Normans and Pisans set to work to build churches on an extensive scale, making in the process a clean sweep of the older buildings. The case bears some analogy to that of the Saxon churches in England, and one may perhaps be permitted to conjecture that as a general rule what the Normans and Pisans found was not in itself, or in a condition that made it, worth keeping. 1 am not disposed to attribute this to mere Vandalism, but rather to the nature and poverty of the materials used in construction and the natural decay owing to the age of the buildings, for during the five centuries that elapsed between the reign of Justinian and the arrival of the Normans and Pisans there is no evidence of any extensive or general church building in these countries. The architecture for this long period is now represented by the chapels at Malvagna and Maccari and the church at Oastiglione. Nothing else remains of the numerous churches and chapels that must have been built during the Byzantine occupation, and we can only guess what they were like from the ecclesiastical archi- tecture of other countries, or from the style of building adopted l)y the Normans after their conquest of Sicily. In Africa the churches of the Byzantine period, like the forts, were built of old Eoman materials, on the plan of the early Christian basilicas in Eome, and decorated with Byzantine carved detail. Those who are acquainted with 6th century work in Constantinople, Ravenna, or Salonica, will recognise some familiar patterns in the capitals of Justinian's ])eriod jireserved in the mosque at Kairouan. Introduction of the Girek liturfjij into Sicilij and Calabria. And here I offer an apology for the reference in these pages to the ecclesiastical controversies of the 7 th and 8tli centuries arising in a great measure out of the change in the official language from Latin to Greek, asking the reader to remember that the arcliitecture of the churches was adapted to the service to be celebrated, and, though for a long time after Justinian's reign the Church was one. Introduction. ix. the form of service used in the Church of Constantinople, commonly ciillod the Greek liturg'y, was considerably altered and amplified and diiTered, not in lano-uase only, from that used in Kome. The alterations tlien made necessitated a particular form of constructing the chancel and east end that may for convenience be called the tri])le apse. Tliis form of construction, still used in the Greek and Eussian Churches of to-day, can be recognised at the first glance, and, in these countries, is one of the few pieces of sure evidence of a Greek community. To be exact the east end termin- ates in three a])artments ; in the centre a chancel for the altar, terminating in a semi-circular apse, and on each side of it a chapel, also terminating in an apse and communicating witli the cliancel ; tliey are called respectively the chapel of the prothesis and tlie diaconicon. These three apartments, shut oft' from the rest of the church by a screen called the iconostasis, are essential for the due ])erformance of the ceremonies of tlie Orthodox Greek liturgy, but for no other. Where they occur in a churcli built for a Western rite, like, for instance, the Galilean or Roman, they have no ritual significance, and their presence is due either to a fancy of the architect, or perhaps to habit if he happened to be a Greek. The student who is not practically acquainted with the interesting ceremonies referred to will find them described in Neale's History of the Holy Eastern Church. The dilTerence between the Byzantine triple apse chancel and the chancel arrangement found in the earlier African churches, like Tebessa, must be distinguished. In the latter the altar stood, as it does in the Western Cliurch, out in the open, and the apse, copied from a secular Roman basilica, was fitted as a kind of theatre witli a bishop's throne in the centre and seats for the clergy on either side of it. This plan is made familiar to us by the basilican churches in Rome, like S. Clemente, or the church at Torcello, Parenzo Cathedral in Istria, or S. Eirene at Constantinople. Rut in the Byzantine arrangement the small central apse was put to enable the celebrant to walk round the altar at the celebration, and the altar itself stood concealed from the view of the congre- gation by the screen or iconostasis. These alterations in the form of service and church design in Constantinople were applied to Sicily and Calabria, first by tlie Byzantine conquest, secondly by the change of the official language from Latin to Greek about the time when the Emperor Constans X, Iiitrodnrtio)!. came to reside for five years at Syracuse, and the Greek element, in importance, if not in numbers, came to prevail over the Latin. Thirdly, through controversies upon matters of faith during tlie reigns of the Herachan Emperors, and lastly by the transfer of the Sicihan and Calabrian Churches from the Eoman jurisdiction to that of the patriarch of Constantinople, and the corresponding confiscation of the endowment or patrimony of the Eoman Church in Sicily and Calaliria by the Emperor Leo III, in the beginning of the Stli century. The severance of these two Churches from the Latins tlien l)ecame and remained complete for three hundred years till tlie Norman conquest in the middle of the 11th century. The Greek Church in Sicily and Calabria during the Saracen and Norman occupations. While the Saracens ruled in Sicily the Church continued to use tlie Greek rite and to depend on the Patriarch of Constantinople. Tlie Saracens forbade the building of new churches, but those decayed could be restored. Only two churclies apparently existed in Palermo when the Normans came, a little church dedicated to S. Chirico (Agia Kyriaki), and tlie principal mosque that had been a church. Almost directly after their conquest the Normans began to build churches. I have selected for description in these pages three Sicilian and five Calabrian churches, built after the conquest, with chancels in tlie Byzantine form ; tlie Sicilian churches are S. Giovanni dei Lep])rosi and tlie chapel in Favara Castle, both near one another in the suburbs of Palermo, and built, according to a well authenticated tradition, by the brothers Eobert Guiscard and Eoger (the Count) either during or directly after the siege of Palermo, and the chapel of the Trinity belonging to a Greek monastery of S. Basil at Deha. The Calabrian churches built in the Byzantine form are Stilo, Eossano, the Eoccelletta at Squillace, the Patire chapel and Gerace Cathedral. By the terms made with the Eoman see at the Council of Melfi (1059), shortly after the schism between the Greek and Latin Churches (1054), the Normans undertook to reclaim for the Latins the patrimony and the spiritural jurisdiction of the Pontiff over the Sicilian and Calabrian Churches that had been taken away hitroductinti. .XI. ))y the Emperor Leo III. in the beginning of the 8th oentiiry. In Sicily lliere was little difficulty in carrying out the undertaking, for the Saracens had dispersed the Christians, and the Xormans found but a single community at Palermo presided over by an archbishop of the Greek Church named l^icodemus. This prelate officiated at the Thanksgiving Service offered by the brothers Guiscard and Roger when the city was taken ; but at an early date he was replaced by a ^N^orman bishop. I should add that upon the introduction and establishment of the Gallican Church and a Norman episcopate by Count Roger, the clergy of the Greek Church in Sicily were put under the jurisdiction of the Norman bishops ; the liturgy, however, was not interfered with, and con- tinued to be used until the scattered Greek population had either emigrated or amalgamated with the Latin population in the times of the Angevines. The conditions in Calabria were, however, altogether different, for it had become the place of refuge for Greek Christian emigrants from xVfrica, Sicily, and the Levant. These migrations, caused partly by tlie iconoclastic persecutions, and partly by the Saracen conquests in the South and East of the Mediterranean extending over three centuries, had resulted in bringing a large and wealthy Greek population to settle in the cliief towns like Gerace, Stilo, Rossano and Reggio, Cotrone and Bari. To concilitate their Greek subjects Guiscard, and the early Norman princes of his family who succeeded him, allowed the Greeks in many dioceses to elect their own bishops, as well as to retain their liturgy ; and these privileges were continued in several instances for three cen- turies after the conquest, liut in the majority of dioceses on the Adriatic side, Latins were appointed as the Greek bishoprics fell vacant. In Sicily, as I have said, the Greek clergy were put under the Norman bishops. In the Basilian monasteries of Calabria the Greek liturgy con- tinued in use long after the times of the Angevines, and in the end of the 18th century Cardinal Sileto, described by M. Battifol as ' prefet de la congregation dite de la reforme des Grecs laquelle embrassait le service de I'orient Grec catholique,' was commissioned to reform them. I have found no reference in the autliorities to the date or Act when the Greek clergy in this part of the world became Uniati, or first received their orders from the Latins instead of th«.» xii. Introduction Patriarch of Constantinople, or accepted the alterations in the creed and prayers, and the position in the Roman Church eventually confirmed in the 18th century and now held by them. In these changes, no doubt gradually effected, the Latin conquest of Con- stantinople played an important part. The Uniate rites and ceremonies remained, and still remain, substantially the same as those of the Ortliodox Church, and the service is still said in Greek, but the structural peculiarities of an Orthodox Church are not maintained, and the altar stands, according to the Western practice, in full view of the congregation. The several congrega- tions of the Uniate profession now in Sicily and Calabria are the descendants of emigrants from Thessaly and Albania in the 18th century, and the Greek clergy who minister to them stand in much tlie same relation to the Roman bishops as the Greek clergy stood to the Xorman bishops under Roger's settlement. Whether the laity fully appreciated the subtlety of the Uniate profession is an open question. But in Norman times they certainly were not in the fortunate position of the Orthodox Ruthenian emigrants to Canada who recently obtained relief from Latin pretensions tlirough the Privy Council of a King of England. I have included tlie cathedral at Gerace for two reasons, first because in common with the majority of churches built in Sicily and Calabria in the beginning of the Norman conquest it Jias a triple apse, and secondly because Gerace was one of the cities where the Greeks first obtained their local independence and Chureli ])rivileges from Robert (luiscard. From similar concessions made to tlu' Greek community at Ros- sano, during a revolt of Norman barons against the suzerainty of Guiscard's sons Roger tlu» Duke, Boliemond, and their uncle Roger the Great Count of Sicily, we learn why tliese privileges were retained by some cities and new grants made to otliers and, incidentally, how it came about that the bargain made at the council of Melfi was not always kept and the Emperor Alexius Komnenos thought it worth wliile to negotiate with Count Roger for a settlement of the difference between the Greek and Sicilian Churches. I conclude that these early churches in Sicily and Calabria built 'ad usum (iraecorum,' or 'more Graecorum,' were intended for the Greek liturgy, and the ])eculiar form of triple apse essential for the Greek service seems to confirm my conclusion. Introduction. xiii. The examples I have chosen from the ' regular ' churches are, the Cattolica at Stilo, S. Mark at Eossano, the chapel of the Patire at Corejfhano, the Eoccelletta at 8quillace, and the chapel of the Trinity attached to the Basilian monastery at Delia : the church of the Lepprosi and the Favara chapel at Palermo are selected because they are supposed to have been built during or just after the siege, at a time when the Christian population was small and must have been almost entirely Greek. We only know imperfectly how these changes affected the Church in Sardinia, I found evidence of Byzantine art in details and ornaments in many churches, but only one church, Sta. Sarbana at Silanus, certainly built with the triple apse for the Greek liturgy. The great majority of Sardinian churches Jiave only one apse and were rebuilt by the Pisans. 1 have dwelt at some length on the triple apse because of the hturgical signiticance. The architect will find in these old buildings other features of interest, as for example the various forms of roof the Voute en berceau and Voute d'aretes, domes of high and low pitch, and the use of squinches and pendentives to support them. I allude frequently to the insertion of small columns of rare or decorative marble into the angles oT pilasters generally, at the entrance of the chancels or the apses. The earliest example recorded here occurs in the Byzantine basilica at El Kef, so that though it frequently occurs in Mahometan buildings, Hke the mihrab or an important doorway of a mosque, it is not a Saracen invention. The position these pillars occupy possibly indicates that they were originally meant to support the chancel screen. The student of early liturgies used in Sicily and Calabria should consult Neale's History of the Holy Eastern Church. He should also consult a learned and most interesting work by M. Battifol on Eossano and the library of the Basihan monastery of the Patire near by. Of this book I have made much use. If the materials for writing the history of these provinces exist at all they will be found in the archives of some of these Basilian monasteries collected in various libraries like that of S. Salvatore dei Greci at Messina. Xlie general history of Sicily and Calabria from Procopius until xiv. tntrodiictiori. the clironicles oi: the Arab writers edited by Amari, especially that about the interesting and important period when the Emperor Constans II. came to live at Syracuse, is very meagre and still remains to be written. In the course of my journeys I came across a church and three chapels in Sicily that are certainly older than Justinian's reign, and though not directly connected with the principal objects of my research they should find a place in these notes. The Early Chridlan period in Sicilij. The trefoil chapel at S. Theresa by Syracuse, the cruciform chapels at Camerina, and the little basilica with a single apse at Priolo, are earlier than the Byzantine conquest, and belong to the period of architecture conveniently termed early Christian. S. Phocas, at Priolo, resembles some of the earliest churches found in the Levant, and the facsimile of the chapel at Sta. Theresa will be found on the Island of S. Honorat, opposite Cannes. These buildings probably date from the early part of the 5th Centmy. The chapels at Camerina in particular are so interesting as to justify more than a passing reference. Camerina. The twin chapels near Sta. Croce, in Camerina, exactly ahke and obviously built by the same person at the same time, are about a mile apart and quite isolated from any buildings or ruins. I find that by inadvertence I have transposed the names, the Vigna di Mare being nearest the sea and the Bagno di Mare nearest to Sta. Croce. They seem to have belonged to a country village district where the habitations, made of rough shingle and rubble, have crumbled away into their component parts, as frequently happens in an arid coimtry. Whatever their age may be they are real archaeological curiosities. The cru(;iform ground plan appears to estabhsh beyond doubt that they were built for Christian worship, and two details of construction, the waggon vault I have already mentioned and the method of rooling the square intersection or lantern of the nave beWeen the transepts and the chancel, point to Introduction XV. iin early date. At present this intersection in both chapels is covered by a dome made of large and well cut stones arranged in rows of diminishing size from base to crown. There are neither sqiiinches nor pendentives to support the dome, and, to use a colloquial phrase, it is merely dumped down on the crowns of the chancel and transept vaults and on the arch in the wall dividing the nave from the lantern. The dome in each case is clearly a later substi- tution for an original roof as appears by four brackets projecting about 18 inches from the corners of the intersection and six inche? l)elow the base of the dome. The two African chapels described in these notes show that these projections, which I have called brackets, are not brackets at all but springs of a cross vault, the voute d 'aretes, the roof being either flat on tJie top as at Maatria, or following the contour of the vault as at Gebioui. The substitution of the dome for the vault can be accounted for by the latter falling in : this frequently happened, for the cross vault in practice does not seem to have been an enduring form of roof and somcAvhere about the 7th century it was abandoned in ^Vfrica in favour of the dome. It is clear that at an early date, probably at least before the Norman conquest, they have been put to secular use. In each case eight holes about six inches in diameter, have been pierced at great pains and with careful precision through the base of the domes ; these holes, in pairs at the angles of the lantern, are carried througli into tlie interior of the building. The fact that in piercing them the corners of the brackets or springs of the cross vault have been mutilated in the driUing shows that the holes were put after the original roof had been replaced by the dome. Professor Orsi suggests that they were intended for poles to support a tent or awning. It seems to me more likely that the chapels were converted into baths by the Saracens and the holes were intended to serve some purpose connected with the heating. The local tradition is that they were built for baths but I must say that, upon consideration and having regard to the cruciform shape, I have come to tlu; con- (;lusion that they were built for Christian worship, and probably by refugees from Africa during one of tli^e Vandal persecutions in the end of the fifth century or beginning of tlie sixth. The site of these chapels would acquire some interest if it could be indentihed Avith the sea port of Caucana, where Procopius tells xvi. Introduction. us the Byzautine fleet sheltered on its way from Constantinople to Africa, and whence he was sent by Beiisarius to Syracuse to ascertain the condition of the Vandal forces. It may be presumed that Caucana was a place of some importance and the choice seems to lie between the roadstead of Sta. Croce in Camerina and the lagoon of Maccari. Procopius is unfortunately not precise enough and the local remains are insufficiently well preserved to identify Caucana with either of the remains of the large cities discovered by Professor Or si near Sta. Croce and Maccari. The mediaeval and modern names given to these two places do not help identifi- cation. It has, however, been suggested that Vindicari, the name given to the port of Maccari, is derived from Khalat ibn dikami, the Saracen name given to it by Edrisi, and that dikami in its turn was a phonetic corruption of Ichana or Caucana. In the chapter on Africa I have pointed out how Saracen names occasionally help to identify Byzantine sites. So, for instance, Caput Vada where Beiisarius' fleet made for when they sailed from Caucana became Eas Capoudia and Kuspe became el Eosfeh. The substitution of dikami for Caucana is, of course, possible, though the pro- nunciation of the u in Caucana as v makes it unlikely ; the modern names of the Sta. Croce shore bear no resemblance whatever to Caucana. But some of the local conditions at Sta. Croce seem to justify a preference for the Camerina roadstead as the site of Caucana. There is plenty of sea room, shelter from the prevailing N.E. winds and abundant drinking water from the river Oanis ; the last could not have been the least important condition to Beiisarius, for Procopius has given a graphic description of what occurred owing to the lack of water on the prolonged passage from Zante. On the other hand the neighbourhood of Maccari is a salt marsh, there is but little fresh water and, so far as access and shelter go, the difficulties of negociating the nan-ow entrance to Porto Vindicari for a large fleet of six hundred vessels are at once apparent, more especially as the entrance faces North East. The illustrations of the bastion and walls of the ancient Greek city of Heraclea Minoa and of the ' gymnasium ' at Tyndaris need a word or two of explanation. The former are in imminent danger of destruction by the subsidence of the cliffs on which they stand, and as they do not appear to have been photographed before, it seemed worth while to record them before they disappear. Many of the stones have mason's marks. Introduction. xvii. Tlie building at Tyndaiis called gymnasium is remarkably like a church and may have been used for one, though there are no Christian emblems about it, l)ut the form of construction is thoroughly Koman. With two or three exceptions the buildings described in these pages have been noticed in foreign pubhcations. The chapels of ^Southern Sicily in Professor Orsi's articles in Byzantinische Zcit- schrift : the churches at Eossano and the Roccelletta by Sig, Abatino, the Inspector of Monuments for South Italy, in the Neapolitan pubhcation of tbe Storia Patria : the churches of Sardinia, in Dr. Scano's book, eicepting S. Sarbana, discovered by Dr. Ashby : the African chapels, by M.M. Saladin and Diehl. I make no claim therefore either to original or exliaustive discoveries. It seemed to me better to record what exists now than to wait indehnitely for a more complete examination by digging, with the risk of tlie buildings being in the meantime entirely destroyed. To the authors of the various books I have consulted I make here my dutiful acknowledgments ; a list of them will be found in its proper place. For the history of Sicily and S. Italy I have consulted : Martroye's VOccident a Vcpoque Byzantine, for the Gothic and Vandal periods : Gay's I'ltalie Meridionale et V Empire Byzantin for the Byzantine period : Amari's IStoria dei Musulmani in iSicilia, for the Saracen period : and for the Normans and their charters, Chalandon's Uistoire de la domination Normande en Italie et la 8icile, and the articles by my friend C. A. Garuli, the Professor of History in the University of Palermo, in Archivio Storico ^iciliano. The Gloria della Chiesa in iiicilia, by M. Lancia di Brolo, actually archbishop of Monreale, is a mine of information about the Latin Church but says little or nothing about the Basilians and other representatives of the Eastern Church, who, as I shall point out, must have been numerous between the 7th and the 12th centuries. For Sardinia I have consulted two admirable books. On the history, La ISardegna Medioevale, by Professor Besta, and on the art Storia delV Arte in Sardegna, by Dr. Scano. For Africa, M. Diehl's, M. Saladin's and M. Merhn's works, and Les monu- ments antiques de VAlgerie, and L'Algerie dans Vantiquit^, by M. GseU. xviii. Iiitrodiu-tloii. For the ;^'eneral history of the tiiueis I have used the Later Roman Empire, and the illustrations to Professor Bury's edition of (jibbon sujjgested to uie the addition of photographs of some coins collected by my wife during our travels. I have to thank many friends abroad who have been at pains to lielj) me. Foremost among them Sir II. Austin Lee, k.c.m.g., c.b., and through him the Governments of Algeria and Tunis, their civil and miUtary officers, especially M. Kleper the Controlleur of Beja, Dr. iind Mme. Boricaud for their kindness and hospitality to us at Feriana, Pere Lemaitre and the White Fathers of Tibar, M. Mussali, Chief of the Constabulary Bureau at Tunis, the Archaeological Society of Constantine, and M.M. Gsell and Merlin of the department of Antiquities in Algiers and Tunis. I received much kind help in Sardinia from Dr. Scano and the Consuls at Cagliari and Sassari ; in Naples from Signor Abatino and the Society of Antiquaries ; in Sicily from my old friend the late Mr, John Sofio of IVIessina, and his family, especially his son Commendatore Luigi Sofio. I-'rom Professors Orsi and C. A. Garuh I received my hrst encouragement to publish these notes. I venture to liojjc that an appeal to the indulgence of the reader for the many errors and omissions which can scarcely be avoided in a book of this description will not be made in vain. E. H. F. Palermo, 12^/? JaitKary, 1913. XIX. ILLUSTRATIONS PLANS AND MAPS. SICILY. Plate Paok. 1 Fa VARA, near Girgenfi : chapel door hi the Norman Castle ... ... ... ... . . 1 2 Priolo : Monastery of S. Phocas ... ... ... ... 3 Elevation and Plan. 3 Plan and Elei'ation. 12 Bronze Lamp, /ownd «/ Set.inunto ... ... ... ... 17 By permission of Professor Salinas). 13 l{ossoT,iNi : rock cut church ... ... ... ... ... 19 Plan. 14 Impehtat. Byzantine Coins of the Seventh Century, struck hy the Heracl ion family ; Heracliusto J uKtinia}) II. 24 1.5 Palermo: Church of S. Giovanni dei l^v^vvno^i, interior. 31 Plan. \{S the same, exterior, a}id a capital tvith an Arabic inscription ... ... ... ... ... 32 17 !*AT.ERMO : Chapel in Fa vara Castle {Castello a marc dolce) 34 15 the same, interior. Plan ... ... ... 3ij 19 Castelvetrano, Chapel of the Trinity of Delia. West front, and Plan ... ... ... ... ... 37 20 the same. East front ... ... ... ... 38 21 :i/l\zz\R.K, Church of S . Egidio .39 22 Heraclea Minoa : The Citadel Walls of the Creek City . . . 43 23 Tyndaris : The Roman Basilica, interior ... ... ... 44 24 Tyndaris: Apseof the Roman Basilica. Cefalu, Ruins of the Church built over the " prehistoric house." West Front and Door. Plan. ... 45 2.'> Cefalu: Interior of the Church ... ... ... ... 40 20 Termini, the Byzantine triptych i}i the Museum ... ... 47 The sources of the illustrations and plans, other than those taken by my wife and me, are acknowledged on the fly sheets to the Pl.at^^s, XX. ILLUSTRATIONS, PLANS AND MAPS. SARDINIA. Plate Pagk. N 27 FouTOTOJiKKSjS.Gavino, Abbey Church, South side ... ... 57 28 the same, North side ... ... ... ... .58 29 the same, interior and plau ... ... ... .59 30 the same. East gable and apse dedication cross ... 60 Byzantine Capital, and plan. 31 SiTANT'S, Chapel of Sta. Sarbana, West and Soath sides ... (51 Sketch. Plan by Dr. Ashby. 32 SiNis, .S". (liovanni Abbey Church ... ... ... ... 02 Plan. 33 the same. East end ... ... ... ... ()3 34 the same, details ... ... ... ... 04 35 the same. North side ... ... ... ... 05 30 the same, interior ... ... ... ... 00 37 Cagt.iari Abbey Church of S. Saturn ino ... ... ... 09 and Plan 38 the same. West front of the Lantern, Corbel in the Dome ... ... ... ... ... 70 39 the same, East end, detail of irall and dedicatio)i cross ... ... ... ... ... ... 71 40 AssEMiNi, Church of S. Giovan^ii Exterior and Interior a}id stone ivith carved crosses from the Clrurcli of S. Pietro 72 Plan. 41 SIT.ANUS r/zu/ Macomer, iVwr/iar/ii 75 42 Impekial Byzantine Coins; Various ... . ... 77 CALABRIA. 43 Roceij.etta, near Squillace, central apse ... .. ... 89 Plans. 44 the same. South side of the nave and Sketch of East end and North side of the Church ... 90 45 the same, interior, looking towards the Chancel !)1 40 the same, interiors ... ... ... ... 92 47 Stit.o, la CattoUca, South Side ... ... ... ... 95 Plan and Elevation. 48 the same, interior ... ... ... ... 90 49 Ros.SANO, Church of S. Mark, East end and interior and Church of the Patire at Coregijano 97 Plan. 50 Ro.ssANO, view from the tiarthex ... ... ... ... 9S 51 the same, view looking across the Clacrcti ... 99 52 Dedication Cross ai Stit.o ... ... ... 102 ILLUSTRATIONS, PLANS AND MAPS. xxi. TUNIS. Plate Page. 53 El Kep, panel ; and Tunis, capital at the Bardo Museum... 102 54 Tebessa ; staircase from the hasilica to the trefoil chapel ... ] 03 55 the same ; west apse and mosaic on the tomh of a Vandal boy ... ... ... ... ... 104 . 50 the same ; south and west apses, and mason's marks. Plan by Mr. Duprat ... ... ... 105 57 ITppENA, Cakthage, AND HiDi Abicii ; Fonts ... ... 100 58 Carthage and Zaghouan ; cella trichora at the Damns el Karita. fieneral view of the Zaghouan nymi)h- Oium. the cella and the roof ... ... ... 107 50 Henchitj M.\atrta, exterior (/eneral view ... ... ... 108 00 the .same; interior ... ... ... ... 109 01 SiDi Mohammed Ex. Gebioih, exterior North and West .sides 110 02 the same ; vietvs of the other sides... ... ... Ill 03 the same; interior ... ... ... ... 112 (54 llENCHTH Maatiua AND El (iEBioi^ ; vicws of the interior ; and tubes of j tottery used in the cor}. si ruction of EKlebioui ... ... ... ... ... ... 113 05 SoussE AND Henchir Maatria ; the coffee house at Sousse called the Koubba, interior ; and the exterior of the Zaouia at Maatria ... ... ... ... 110 I^]l Kef chiirch called Dar el Kus ; view of the nave and. apse ; the West end or narthex, and a key stone of an arch 120 The ap,se of the (/arrison chapel in the Byzantine fori, and one of the pillars sirpportinr/ the chavcel arch ... ... ... ... ... ... 121 Roman and Native ... ... ... ... 1 22 thesam,e; Byzantine ... ... ... ... 123 and Kairouan. Marble panel over the door of the (treat Mosque at Sfax ; and terra coiia plaques from near Kairouan ... ... ... ... 124 Central Tunis ... ... ... ... ... 125 Enfida church ; from Uppena and. Sidi Abich ; and from Sfax ... ... ... ... ... 120 Central 'Tunis ; Bou Ficha, Sbeitla, and two tomb stones from the Caucasus. Cross on one of the pillars in the French churcJi at Sfax, ob- tained from. BAispe ... ... ... ... 127 00 I^]l Kef cl 07 Haidra. 08 Capitals. m 70 Sfax anii 71 Consoles, 72 .MosAT(;s. 73 Crosses. MAPS. Sicily 15 Sardinia ... .^o Calabria ... 84 Tunis 114 xxu. BOOKS. Abattxo. Pr. : Puhlication on Stilo in ' Napoli Nobilissimfi.' Vol. xii., paf/r 11, Fehriiary 1903. Amari : Storin del Musulmnni di Sicilia. Florence, 1854. Antoniades : Ekphrnsis fes Agios Sophias. ConslnriiUmplc. lOOS. Barreca : Caiacomhe di S. Giovanni in Siracusa. Syracuse. 1900. Barth : Travels and discoveries in North and Cenlral Africa. 1857. Battifol : L'Abbaye de Rossano. Paris, 1891. Bertaux : L'Arf dans VHalie Meridian ale. Paris, 1904. Besta : Tm Sardegna Medioevale. 2 Vols. Rebcr, Palermo, 1008. BiscARi : Paferno, Prince of : Viaggio per tutti le antichifa della Sicilia, 1817 Brolo : Lancia di : Archbishop of ^fonreale. Sforia della Chiesa in Sicilia. 4 Vols. Palermo, 1880. Butler : Ancient Coptic Churches of Egypt. 2 Vols. 1884. Bury : Professor, Later Boman Empire and Eastern Roman Empire. Chat.axdon : Histoire dc la domination Normande en Italic ct en Sicile. Paris, 1906. Clarixval Coaimaxder: Article on Tebessa in the R^'ceuil of the Archico- logical Society of Co}istantine, Algeria, 1870. DElJ^TTRE : Rev. Pere. Cn PeJeri^iage aux rui)ies dc Cartage. Lyons, 1906. DiMARZO : Dizionario Topographico della Sicilia. Palermo, 1856. DiEiiL : L'Afriqve Byzantine. Paris, 1896. DUPRAT, M. : Article on Tebessa in the Receuil of the Archreological Society of Constantine, Algeria, 1897. PiNLAY : Greece under the Romans. Freshpield, Dr. E. : In Archrpologia. Vol. xliv., p. '.iH'.i. Gay : Ij'Italie Meridionale et Vcmpire Byzantin. Paris. 1904. Garufi, Prof. C. A. : L'Archivio Capitolarc di Girgenti. In Archivio Storico Siciliano, 1903. GiovANXi. V. DI : Castello e la. chiesa della Ii'avara di S. Filippo a nmrc dolce in Palermo, in Archivio, Storico Siciliano. Palermo. 1897. Gregorovius : Rome in the Middle Ages. Trans. Hamilton. GsELL. S. : Les monuments antiques dc VAlgerie. Paris. 1901. GSELT,. S. : LJAlgerie dans Vaniiquite. Algiers, 1903. BOOKS. xxiiL (^LTEitiN, V. : Voyaye dund Id Iie(/eiice de Tunis. 2 Vols. Pari', 1802. TloDGKix : lluly and kcr invaders. Hoi/rziNGEK, H. : Die AKchrisllichc Archilectur. Slulliiarl, ISS!). .JacIvSox, T. Cj. : DaJnuilia the Quarnero and Isiriu. ;i Vols., 1S87, Kocii, 1*. : Byzanlini^che BeamtentUcl. Jena, 1903. Kaestner : De hnperio Consiuntini 111. Lcipsic, 1907. La Marmoua : lie de Sardaujne. Turin, 1851. Lai'otkk. a. : U Europe et le Saint Siege (Lv Papc .Jean VI 11.). Paris, 189.J. Lenormant : Jm Grande (Irtee. Paris. 1881. .Maktijoye, F. : L'Oeeident a Vvpoque Byzantine. Paris, l!»Ol. Mas Latrie : Tresor de Chronoloartment of .intiquities in the Re(fe)icy of Tunis. Mli.AiAN : History of Christianity. Mo.sSHEl.M : Eeetesiasti<:al History. Indited by liistiop Stubbs. Murray : Handboolc for Sicily, 1890. Neale, .1. M. : History of the Holy Eastern Church. Orsi, Professor P. : Articles on Canierina and Muccari in Byxayitiniselie Zeitschift, 1898-1899. Pas1'ATE.s : By-antince Meletai, Constantinople, 1877. Patricolo : Article on Delia Chapel in Archivio Storico Siriliano, N.S. Anno V, Palermo, 1880. Pirri : Sicilia Sacra. Piedesel, .J. H. Vox : Travels Ihrout/h Sicily, translated by J . II. Eorslcr. London, 1772. PoDOTA : Dell Oriyine stafo e proyresso del rito Greco, 1758-0!. PoDU, Sir Rexxell : Princes of Achaia and the Chronicles of Morea, 1890, 8AIXT NoM, ABBii UE : Voyoye Pitloresqae Naples et la Sicile, Paris, 11 So. Salinas, Professor : On the Christian Lamp in the Puler)no Museum and on the walls of Mount Eryjc, in Archivio Storico Siciliano, 188o. ScAXo, Dr. D. : Storia dell Arte in Sardeyna, 1907. ScoBAR : Caialoyus Episcoporuni Ecclcsiw Syracusance. Seriziat, Commaxder : Article on Tebcssa in the lieceuil of the Archmo- loyical Society of Constajiline, Alyeria, 1868. Sladex : Sicily, 1904. Theouoret : History of the Church. Bohn, 1851. Wroth, W. : imperial Byzantine Coins in the British Museum. xxiv. ADDITIONS AND COEEECTIONS. Page. Line. 23 There are no signs of a rib. See explanation in i'oolnote 5 to page 104. 12-16 The Cathedral at Syracuse is now in process of restoration and some recent excavations at the Chapel of S. Alarcian have brought to light the remains of the Noruiaii basilica. These works and the excavations at S. Theresa will be noticed in Vol. II. 24 Footnote 1., omit " tmd OSO." Footnote iii. in the third line for " Constans II." read Constantine III." As the result of further research I have come to the conclusion that the Sta. Theresa Chapel belongs to the 5th Century and is consequently earlier than the others. 26 Footnote 1., for " 598 " read " 098." 31 9 For " Amalo " read " ^Vmato." The authority for this statement is not Amato but Geoffrey Malaterra. In the last line but 7 I should have said that the stone vault now appears to be a late addition. 36 18 For " S. Orsola " read " S. Spirito." 37 Views of the interior of the Trinita de Delia and a note on the inscription on the apse will be found in Vol. II. 71 1 These are real mason's marks. What I took to be a spear and the reed and sponge are a chisel and mallet. Simi- ' lar marks in the cloister at S. Honorat and in the Kircherian Museum at Rome are illustrated in Vol. II. 87 The site of Cassiodorus' Monastery seems to have been at u place called Copauello, where Professor Orsi recently found the foundations of a trefoil chapel. It lies just below Staletti. 99 The Chm'ches of S. Joseph at (J acta, S. Costanzo at Capri and S. Maria delle cinque torri at Monte Cassino are illustrated and described in Vol. II. 122 bvit three The French name for this ornament is Coulicole. 123 Capital 23. Dviplicates of tliib pattern will be found in the Capella Palatina at Palermo and at Caix'o in the Mosque of Touloun. I i'AYWw^-vV 1. t'*M '^-^ J^A VARA, (fiear Girgenti ) Chapel door in the N'ormon Castle \ Frontispiece NOTES UPON SIX ANCIENT CHURCHES IN EASTERN SICILY. The ancient episcopal sees in Sicily were Syracuse, Camerina, Girgenti, Triocala, Lilybeo, Palermo, Termini, Cefalu, Tusa, Tyndaris, Messina, Catania and Lentini.^ Syracuse as the oldest and as the seat of government till the Saracen conquest became the most important and in the 8th century the see was raised to an archbishopric with metropolitical jurisdiction over Sicily. In Syracuse itself the principal churches of interest are the rock cut chapel of S. Marcian, an ancient basilica above it now forming part of the nave of the Norman church of S. Giovanni, the cathedral, and fragments of the Norman church of S. Lucia. The following ancient churches described in these notes are in the outlying parts of the diocese : at Priolo a village about half way between Syracuse and Porta Augusta, at S. Theresa, a village on the railway to Noto, and Maccari, a deserted city on the lagoons near Pacehino and cape Passaro. The other churches described in this chapter are situated in the dioceses of Camerina and Taormina. PRIOLO. Priolo is a station twelve kilometres north of Syracuse on the railway to Catania. The hamlet, a poor place with primitive houses, stands near the sea shore at the head of a wide bay between the promontories of Magnisi and Augusta. The former, known in classical times as Thapsos, claims to be the site of the earliest Greek colony in Sicily. The Athenian fleet of the great expedition against Syracuse is supposed to have sheltered in the roadstead protected by a sandspit from the S. and S.E. winds. The neigh- bourhood is full of classical remains, and one of the most important, 1. They varied from time to time, and at the Vandal Conquest of Sicily there were only nine. '^": :,:■'': .' PRIOLO a mass of ruined masonry locally known as the torre de Marcello is, according to legend, either a trophy to celebrate the victory of Syracuse over the Athenians or a monument to commemorate Marcellus' expedition against Syracuse. Dimarzo describes Priolo under the article on Mellili, the site of the ancient Hybla in the hills behind the village, and gives the district two other names, Aguglia and Mostragiano. The church and the adjoining monastery of six cells, stand in a little close a few yards from the high road leading to Catania, about three quarters of a mile south of Priolo. Dimarzo describes it as a noble and most ancient temple, dedicated to S. Phocas, built of square blocks made of stone, and spoken of by Scobar and Pirri as built by one Germanos, a bishop of Syracuse in the fourth century.^ The local legend about this S. Phocas agrees substantially with the life of S. Phocas, martyr, bishop of Sinope in the second century. He was a patron of sailors."-^ Both church and monastery are abandoned though the former is occasionally used and served by the parish priest of Priolo. The latter is resorted to by pilgrims at certain seasons of the year in connection with the agricultural festivals and thanksgivings. The old custodian or lay brother keeps the key and lives in Priolo. According to Dimarzo the church was damaged by earthquake, possibly in the twelfth century, when the Dasilica over S. Marcian's chapel was destroyed. The ground plan of the church is of the simplest character. It is a basilica with a nave of five bays terminating in a semicular apse covered with a sernidome. The aisles on each side had no apses but ended in square walls. The north aisle has been entirely destroyed exposing the nave arches. These arches are round and the square piers supporting them are made of massive well-cut stones in the style of good Roman masonry. The accom- panying sketch of them is taken from Professor Orsi's article. 1. For the reference to S. Phocas church see Sicilia Sacra by Pirri, ed. 1733, vol. I, }i. 603. Also Dizionario Topograjico della Sicilia by Vtto Amico, edited and translated by G. Dimarzo, Palermo, 1856, vol. TI, p, 77, Article Melilli "nel Masrojanno linalmente si osserva il nobile ed antichissimo tempio di S. Foca Martire fabbricato di pietre quadrate di cui parlano Scobar nel Catal : ed il Pirri che riconosce orij^ine di Germano Vescovo di Siracusa verso il IV Sec. di Cristo e dove riposano le spoglie del S. Martire " 2. Dictionary of Christian Biography. M^ 'Jh ■'i. v;^Js\ ^-itoX <{\ PRIOLO. ^. ,;, . ,-, _ ( S. Phocas.) above: — West front: cloister sho7ving south door of the church and small 7vood door on the right leading to the cells above. below: — N'ave arches on the north side showing the three low'er courses of the north aisle roof now ruined: main entrance unth a cross above the arch. \,^^ -J' *r Elevation N. side (OrsiJ. V~A Plan f OrsiJ. To face page 3. PRIOLO 8 They are no doubt the square stones referred to in Dimarzo's description of the church. The nave and aisles were originally roofed with stone barrel vaults.^ The nave vault sprang from a projecting stone cornice at 2'90 met. above the floor and close above the crown of the arches. These arches separating the nave from the aisles have been filled up with loose rubble, covered with cement and whitewash, and are now only 2*25 met. high above the floor ; they are also the same width at the spring of the arch. From these measurements, I conclude that the present tiled floor is about a metre and a half above the original floor. When the vault was broken, the nave was raised to almost twice its heiglit and covered with a plain timber and tile roof. In this addition there are three windows, one over the west door, and two others opposite one another on the north and south sides in the clerestory about midway the length of the nave. The window on the south side is now blocked up by the monastic buildings, showing that they are later than the raised part of the nave. The apse arch is 4'48 met. high from floor to crown, 3*83 met. wide and springs from two single moulded cornice stones, five inches thick 2*90 met. above the floor. The semi-circular apse, 2 met. deep, is covered with a semi-dome, and had one small square-headed light in the centre, now blocked up. "75 met. from the eaves outside, the apse wall recedes slightly from the plumb, and is composed of cement, small stones and pieces of brick. The rest of the apse is made of rough hewn stones and rubble and has a low flat tiled roof. Almost all the outer casing of dressed stones has been removed, and only portions of it remain on each side of the base to a height of a metre above the ground. The main door at the west end is 1*95 met. high and 1*32 broad, with a round discharging arch over it, not visible from outside. The jambs of dressed stone moulded plainly are comparatively modern. The letter " I," part of an inscription, is cut on the lintel. The west front outside is quite plain and covered with plaster and on each side of the door are two rough benches of stone. 1. This feature points to a very early date and possibly Coptic origin. Ancient Coptic Churches of Egypt. A. J. Butler, 1884, vol. I, p. 9. 4 PEIOLO The sketch shows the ruined north aisle. Above the arches there are three rows of large dressed stones inclining slightly inwards, forming the lower courses of the vaulted roof of this aisle. Above these stones on the roof is a heavy mass of cement inclining gradually from east to west and showing traces of a gulley pipe or conduit to carry water from the roof to a well, on the south side of the west door. The entrance to the monastery is at the south-west angle of the nave, by a round-headed door leading to a square vestibule. Above the door is a lozenge-shaped stone, with a cross carved in high relief. This stone is set in cement. The vestibule has a plain vaulted barrel roof of stone and leads into a courtyard, closed on the south side by a wall, on the west side by the wing of the monastery, on the east by outbuildings, and on the north by the refectory and cells over it. The refectory with the cellar and two vestibules adjoining it originally formed the south aisle of the church. The west bay is a cellar, the second and third bays are the refectory, the fourth a porch giving access to a door in the south side of the nave and into the cloister. The fifth bay was apparently used as a storeroom and led to the kitchen and scullery. The arches on the south side of the refectory are partly filled with masonry, leaving square windows and arranged so as to provide window seats between the piers. The tables on either side are raised on a low dais, leaving a tiled gangway in the centre. A small cloister of three round-headed arches runs along the north side of the court in front of the refectory, extending from the entrance to a stairway leading up to the cells on the first floor above the refectory. This stairway consists of six stone steps rising to a landing and then nine steps more at right angles lead up to a corridor on the first floor. This corridor is lighted by windows at either end and gives access to four cells over the refectory, and leads to two more cells in a wing on the south side. A narrow door on the north side leads to a small wood gallery at the west end of the church. I should add that, excepting the cross over the main entrance and a raised cross on a trefoil base over the store room door, 'f^v^'ovT'T^BE^)*" ' , ^ r^'V:'-^' -'rv'':?!ftp^ ■ " *r°5y,> .C' ,.f P» f , ^■#¥-^^ .»• A f, X \-^'iV>\ tj\ 3 & 4. STA. CROCE IN CAMERINA. f Bagfio di Mare ) J. North side; and one of the circular vents drilled through the masonry. 4. My wife holding a reed in one of the vents on the south side. Plan ( Orsi). To face page j. CAMERINA 5 there are no traces of ornament, decoration or mason's marks about the buildings. Professor Orsi visited the monastery more than once and conducted a series of excavations ; he has described it in detail and I have used his plan.^ In its original state the exterior of the church probably resembled the old Servian church at Mostar in the Herzegovina. A good idea of the interior can be obtained from the photograph of the nave of S. Giovanni in Sinis in the chapter on Sardinia. S. CROCE IN CAMERINA. The so-called Camerina chapels lie on the bank of a small stream between the village of S. Croce in Camerina and the sea. Dimarzo in a footnote to the article on this place, says : " Non lungi della vasca (a fine fountain just outside the village) " osservansi por presso un orto avanzi di antico bagno di tre stanze " composto di pietre quadrate senza calce e macerie di simili " fabbricati sin al mare ; puo credersi aver ricevuto acque dalla " vicina conserva." The antico bagno is, I have no doubt, the Vigna di Mare chapel and that is the only mention of these remains I have found in the authorities. There are two chapels, one called Vigna di Mare, on the bank of the stream a short distance west of the town near this fountain, the other called Bagno di Mare is about a mile and a half further south on the bank of the stream and among the sand dunes, about a quarter of a mile from the seashore. The chapels and adjoining buildings are so similar in character and construction as to leave no doubt that they were built at the same time and by the same architect. They have the following principal features in common ; a square nave covered by a dome, chancel, transepts and an annexed building on the fourth side consisting of two chambers, the ground plan thus forming a Latin 1. For his article on this and the other churches, see Byzanlinishe Zeitschii/t, p. 1 publication of 18th January, 1898, and p. 613 publication 1899 iVlII Band) Leipsic Teubner. 6 CAMEEINA cross. The only diflference between them is, that the chancel at Vigna di Mare is on the north side, and at Bagno di Mare it is on the south. The land they stand upon belongs to the princely family of St. Elia, to whose agent at Vittoria 1 was indebted for permission to visit the locality. The nave of the Bagno di Mare is a little square chamber, built of large blocks of cut stone, covered with a dome made of eleven rows of square blocks, in decreasing sizes from spring to crown. The chancel is on the south side and the transepts on the east and west sides ; but only the west transept remains perfect. The dome is not supported on either squinches or pendentives. It is to use a colloquial phrase simply dumped down on to the vaults of the chancel and of the two transepts on three sides, and on the wall of the narthex on the fourth side. Though the diameter is considerably greater than the square nave little brackets or consoles are introduced into the angles of the nave as if to support the dome. In fact they are quite meaningless and in the existing edifice support nothing. The peculiarity will be best understood by reference to my sketch and photograph of the bracket at Vigna di Mare, where I have placed a white cloth on the top to show that it does not support any superstructure. Possibly these brackets were intended either to carry the stone ribs of a cross vault or wood beams for a flat roof. In either case the dome would then be an addition. Another curious feature about them is that each has two large holes on either side bored right through and also through the dome. It will be noticed that the transepts are not built true to the centre of the nave. The east transept has disappeared altogether, and the arch leading to it is blocked up. The west transept is perfect, has a square end and is covered with a fine barrel roof of well cut stones laid in symmetrical rows ; only a portion of the west wall of the chancel stands, and the arch to it has been filled up and a modern door made in it. A small door in the north wall of the square nave and on the east side of it leads to the two adjoining chambers. Neither this nor the present main door are in their original position, but both have old jambs with mortice holes. The original access to the STA. CROCE IN CAMERINA. ( Vtgna di Mare) North side spring of the cross vault ( Vonte d'' aretes) which originally covered the nave. There is no rib to the vault (p. 104 foot note ^ ). Above the spring a white cloth; above the cloth the square stones of the dome. The t7Vo breaks on either side of the spring are the vents pierced through the masonry showing that they were drilled after the building was erected ; on the left and right the chancel and transept arches. Plan f Or si J. To face page 6. ;^^\AL wu M-^f^^/ CAMERINA 7 church seems to have been by a door in the east wall of the first chamber or narthex, and then by the door m the partition wall between the narthex and the nave. The main building was lighted by little square windows in the dome over the chancel and transepts. The two chambers are nearly square, that adjoining the nave being slightly larger than the other. Both are covered with waggon vaulting of well-cut square stones, the spring of the vault receding three inches from the wall and making a kind of cornice. The former may have been the narthex, the latter the baptistry. Each chamber was lighted by little windows or apertures of no interest. The window on the east side of the narthex above the main door has a bevelled cill and below it are two little apertures vhich look very like medieval lepers squints cut slantwise. The following are the measurements : — The nave 3'80 met. square. The east and west arches are 2 met. wide, and 2*50 met. high above the present floor. The west apse is 2"80 met, long. The crown of the dome stands 5 met. above the present floor level ; the latter has not risen much above the original. The first chamber is 2*90 met. long from south to north, and 2*80 met. broad ; the height of the vault 4'55 met. from the floor ; the second chamber is 2*60 met. square and 4"5 met. from vault to floor. The arch between the chambers has been broken, leaving an aperture 7 feet wide. The average width of the walls is 0'65 met. The plan and elevation of the chapel of Vigna di Mare are similar to the Bagno. The two adjoining chambers are at the south end of the building instead of the north, and the dome is of higher pitch and more perfect than the Bagno. The chancel and transepts have been entirely destroyed. The dome is composed outside of large blocks of cut stone, and inside of smaller cut stones arranged in nine rows and diminishing in size from spring to crown. The lowest row is composed of narrower stones, and they rest on the chancel and transept arches and on the south wall. The peculiar consoles or brackets in the angles of the square occur in this building also and they have the same holes bored through them. These holes are about 6 inches in diameter and are carried right through to the outside of the dome. In the photograph my wife is holding a reed in one of them. Professor 8 CAMERINA Orsi suggests that they were intended for poles to support a tent or awning, but I think they were more probably cut when the Saracens turned them into baths, and that they served some purpose connected with the heating. The first chamber or narthex adjoining the nave is in fair condition, but the second is almost gone. The entrance to the chapel was by a door in the wall of the narthex and the second chamber was connected with the first by a door in the party wall. Both chambers have strong barrel vaults made of well -cut square stones, and there are traces of cement on the roofs outside. The nave inside was lighted by long narrow slits in the dome placed at the cardinal points over the transept and chancel arches and the south wall. The original floor of the nave now covered with fallen stones and rubble, was about three quarters of a metre below the present level. The measurements are as follows : — The nave 3*65 met. square ; average width of walls, '90 met. ; chancel and transept arches, width 2'10 met. ; height above the present floor, 1"70 met. ; height of the dome inside above the present floor, 4'30 met. ; length and breadth of the first chamber, 3*40 met. and 3"75 met., and of the second chamber, 2*5 met. and 3"7o met. The chapels at Malvagna and Maccari are so similar in general plan, appearance and construction as to justify a conjecture that they were built about the same period and by the same architect. The materials used in the latter are chiefly large cut stones, probably taken from earlier buildings, and in the former, rubble, bricks, blocks of lava and cement. The Maccari chapel is consequently in much better condition ; moreover it is not exposed like Malvagna, to the disintegrating effect of frost and snow. Both buildings owe their comparatively good state of preservation to a thick layer of cement of a very hard kind covering the domes. MALVAGNA. The village of Malvagna is situated about half-way between Castiglione and Eandazzo, on a spur of the Nebrodian hills, facing the northern slope of Etna. Edrisi does not mention the village, ■•.^'^. ..A ■S^^':.^!^ MALVAGNA. The S. W. angle of the chapel; the S. apse pierced by a modern door. My wife watching an eruption on Etna On the plan ; a, modern door; b, ancient door now blocked; c, niche. The fragments on the East side mark the room or cell: see plate 7. N Plan (E.F.J To face page 8. nave W Ok :i ^'ioi^*: rn \'r "A \\'t \j<\\iiVb !\\\ .\ji TjtV ' ■:\\ ^ 3iVii»<^V«\*>\ 7. MALVAGNA The S.E. corner of. the chapel showing the S. apse with a modern door, the E. apse and a part of the wall of the cell. To face pai:[c g. MALVAGNA 9 though he speaks of the neighbouring township of Mojo. Dimarzo mentions Malvagna and also the chapel as " una fabbrica quasi intera appellata cuba dai terrazzini,'' and speaks of extensive ruins of walls, of cisterns, tombs, urns, lamps, and money found in the district, indicating an ancient locality of importance. The chapel is not mentioned by either Fazello or the Principe di Biscari ; Murray's handbook speaks of it as one of the few relics of the Greek Empire now extant in Sicily. The chapel, locally called the Cuba, is situated (and I describe its position with some detail, since I had much difficulty in finding it) on a plateau at the foot of a hill below the village of Malvagna. It is about half a mile north of the high road leading from Francavilla to Randazzo and is reached by walking up a little brook spanned by a culvert two hundred yards east of the entrance to the village of Mojo. The latter and Malvagna gives their names to a station on the Circum-iEtnea railway, about two miles south of Mojo. The chapel is a square chamber with semi-circular semi-domed apses on the west, south and east sides ; the north side is a plain wall and contained the main entrance which was not placed in the centre, but a little on one side so as not to face directly on to the altar in the central apse. The square is covered by a flat dome resting on the north wall, and on the arches of the apses, and supported in the four angles on squinches. The door cut into the central or south apse, shown in the photograph, and the remains of a chamber on the east side are of later date. The features immediately noticeable inside the building are that the lower half of the flat dome is octagonal, and that the eastern apse is not only smaller than the others, but is on the south side of the east wall leaving room for a little niche on the north side. What this niche may have been intended for it is impossible to say. But if these apses were intended to be used as prothesis, altar and diaconicon respectively, then the eastern apse would be the prothesis and the niche may have some ritual significance. The dome is well-made of small blocks of stone and lava, roughly cut, arranged in eleven rows, and set in cement. The roof outside is also covered with a thick layer of this material. The squinch arches in the four angles are made of lava blocks similar in construction, but of rougher work than those at Maccari. 10 MALVAGNA The arch of the west apse is made of thirteen well-cut blocks of stone and lava placed alternately like those in the S. Theresa, Cuba. Of the little square chamber on the east side, only the north wall and a fragment of the south wall on the lop of the roof of the east apse remain. From the latter circumstance I conjecture the chamber to be a later addition. The chapel is now used as a tool shed, but from a circular tank sunk in the floor and lined with cement, and portions of a square tank with a lipped gutter close by, it would seem that at one time it was used for a wine press. The following dimensions must be taken as approximate only : — The nave 4*48 met. square ; average width of the walls 1"75 met. ; width of the two larger apses 2"75 met. ; depth of the same ISO met. ; height of two arches 3"25 met. ; width of the east apse 210 met. ; depth of the same 1*30 met. ; height of this apse arch 3 met. ; main door, height 2'50 met., width 1*30 met. ; height of the dome above the existing floor 5*90 met. ; and it has eleven rows of stones from spring to crown. MACCAEI. Maccari is situated about ten miles south of Noto on the road to Pacchino and close to the seashore. The chapel stands on the north end of a narrow peninsula surrounded by the lagoon of Vindicari and connected with the mainland at the south end. The spot is locally known as " citadella," and, judging by the remains in the adjacent fields, must have been a place of considerable importance. This spot commands a fine view of the seashore to the east, the lighthouse and headland at cape Passaro to the south, and the hills over Noto towards the north. The country to the west, well cultivated, enclosed in stone fences and wooded with large olive and carob trees, has an appear- ance of considerable prosperity. There is now much malaria in the district, and the desertion of the town is no doubt to be attributed to it. Edrisi mentions it thus in his itinerary : " Da Marzameni a dahlat'ibn dikami (cala d'ibn dikami anche oggi " Porto Vindicari) sei miglia." According to Dimarzo this spot is the site of the ancient Ichana MACCARL S.E. front of the chapel. 17^1 Elevatioji and Plan ( Orsi ). To face page lo. ^M^>^'^ m 60 111 to ' ^ -y siderabl T1S it til m r , J •wll M. 1 .Q .W ^-^ft^ ^%»\ «i\. 9. MAC CARL South apse and side door, showing the step buttresses and the square projecting stones in the drum of the dome. To face page ii. MACCAEI 11 mentioned by Pliny, and not to be confused with Machara mentioned in Cicero's oration against Yerres with which Fazello identifies it. The latter, quoted by Dimarzo, describes it as " una salina .... alia dicui bocca il porto detto Fenico da Tolomeo, " Nanstatmo da Plinio ed oggi Vindicari .... Vedoiisi molte " vestigi;'. di edifizii .... Vi ha un tempio oibicvdare ed a volta " travagliato con antique lavoro da piette quadrate e talmente ancora " intero che non in antiche ma nei tempi del Cristiani sembia " costruito al Salvatore cui ora e addetto.' He also mentions another building of the same form which has fallen in ruins. Paterno drily observes that the ruins do not repay the fatigue a visit entails. This chapel is also a square chamber covered by a flat dome and semi-circular half domed apses on the south, west and north sides, forming chancel and transepts. It is built of large blocks of local stone roughly cut, and the dome and apse roofs are covered with a very hard cement. Some of the stones are so large as to suggest that they have been taken from older buildings, though Professor Orsi does not consider that there are any traces of a classical town on this site. There are three entrances. The main door is placed a little to the side of the east front, and not in the centre. There are two small round-headed doors in the south and north walls on the east side of the respective transept apses. This arrangement of side doors will be found in the Norman Greek chapel, near Castelvetrano, called the Trinita di Delia. The top of the east front has perished considerably, but enough remains to show that the wall was carried up towards the centre in the way of a rudimentary fa9ade. The dome is counterweighed by a low circular drum outside, and further supported at the four angles of the square by little step- shaped buttresses. On the south side of the drum there are two projecting square stones, visible in the photograph, but their purpose is not clear. The dome inside is well built of ten rows of cut stones of diminishing size from spring to crown, supported by squinches in the angles. It is in every respect better finished and more substantial than the dome of the chapel at Malvagna. 12 MACCAEI The interior was lighted by little square windows above the apse arches and in the centre of the east front. There are also traces of a window in the north apse visible on the outside, though I could find none inside ; but it must be remembered that all these windows are merely small square apertures without ornament, and they may have been blocked up in later days when the buildings were used as storehouses and shelters for cattle. The whole of the inside of the church is blackened with smoke, and there is no trace of decoration or painting. The dimensions are as follows : — The nave 6*40 met, square ; average width of the walls 1*5 met. The north apse 3"50 met. wide, 2'50 deep, and the height of its arch 4*10 met. The other apses are 2 '75 met. wide and 2 met. deep, and their arches are 4"10 met. high. The main door is 3*50 met. high and 2 met. wide ; the side doors are 2" 10 met. high and •98 met. wide. There are 10 rows of stones in the dome from spring to crown, and the latter is about 6*50 met. high above the existing floor level. I should add that the floor in the apses is raised a step above the nave. S. THEEESA. The S. Theresa chapel, locally known as the Cuba, stands in the middle of the farm-yard or fattoria of Mr. Vincent Vinci. It it an hour's drive south of Syracuse, near the railway to Noto, and about a quarter of a mile from the sea shore. It is almost completely buried in the ground, and a sixteenth century martello or watch tower has been built on the dome. This is now used as a country house by the Vinci family. I have been unable to identify this chapel in any of the old authorities. It is situated a short half-mile to the east of the railway from Syracuse to Noto, near the S. Theresa station. According to Dimarzo all this district, bordering on the sea shore from Lognina by Longarino to Cassibile is full of ruins, aqueducts, tombs and baths of the classical period. The ground plan of the S. Theresa Cuba, is similar to that of the last two chapels, that is to say, the square nave is covered with a low dome and has on three sides chancel and transept apses, and a porch or narthex on the south side. The roof of the latter 10. MACCARI. N. W. angle of the chapel, showing the S. 6- W. apses: the step buttresses,: and on the right, one of the square stones let into the di-uni of the dome. To face page 12. Ait ire 10 1 the latter is ai S. TH .s\ v>,»\vk»\(r^; ??4^ ■■■^■ssnaffUKi-' (X /.\ ■,A .11 1 ^'iVA orsJb>i7s?,t\\v^^v\ a" Mazzara, the 16th century chapel in the Cathedral at Mte. S. Giuliano, a chapel in the church of the Addolorata, also o:i Mte. S. Giuliano and the chapel of S. John, attached to the famou-< Norman church of the Madonna oiitsiile Trapani. ^5 O £? (O^ •x. ■^ SICIL Y. V To face page ij^. V /,OS\H .\ \ ^o Vi<\ Yi'B^ CsX. 12. To face page ly. THE SELINUNTO LAMP 17 Blessed Virgin should be called the Mother of God or as Nestorius held only the Mother of Christ.^ During the fourth and fifth centuries the Church was distracted by the Donatist and the Arian heresies. The Sicilian church was directly concerned in the former, and Bishop Crestus of Syracuse, the first Sicilian bishop whose name is recorded as taking part in a Council, was summoned by Constantine to the Council held at Aries in 314, to judge the Donatists. He attended with his deacon, and the terms of the Emperor's letter of summons have come down to us.^ There is an interesting relic in the museum at Palermo relating to the Donatist controversy. During some excavations undertaken at Selinunto in 1882 in the temple ascribed by Schubring to Hercules, a bronze lamp was discovered between the tenth and eleventh pillars on the north side of the nave counting from the north-east corner. This interesting object, found broken into two pieces, was restored by Professor Salinas and deposited in the museum of Palermo. Owing to the inscription Deo Gratias, he attributes it to an orthodox community in the fourth or, at the latest, the fifth century.^ Deo Gratias was the catchword or war-cry of the orthodox party, Deo Laudes that of the Donatists, as appears from a passage in S. Augustine of Hippo's psalm giving a history of the heresy. " Vos Deo Gratias nostrum ridetis. Deo Laudes vestrum homines " plorant." ^ There is no evidence to show whether the Christian community here survived either at the time of the Saracen invasion or the Norman conquest. But this lamp, a gravestone in the Palermo museum and numerous tombs and fragments, prove the existence of what must have been one of the very early Christian colonies, and the explorer Benndorf claims to have found traces of a chapel in the ruins of the temple locally identified as A. 2. Bury's Later Roman Empire, vol. I., p. Ifc9. 3. Storia della Ohiesa, vol. I., p. 191, from Eusebius. 1. Archivio Storico Siciliano, New Series, Anno VII., fasc. I.-V., p. 13;.'. 8(3 Article, by Prof. Salinas. There is another lamp of the same pattei n, but without the inscription, in the museum at Cagliari in Sardinia. U[on ihe controversy, see Bury, vol. 1., p. 194. 2. Pt\^ 13. ROSSO LINT. Rock ciit church. I. Interior from A. E.F. 2. The main and 2 side doors. J. Interior from B. Plan (Or si J. S. HILARION 19 Whoever the founder of S. Phocas may have been, it is certainly a building of great antiquity. From the character of the masonry and the stone waggon vault of the roof it might certainly be earlier than the reign of Justinian. I hazard a conjecture that it was built by some of the monks who sought refuge in Sicily during the fifth century, either from Italy after the sack of Rome by Alaric, or from Egypt where the Eutychian heresy was then the national religion^ or, as seems to be more likely, from Africa after Carthage was taken by Genseric.^ It will be remembered that Sicily, and in particular this part of it near Syracuse, was on the trade route between Ephesus, Syria, and Egypt, and Rome and Gaul, and in the way of constant and considerable intercourse with the great centres of Christian activity at Antioch and Alexandria. Records of migrations of the more celebrated of these monks have come down to us. S. Hilarion and his deacon visited southern Sicily in 363, and appear to have spent a year in the island in the neighbourhood of the modern Pacchino on the coast, or in the hills behind Spaccaforno. The locality usually pointed out as the scene of his labours is the Val d'Ispica, where there are numerous caves The rock-cut chapels and dwellings at Pantalica and Pallazuolo Acreide may belong to this date. One of the largest and most accessible of these rock-cut churches is at Rossolini, a station on the Syracuse Vittoria Railway about half-way between No to and Modica. This church is a basilica of irregular shape with a nave terminating in a semi- circular apse and divided from the aisles on both sides by round arches on square piers. In one aisle there are some tombs cut in the side, and at the end of the other aisle there are the remains of what seems to have been a kind of canopy in the roof. This church has now unfortunately been turned into a wine cellar, and the paintings which are said to have existed, completely destroyed by smoke. I reproduce the plan given in Professor Orsi's article in Byzantinishe Zeitschrift, and a photograph of the front and of an arch and pier in the nave. 1. Finlay, Greece under the Roman$, vol. I. , p. 213. 2. Storia della Chiesa, vol. I., p. 404 : S. Eufinianus, mentioned in the life of S. FulgentiuB, was a refugee from Africa. 20 S. PHILLIP D'AGIEA About fifty years after S. Hilarion another noted man known in Sicily as S. Philip d'Agira or Argiro, visited the island ; the country also is full of places dedicated to or named after S. Calogero, which I take to be merely a generic name and to indicate the presence of a monastic settlement or a hermit's cell. The precise date of S. Phillip's arrival in Sicily is not known but believed to be shortly after Alaric's sack of Eomewhen many priests and monks fled from Italy to Sicily. S. Phillip came from Rome by way of Eeggio and settled at Agira on the west slope of Etna. The tradition concern- ing him is that he was sent to purge the regions of Etna and Agira of the wicked spirits who had come to infest it after the destruction of Solomon's temple in Jerusalem.^ This legend refers no doubt to paganism which lingered for a long time in the more remote parts of Sicily and in spots where the later Greeks and the Romans feigned to locate the home and doings of the gods of mythology. I can find no evidence of a persecution by the Yandals, at any rate in this part of Sicily, so continued and systematic as to make it impossible for S. Phocas and the Camerina churches to have been erected during the fifth century. Finlay observes that Genseric's jDiratical incursions into Greece were neither very extensive nor very successful^ and the same I think may be said of the raids into Sicily, for Marsala taken in the first raid (440) and retained by Genseric when he purported to cede Sicily to Odoacer (476) is the only permanent settlement recorded. Upon this transaction Martroye observes that there is reason to think that Genseric did not cede Sicily, which he no longer possessed, but only renounced, in consideration of the tribute, all claims to the territories which he had previously occupied.^ There is, however, no satisfactory evidence that the Vandals ever had a permanent occu- pation, if indeed their naval forces were large enough to effect and retain it.^ In any case their domination over Sicily was too brief and intermittent to admit of a systematic persecution of the orthodox Church in Sicily as in Africa. Moreover, persecutions were not always unconnected with political motives, and in Africa the conciliation of a considerable section of the population may have made persecution expedient, if not a necessity, for I gather tbat the 1. Storia della CMesa, vol. I., p. 211, footnote 2. 2. Finlay, Greece under the Romans, vol. I., p. 271. 3. Martroye, p. 190. 4. Diehl, I'Afrique Byzantine, p. 13. THE VANDALS IN SICILY. 21 successful occupation of Africa, if not the coming of Genseric him- self, may probably be attributed to the Donatists. The Vandals therefore had a local and political reason for oppressing the Orthodox Church in Africa which did not exist in Sicily. Genseric died in 477 and was succeeded by his son Hunneric and his grandsons Gundamund and Trasamund. Hunneric occupied the Vandal throne for seven years and his reign was marked by severe persecution of the catholics in Africa. Gundamund succeeded him in 484 ; and in the same year Theodoric became the ruler of Italy, giving his sister Amalafreda to Trasamund in marriage and settling Marsala on her as her dower. Trasamund succeeded Gundamund in 496 and reigned in Africa till 523, dying three years before Theodoric, and leaving his wife Amalafreda surviving. These two princes, Theodoric and Trasamund, reigned over Italy, Sicily and Africa during the same period and for nearly thirty years. During that time an insurrection in Sicily (522) was crushed by Theodoric. Trasamund in Africa was succeeded by Hilderic, and Theodoric in Italy by his grandson Athalaric then a boy, Amalasuntha his mother being appointed Kegent. The feud between the reigning houses of the Goths and Vandals came about by Hilderic's treat- ment of Amalafreda. He had imprisoned her for conspiracy and, so soon as her brother Theodoric was dead, he put her to death. Hilderic continued to rule Africa till 531 when Gelimer seized his throne and held it for three years, being in his turn conquered and sent a prisoner to Constantinople by Belisarius. Belisarius then sent a mission to the Goths in Sicily claiming and eventually seizing Marsala upon the ground that by Genseric's cession to Odoacer it formed part of the Vandal kingdom. Justinian ignored Amalasuntha's protest^ and the conquest of the rest of Sicily was proceeded with and accomplished by Belisarius in 536. It was not, however, till sixteen years later (552) that the Goths were finally driven out and during that time Totila had conducted a successful campaign against the Greeks and raided Sicily for a short time in 548. There is very little evidence as to the condition of the Church in Sicily during this long period. S. Gelasius was Bishop of Rome 1. Martroye, pp. 173-174, translates (from Procopius) Amalasuntha's letter to Justinian. 22 THEODORIC from 492-496, that is at the commencement of Theodoric's reign. In 494 he addressed a long letter to the Bishops of the Sicilian Church reciting the desolation caused by the long wars and the insufficiency of clergy to minister in many churches. If the remedies prescribed in this letter may be taken to indicate the special abuses which had crept into the Church, the conditions must have been miserable enough. There is no evidence to show whether reforms were carried out or what ensued in the later part of Theodoric's reign, in Amalasuntha's regency, or during the sixteen years' war between the Goths and the imperial troops. The Church historian of Sicily concludes that in their dealings with the Orthodox the Goths were more tolerant than the Vandals, and so far as the early part of Theodoric's reign is concerned the conclusion is probably right. Theodoric found it expedient for political reasons to favour the Latin Church until the close of his reign when the relations between the Roman See and the East had changed. The severe laws against the Arians prompted Theodoric to retaliate with equally oppressive laws against the Orthodox in Italy. But he died before they were put in force, and his successors were otherwise engaged in a struggle with Justinian which ended in the expulsion of the Goths from Sicily and Italy altogether. I conclude that these three churches were built during the Gothic dominion and that the two at Camerina were copied from an African model like the chapels at Maatria and Hadjla (p. 108), possibly by Tefugees from Hunneric's persecution in 484. Belisarius came to Sicily in 534. He was then leading the Imperial fleet to the conquest of the Vandal kingdom in Africa. On the way the fleet harboured in Caucana, and though the precise locality is not known, it is believed to have been in the bay of S. Croce in Camerina. The plan and construction of the two little churches at S. Croce in Camerina are so peculiar and so unlike anything Byzantine as to make it reasonably certain that the architect was unfamiliar with, or at least uninfluenced by the Byzantine style developed in Justinian's reign. I have adopted the name for them chosen by Professor Orsi in his articles, and it is not inappropriate since the nearest village or township is S. Croce, but in fact they, as well as S. Croce, are some little distance from the BELISARIUS 23 site of the classical Camarana on the river Hii^paris, a place of sufficient imj)ortance to have had a coinage of its own.^ Moreover the village and the chapels are on a little stream in an adjoining valley, known in classical times as Oanis, and mentioned by Pindar. This stream flows down to the seashore between sandhills into an open roadstead, and the estuary seems to be the site of a town known under the names of Porto Longobardo,- Easacambro and Caucana. The roadstead was also called the asilo of Bricia. To this roadstead, as I have said, Belisarius brought his fleet from Constantinople on the way to Africa, and it was from Caucana that he sent Procopius to Syracuse to ascertain the condition of the military preparations of his opponents. Caucana has been described as Belisarius' quarters,^ and presumably must have been a place of some importance at that time ; but though the name sounds Roman, there is no evidence, so far as I know that is was a classical City, nor is it mentioned in the itinerary of Edrisi, though Easacambre is a Saracen name. On the other hand the name Porto Longobardo speaks for itself, and Dimarzo calling the village S. Croce di Easacambre says it contained a manor belonging to the priory of S. Lawrence and S. Phillip of Scicli which was suffragan to the Norman monastery of S. Phillip d'Agira,^ It occurred to me that Longobardo might mean Norman, and that there might be traces of Norman work in these chapels, but I could find none, and their general appearance is conclusive that they are much older than the twelfth century. The conclusion I have arrived at is that they are of Carthaginian or Egyptian origin and belong either to the period of the Gothic occupation, or that immediately succeeding the arrival of Belisarius, and before the importation of Byzantine ideas and models had time to influence local architecture. 1. Article in Byzantinishe Zeitschrift, p. 1. 'Publication of 18th Jan., 1898. The Italian Military survey of this district gives no help in identifying the ancient name^ and does not mark the site of these chapels. Camarana was in ruins by Strabo's time. It continued to exist till the second century. 2. For this name see Gibbon : ed.. Bury, vol. IV., p. 279, 1898. Walter of Malaterra IV. 16, cited as calling this spot Kesacramba. 3. Hodgkin, Italy and her Invaders, vol. IV., p. 602, and Martroye, L'occident a Vepoque Byzantine, Paris, 1904, p. 228. 4. Dimarzo Diet, of Sicily, vol. 1., p. 361, vol. II., p. 179, Ebbe il nome da un imagine di S. Elena madre di Costantino expressa colla croce nel anticbissimo castello come Pirri attesta. Del Casale di S. Croce di Easacambre fa menzione il diploma dell' Imperatore Errico V. nel anno 1195 datatoinKagusadove enumera i beni del convento di S. Maria Latina presso Gerusalemme cui successe quello di S. Filippo d'Argiro' dietro di essere Stata dai Turchi devastata la Palestina. 24 THE BYZANTirE PEKIOD. The three chapels at Malvagna, Maccari and St. Theresa are unmistakably Byzantine in design, and built to suit the Greek rite requiring prothesis and diaconicon. As to their date all that can be said is that they are certainly not older than the reign of Justinian or probably later than the occupation of eastern Sicily by the Saracens in the tenth century. My opinion is that they are not earlier than the middle of the seventh century for the following reasons. By the end of the sixth century the official language of the Empire, which in Justinian's reign was Latin, had become Greek. But the change from the Latin to the Greek rite in the Church of Sicily was much more gradual, and the final sex)aration from the jurisdiction of the Patriarchate of Kome, to that of Constantinople was not accomplished till well into the eighth century.^ It is no doubt the fact that during the early christian and mediaeval days Sicily was a bilingual country. From Belisarius' conquest it is easy to trace the gradual steps by which Greek took tJie leading place. Perhaps the most important of these steps was the expedition to Italy and Sicily of the Emperor Constans, the grandson of Heraclius,'^ and his residence for six years in Syracuse with a large part of his army and a retinue of officials and clergy who accom- panied the Imperial Court.^ At that time (663 to 668) the Greek element, if not in numbers, at any rate in importance, came to prevail over the Latin. Professor Bury introduces this Emperor's reign with the remark that the history of the successors of Heraclius is veiled in the most profound obscurity."* But from what we do know of this interesting Emperor, his eventful career, his remarkable personality and his 1. Gregory the Great, in a letter to the Bishop of Syracuse speaks of Sicilians, Latins or Greeks who accuse the Bishop of Rome of servilely imitating the customs of the Eastern Church; and the clergy in Sicily sided in the Councils of 649 and 680 with the Eoman See against the Byzantine authority. Gay, Vltalie Meri- dionale et V Empire Byzantin, p. 9. 2. The chapter upon this Emperor is one of the most interesting in the Later Roman Empire. A monograph about him has been written by I. Kcestner, of Jena, published at Leipsic by Teubner in 1907. 3. Storia della Ghiesa, vol. II., p. 23. 4. Theophanes calls him Constans; on his coins he is called Constantine. Professor Bury (vol. II., p. 281 ) calls him Constans II. In the latest publication about him by I. Kcestner at Leipsic in 1907 he is called Constans II. His son was known as Constantine IV., Pogonatus, the Bearded. But on his coins Constans too is represented with an enormous beard. 14. IMPERIAL BYZANTINE GOLD COINS. Of the Heraclian family,, from 6ij to 6g5. 1. Heraclms and Heraclius Constantt'ims, 6lJ — 614. 2. The same^ 6jo. J. Constats II ( Constantme III) and Consiantme IF', and the young princes, Heraclins and Tiberius 659—665. 4. Constans II and Constantine IV, 654 — 65g. 5. Constantine IV, 6^0 — 685. 6. Justinian II, 685 — 695. 7. Justinian II, 685— 6(; 5. ^^^sS'Sm^. These coins are described on p.p. I2g — 133. Nos. I, J, 4 and J are taken from the coins, and Nos. 2, 5 and 6 from casts. Constantine IV is named ^ Pogonatns,^ the bearded ; but the coins of his grandfather Heraclius, No. 2, and of his father Constans II Nos. J and 4, represent those sovereigns wearing enormous beards. Constantine IV himself has a short beard. The name ' Pogonatus'' was probably m,is-applied through a mistake of the historians. To face page 24. .i^r w",- ^?^?i v?.^ AYA-XYik^X ^^ A \ A^^v^^WiA ??i?i — Q7,^ 2, T •V^ vjwA ■>-vtv\ oT CONSTANS II. AT SYRACUSE 26 religious opinions, we may infer that his reception by the bishops of the Church in Sicily can scarcely have been cordial, more especially when we consider his edict upon the contro\ersial quest- tion of the day, the steps he took to enforce it, the interpretation the western Cliurch put upon it, and the part the Sicilian bishops played in resisting it.^ There are certainly no churches, nor for that matter are there buildings of any kind in Syracuse which can with certainty be attributed to the Greeks in this Emperor's reign, and lam inclined to the conclusion in view of the conditions and the disturbing events in the succeeding reigns, including at least one raid by the Saracens on this part of Sicily, that these churches are not older than the reign of the Emperor Leo the Isaurian, sixty years later. During those sixty years the Greek influence continued to increase. Bishop Zosimus who is credited with converting the temple of Minerva into the cathedral of Syracuse died two years before Constans arrived there. He was succeeded by one Elias, of whom nothing is known beyond the fact that he was Archdeacon to his predecessor and occupied the see for a few months. It was then occupied successively by George and Theodosius, both Greeks. The former, probably nominated by the Emperor who claimed the patronage, was afterwards killed in a raid made by the Saracens in 669.2 But it was not for some time after Constans' reign that the rite became Greek. Though the writings of Gregory, bishop of Girgenti, a noted scholar and composer of hymns, show that he was a Greek addressing himself in Greek to a flock who followed the Greek rite, he obeyed a citation to Eome upon a charge of 1. His edict was called the Type. The Lateran Council presided over by Martin which condemned the Type, was attended by the bishops of Lentini, Messina, Girgenti, Trioeala, Lilybeo, Taormina, Palermo, Tyndaris, and Lipari. It was held in 649. Storia della Chiesa, vol. II., p. 7. 2. Storia della Chiesa, vol. II., pp. 34 & 36. From that time onward the see was occupied by Greeks, and eventually was raised to an Archbishopric with jurisdiction over all Sicily. This precedence it retained, nominally at any rate, until the Norman conquest, when Count Roger appointed a Latin Bishop of Syracuse who had been consecrated by Urban II. to preside over the clergy of both Greek and Latin rite. See also Eocco Pirri, vol. I., p. 617 quoted by Barreca p. 89. And as to these two bishops see Barrecca p. 49, and Storia della Chiesa, vol. I., p. 43. 26 LEO THE ISAUEIAN heterodoxy concerning his writings, and was acquitted.^ The fact points to this bishop having lived during the sixty years between the reigns of Constans and Leo, when both Greek and Latin rite pertained in the Church in Sicily and before the connection with the Eoman See was severed.^ The ecclesiastical annals of Sicily between 600 and 800 show the extent to which the Latin and Greek rites were practised together. Basilian and Latin monas- teries existed side by side. During this period the Eoman See was occupied by the following Sicilians : S. Agatho 678, S. Leo IL 682, Conon 686, Sergius 687, and Stephen IV. 786. Conon was educated in Sicily and Leo IL was famous for his eloquence in both Greek and Latin. At the same time the church in Antioch had two Sicilian patriarchs, Teofanes (an Abbot of Baya, near Syracuse) in 681, and Constantine (a deacon, also of Syracuse) in 683. The dispute between the Eoman See and the Emperor Leo respecting the images is a matter of history. It affected the Church in Sicily indirectly. The Emperor, after being ex-communicated by the Bishop of Eome, retaliated by confiscating the patrimony in Sicily. I say indirectly, because the authorities do not seem to agree whether the Emperor in terms forbade any intercourse between the bishops and heads of monasteries in Sicily and Eome.^ But the practical result of the Emperor's confiscation was to place the Church in Sicily under the patriarchate of Constantinople, and so it came about that the liturgy of S. Chrysostom prevailed and continued in Sicily throughout the Saracen occupation till the Norman conquest, when Count Eoger and his successors introduced not only bishops from Gaul and 1. Gay, p. 10, and Storia della Chiesa, vol. II., p. 46. The date was 598. Among the Sicilians, who were prominent after the separation from Eome, should be mentioned S. Gregory Asbesta (Bishop of Syracuse), S. Joseph (the hymn writer), and Melo (Patriarch of Constantinople). Amari, a-o1. I., pp. 29 and 197. 2. About this time also the Sicilian bishops adopted seals with legends in Greek. Storia della Chiesa, vol. II., p. 72. 3. Storia della Chiesa, vol. I., p. 384, and vol. II., p. 101. See also Diehl, p. 509. It appears to have been the practice for the Sicilian bishops to meet under the presidency of the rectors of the patrimony. I should, pf-rhaps, not omit to mention an episode regarding the appointment of a rector by Conon, who occupied the Eoman See for a short period in 686, and the opposition of the Sicilians. The clergy claimed a right in respect to it which had been disregarded. THE BYZANTINE PERIOD 27 England,^ but also the Gallican liturgy, and that rite continued till the Council of Trent in the sixteenth century.^ It must be remembered that, though the liturgy used by the Greek churches in S. Italy before the Norman conquest was in calendar and ritual the liturgy of Constantinople, there is evidence that other liturgies were introduced by refugees from Syria and Alexandria after the Mahometan conquest of those countries. For instance, the only ancient manuscript that we have of the Syrian liturgy, of S. James, was found at Rossano, and the only manuscript of the Alexandrian liturgy, of S. Mark, was found at Messina. The date of these three chapels should be placed between the arrival of Belisarius or, more probably, of Constans, and the Norman conquest, that is from 650 to 1050. For the last 200 years of that time Sicily was actually occupied, or in continual invasion, by the Saracens, and I should prefer the second quarter of that period, say from 700 to 850. The chapels at Maatria and El Gebioui, described in the chapter on Tunis, were built on this same plan.^ They belong to the class of African building that may have served as model for these chapels, which were built, I con- jecture, by refugees from Carthage after the Arab conquest in the last years of the seventh century.t 1. A list of Bishops of Girgenti will be found in Arch. Stor. Siciliano N.S. Ann. XXVIII., pub. Palermo 1903 by my friend, A. Garufi, 2. Storia della Chiesa, vol. II., p. 448. For the relation in later times between Greek and Latin cler<^y iu the Levant, see Rodd's Princes of Achaia, vol. II., p. 272, in the Appendix ; and also an interesting? case before the Privy Coiinci] upon an appeal from Canada between Zaoklynski and Polushie, two Ruthenian emigrants. The judgment is reported in The Times of 5th December, 1907. The Greek rite is now followed, according to the Uniate profession, by a few scattered communities, chiefly Albanian, in Calabria and Sicily, and by the Greek settlement at Cargese in Corsica. There is, or was, an orthodox church and community at Messina. P.P. Rodota, in DelV origine, progresso e staio . . . del Hto Greco in Italia. Pub. 1758-63, may be consulted. Also Storia della Chiesa, vol. I., p. 176. 3. The name given by French architects to this kind of building is colla trichora or chapelle trilohie, and the trefoil plan was no doubt copied from such buildings as the chajjels in the cemetery of S. Callixtus on the Via Appia at Rome. 4. See footnotes 1 and 3 on p. 24; in the former omit 'and 680' ; and in the latter read ' ConBtantine iii.' for * Constans ii.' in the third line. In footnote 1 on p. 26, 598 should be 698. 28 WESTERN SICILY. From the landing of the Saracens at Mazzara in 829 until the surrender of Palermo in 1071, the Church in Sicily continued to use the Greek rite and to be, nominally at any rate, under the jurisdiction of the patriarch of Constantinople. There is an edict attributed to the Emperor Leo the Wise and published with variations between the ninth and the thirteenth centuries containing a list of the original Sicilian bishoprics. But as some cities were destroyed and others fell into the hands of the Saracens the bishoprics either ceased to exist or continued as titular only and by degrees the names of Sicilian bishops disappear altogether from the proceedings of the eastern Councils. Their position in the tenth century would I suppose be analogous to that of a Latin bishop in partibus infidelium.^ Nor do the Sicilian bishoprics appear in the Roman annals after a certain date for it will be remembered that the official connection between the Church in Sicily and Rome had been suspended since the beginning of the eighth century when the Emperor Leo confiscated the Sicilian patrimony of the Roman see. The two principal events in the purely church history of this period are the restoration of the images and the controversy respecting the patriarch Photius. Neither of these affected Sicily directly for the Sicilians favoured the images and the Emperor Leo and his successors, possibly for political reasons, took no steps to press their views upon them. The controversy regarding Photius developed into a political dispute between Constantinople and Rome and with the latter, as I have said, all official connection had been suspended for some years.'^ It would seem that after the fall of Rometta, the last Byzantine stronghold in the hills behind Messina, the Church in Sicily was left to itself, and all we know is derived either from the lives of the Calabrian monks, from the Arab chroniclers or from what the Normans found when they conquered the island. Of the many 1. Amari, vol. II., p. 402. 2. Amari, vol. I., p. 485. ARCHBISHOP NICODEMUS 29 monks who migrated from Sicily to Calabria apparently only one returned to Sicily as a bishop, and there is considerable doubt even about him.^ The gradual conquest of Sicily and the iconoclastic persecutions in the Levant caused a large imigration of monks into south west Italy with the result that much tradition and church learning were concentrated in Calabria during this period. With the interesting political consequences of this movement I am not now concerned. They have been described by both Professor Bury and M. Gay. The Christians formed the largest part of the population of Sicily in the tenth century, and were divided by the Saracens into four classes ; the independents, the tributaries, the vassals, and the slaves. The last two classes existed chiefly in western Sicily, called the Val di Mazzara, where the Saracens first landed. In the south and south-east, called the Val di Noto, the population was almost entirely Christian, in part tributory but chiefly vassal. The independent Christians occupied the mountains of north-east Sicily, called the Val Demone. Though nominally Byzantine subjects, they formed themselves into quasi- independent municipalities and leagues. The terms upon which these vassal Christians, called Dsimmi, were allowed to practice their religion and d« al with their property were apparently similar to those granted to the Christian subjects of the Porte by the Sultan Mahomet the Second, and still enjoyed by tiieir descendants to-day, if indeed such a term as enjoyment can be appropriately used. Among a number of rules relating to dress, personal conduct, iiames, and the use of seals, the Christians were forbidden to build new churches or monasteries though they were apparently permitted to restore those that fell into decay. When the Normans took Palermo they found a Christian com- munity presided over by one Nicodemus, who bore the title of archbishop, and officiated at a church called S. Ciriaco. He is, I believe, the only bishop in Sicily of whom there is any record during the Saracen dominion. How he came to be appointed and by whom he was consecrated is not known. Amari considers that he was probably elected by his flock, and not only recognised by the 1. Amari, vol. II, p. 402, 30 HIS THANKSGIVING SERVICE Emirs, but compelled by them to reside near their court at Palermo to watch over, and be responsible for, the welfare of his community. If that were so, he must have stood in much the same relation to the Emir's court as the patriarchs of the Greek and Armenian communities stand to the Sublime Porte to-day ; Amari also cites, in support of his view, the positions of the Jacobite patriarch of Alexandria and the Nestorian patriarch of Seleucia, who were compelled by the Arab Emirs at that time to reside respectively at the courts of Cairo and Bagdad.^ Nicodemus' name is recorded for us through his officiating at a thanksgiving service held by the Normans after the conquest in Palermo. The service, presumably in Greek, was held in the ancient basilica which had been converted into a mosque, the Mahometan fittings were removed, the Christian emblems reinstated and the church was reconsecrated and dedicated to the Blessed Virgin.^ But though the Normans availed themselves of Nicodemus on this occasion, they looked upon him, according to Amari, in much the same light as they did the imam of a mosque. I do not know what authority Amari has for saying this, and he does not cite the source of his information. Shortly afterwards the Normans appointed an archbishop of their own, disregarding not only Nicodemus but also Humbert of Loraine, who at that time held from bishop Leo IX. of Rome the titular rank of archbishop of Sicily. I shall have occasion later on to give a list of some of the bishops appointed by the conquerors, and it is interesting to note that however much the Normans may have been attached to the Roman see they almost invariably appointed their own countrymen to the Sicilian bishoprics ; one of the best known of them, Walter of the Mill, the builder of Palermo cathedral, was an Englishman. There is no record of the number or condition of the clergy in Sicily when the Normans arrived. Amari considers that by the middle of the eleventh century there were not more than about six monasteries left, among them being S. Maria of Vicari, S. Angelo of Lisico near Brolo, and S. Phillip of Demona. In one case the monks appear to have petitioned the Normans for a renewal of the J. Amari, vol. II., p. 402. g. Amari, vol. III., p. 131. char Let' unu ai lids froii .21 ( ruiscar* \i09^«\'\c A^ 'b"^Viora, Palermo. Plan by V. di Giovanni. To face page J4. vl .O^^iV:\K^\ .K?iK^K\ '\Q 3. .81 .1 M\' AWS \*t'<\ - 7 yi.fi\ s"»$\ vT^ 18. CASTLE OF FAVARA. PALERMO. Interior from the nave showing the chancel arch, the lantern, the main apse and one of the side apse niches. Another view from the floor of the lantern looking np into the dome shotving the squinches ; on the left the semi dome of the apse ; on the right the cross vault of the nave; above and below the barrel vaults of the transepts. To face page 3$, THE EREMITI 35 the latter consists of a square lantern supported on pointed arches. The dome of the usual Palermo variety standing above it rests on arched lights in the lantern and squinches in the four angles, and is also lighted by four round-headed windows above the squinches. On each side of the square chancel there are shallow transepts roofed with cross vaulting and the east wall has a shallow semi- circular apse recessed in the depth of the wall, lighted by a round headed window now blocked up. On either side of it two long narrow niches are cut in the walls of the transepts to represent the lateral apses used for the Greek rite as prothesis and diaconicon. The chancel outside is almost entirely hidden by outbuildings in the courtyard of the castle ; it shows no trace of the apse or niches. The dome has been much pulled about and at some time apparently converted into a look-out tower by the insertion near the top of large round headed stones projecting 25 centimeters and supporting a gallery or platform on the crown of the dome. These projecting stones seen from a distance look like the eaves of the familiar Byzantine pepperpot dome with a flat roof. But in point of fact the dome is of the usual Palermo variety, if anything rather higher than usual. The dome over the chapel at the Ziza, now buried under a modern campanile, is still visible from the street in front of the keep of the castle. The chapel was dedicated to the Holy Trinity. I have mentioned S. Giovanni dei Lepprosi and the Favara chapel together because the former is supposed to be the first church built by the Normans in Palermo, and the chancel of the latter to have been copied at a later date from the chancel of S. Giovanni. In studying this style of architecture much interest- ing material for comparison will be found in the church of the Eremiti, S. Orsola the church of the Vespers, and a little pavilion called the Cubola, in the garden of the Cuba. The church of the Eremiti, like the chapel at Mare dolce, is built into an earlier piece of Saracen work, probably part of a mosque. There is a tradition that the site was occupied by a monastery founded by Gregory the Great, but no portion of the present church or of the adjoining building dates from the sixth century, and an existing charter records that the church was founded by King Boger in 1148. This church and the adjoining Arab mosque, 36 THE CUBA AND CUBOLA now restored, are among the most interesting buildings in Palermo. Patricolo has written an article upon them in Archivio Storico Siciliano, accompanied by a plan showing the Arab construction and the twelfth century church with three apses built into it. The earlier church was no doubt destroyed by the Saracens. I have already spoken of the Cuba, the third palace in the suburbs of Palermo. The Arab inscription in it is rendered by Sladen^ in English thus : — " In the name of God, Clement Merciful pay attention, Here halt "and admire. You will see the illustrious dwelling of the most " illustrious of the kings of the earth, William II." A portion of the date also survives — " And of our Lord the Messiah, a thousand and a hundred, add three " to four score." On the opposite side of the road, in a garden once forming part of the palace grounds, stands the little pavilion known as the Cubola. The ornament over the arches should be compared with that over the windows in the east apse of S. Orsola also known as the church of the Vesperi built in 1170 by the founder of the cathedral, Archbishop Walter of the Mill. They are probably of the same date through the Cubola is usually represented to be the only purely Saracen piece of architecture in Palermo. It is beyond my purpose to describe the Capella Palatina or the Martorana and the little chapel of S. Cataldo next to it. The chapels of the Favara and Ziza, being attached to royal palaces, were suffragan to the Capella Palatina, and received a grant of tunny as well as an endowment from the royal revenues. The chapel at the Favara was entitled to a fish from the tunny fishery at Solunto, beyond Bagheria on the road to Messina, as the following passage in the Charter shows : — " Fro prime piscacione Ecclesia sanctorum Phillipi et Jacobi di " Fabaria piscem unum." There is another chapel in this part of Sicily in the Byzantine Arab-style like S. Cataldo and some of the other buildings I have 1. Sicily, p. 4.23. k^YiVi i )'mw '' ^mm^wm 19. CHURCH OF THE TRINITA DI DELIA CA STEL VE TRA NO. ■ West front. Plan by Patricolo. Phot : Incorpora., Palermo. To face page jy. CHAPEL AT DELIA 37 been describing, known as the Trinita di Delia about a mile from Castelvetrano. The earliest record of this church is in a royal letter dated from Palermo on the 17th of June, 1392, written to the Bishop of Mazzara in these words : — " Commendamns dilecto Capellae nostrae Bernardo Figuera " Ecclesiam Sanctae Trinitatis Ficani in territorio Castri Veterani " cum feudo eodem vocato Delia, et omnibus iuribus, et perti- " nentiis suis ecc. et quia Baro Castri veterani asserebat feudum " Deliae Sui fui.«se juris, idem Rex perscribit Episc. Mazarensi ex "htt. Oatanae 12 febr. 1 indict. 1392, ne permittat in possessione " molestari dictum Bene'ficiarium." According to Pirri this church was attached to a Benedictine monastery, and in 1474 Cardinal Giovanni Nicola degli Orsini, occupying the episcopal see of Teano and titular of S. Cecilia in Trastevere, wished the priorship of Delia annexed to his abbey of S. Giovanni degli Eremiti in Palermo. Fazello, on the other hand, considers that the church may have been attached to a priory of the Basilian order. The plan of the church is almost square and orientated. The interior is divided into nine compartments, forming a nave and aisles terminating in three semi-circular apses. The central compartment of the nave is covered by a dome, the four corner compartments are covered with cross vaults, the other four compartments forming the nave, chancel and transepts, have barrel vaults. The apses are covered with semi-circular semi- domes. The material used is a kind of calcareous tufa quarried in the neighbourhood, and the interior does not appear to have been decorated, for there is no trace of mosaics, paintings, or of a marble pavement or marbles on the walls. The dome is supported on pointed arches springing from four marble pillars. Two on the east side nearest the apse are made of cippolino, and the others of red granite. The bases and caps are of white marble cut in the usual Byzantine-Norman form. Spaces are also provided for six small decorative columns by cutting away the corners of the pilasters supporting the apse arches. These recesses are now unfortunately empty, and with these exceptions the interior appears to have been quite plain. 88 CHAPEL AT DELTA All the doors and windows are made in the blunt pointed style usually found in Sicilian buildings of this date. There are three windows on each side of the church, including one in each of the three apses. There are also some small lights in each side of the square drum supporting the dome. The same kind of windows will be found in the churches of S. Giovanni dei Lepprosi and S. Orsola in Palermo. The main door is in the west front and there are also two smaller doors one on each side in the north and south sides near the west-end. This arrangement of three doors is peculiar and similar to that in the little chapel at Maccari. It has been suggested by Patricolo that it was intended to provide separate access to the nave reserved for the women and to the two aisles reserved for the men, the respective portions of the church being divided off by wooden barriers. He argues that this arrangement points to the church being built to suit the Greek requirements. Patricolo notes that the elevation of the church outside differs from that of the Martorana and is more graceful. Instead of carrying the roof of the body of the church at one level, the cross vaulting of the four corner compartments of the building has been lowered so that the nave chancel and transepts stand out and a cruciform appearance obtained. The dome of the usual Palermo variety and the drum below it occupy the centre of the cross, and are raised above the rest of the building. The dome is supported inside on squinches. The plan, the elevation, and the general design of the Trinita di Delia are, of course, copied from an earlier building, but I found no early domed church or chapel in this part of Sicily which could have served as a model. On the other hand there are a number of chapels copied from the Trinita di Delia during the renaissance period, and they are extremely interesting owing to the adaptation of the new style of ornament to the Byzantine-Arab method of constructing a dome supported on squinches. I found the following five examples ; at S. Egidio and S. Catharina at Mazzara, in the abbey church on Monte San Giuliano (Mount Eryx), in the church 1. I have taken the plan of the church from the description of Patricolo, " La Chiesa della Trinita di Delia presso Castelvetrano. Monumento del " XII. Secolo scoverto di 31 marzo 1880. (Estratto dall' Archivio Storico " Siciliano N.S. Anno V. Fasc 1-11.) Palermo. 1880." 20. CHURCH OF THE TRINITA DI DELIA. East front. Phot: by In corpora, Palermo. To face pai^e j8. .(>£ lien aud the chnrch being built ( .... ■ .■ . " I .. , ■/ I \ \ \ ^ IS 0^ %-^fe\ *;>K\ 0\ 21. Church of S. Egtdio. The interior of the square chancel with octagonal superstructure and 1 6 sided dome above. The South-side of the exterior. I To face page jg. CHAPELS IN WEST SICILY 39 of the Addolorata, in the same town, and in the chapel of S. John, attached to the famous church of the Virgin of Trapani. All these chapels are inserted or added to older buildings. I do not recollect to have seen any example of this style in Palermo itself attributable to the period of the Spanish occupation, but there are several instances in Spain, and it may well be that they originated in this corner of Sicily. Where the space for the squinch was insufficient, as in the cathedral at Burgos, the ribs of the splayed scallop-shell were squeezed up to resemble those of a half-closed fan. The chapel in the abbey at Monte San Gialiano has an inscription recording that it was built in the sixteenth century, and I am inclined to think that all these Sicilian chapels with scallop squinches are of that period. The most elaborate of them are the chapels of S. Egidio at Mazzara and of S. John attached to the Norman church of the Virgin of Trapani in the suburb of that town. Norman portions of S. Egidio can still be seen in the nave. The chapel is in reality the chancel of the church, and consists of a square chamber covered by a dome resting on an octagon lantern. The accompanying sketch of the squinches and part of the vault will convey an idea of the general appearance and character of this as well as of these other domed chapels in the renaissance style. But they are really only interesting because they show that the Byzantine and Arab methods of construction were continued here as late as the sixteenth century. Before leaving this part of Sicily I should not forget to mention a few fragments of some early churches. On the promontory at Bonagia, obviously a corruption of Panagia, a small tunny fishing station on the north-east side of Mte. S. Giulano, there are foundations of an early chapel, but the remains are insufficient to give any idea of either tlie style of architecture or the date. On Mte. S. Giuliano, close to the site of the celebrated temple and behind the prison in the citadel, there are the walls of the nave and square apse of an early church. There are also a few minute pieces of coloured plaster, some fragments of pavement and the jamb and spring stones of round arches over two windows on each side of the apse. The church had a plain rectangular nave 16 met. long, 7 met. broad, and the apse is about 4 met. square. The former 46 CHABTEBS OF GIBGENTI has been completely destroyed, and nothing remains but fragments of the walls about 1"25 met. above the ground. The apse walls are still standing to a height of 4-50 met. The absence of any considerable remains of an early church at Mte. S. Giuliano was a great disappointment, more especially as portions of the city walls seem to have been built, or at least restored by the Byzantines. These walls are in a way an epitome of Sicilian history, and contain Phoenician, Byzantine, Saracen, Norman, and later work.^ The chapel of S. John on Cape Lilybeo, near Marsala, also seems to have traces of early work which may be Norman. But the most important remains there are the wall paintings in the catacombs supposed to be the work of early Christians, though the general style and the character of the lettering of the Saints' names, points to the Norman period. I am inclined to attribute them to the Greek monks who returned to Sicily after the Norman conquest rather than to the early Christians. The diocese of Girgenti was one of the most important in Norman times, and according to the charters granted by Count Boger and confirmed by his successors it comprised all the central part of the island. The first of these charters is a grant of tithe to the church of Girgenti by Count Boger dated in 1093. The second is an inspeximus charter dated 9th February, 1244. Both have been transcribed by my friend Professor Garufi from the originals in the muniment room of Girgenti cathedral, and contain much interesting information about this part of Sicily.^ The recitals in the first charter are couched in grandiloquent terms, the more interesting parts of it may be translated thus : — " I, Roger, by Divine Grace Count of Calabria, girt with the "sword of Grace from above, and adorned with the helmet and 1. Professor Salinas records the discovery of Phoenician letters on portions of the walls. I was fortunately able to find them from the information giren in his article, and they appear to me to be masons' marks. Archivio Storico Siciliano. N.S. Ann vii. Fasc I.-IV., p. 411, 1883. 1. " L' Archivio Capitolare di Girgenti 1 Documenti del tempo Normanno — "Suevo e il 'Cnrtnlarium' del sec. XIII. (Estratto dall' Arch : Stor : Sic : N.S. "Anno XXVIII.) C. A. Garufi. Palermo. 1903." CHARTEES OF GIRGENTI 41 " shield of good and laudable intention .... in the year 1093 "of the Incarnation of our Lord Jesus Christ, Urban the Second, " occupying the Apostolical See .... established the episcopal "sees of which one is the church at Girgenti, whose bishop is " called Gii^landus, and to whose jurisdiction I assign whatever is "contained in the underwritten boundaries with all the right of "tithe .... as well in the city of Girgenti as in the said , " diocese. That is to f;ay from the place where the river rises below " Corleone to above the rock of Zinneth and then stretches through " the boundaries of Jato and Cephala and then to the boundary of " Vicari and thence to the river Salso, which is the division of "Palermo and Termini, and from the mouth of this river where it " flows into the sea to the river Torto, and from there to the place " where it i-ises at Pirri above Petralia, and then to a high mountain " which is above Pirri, and thence to the river Salso where it joins " with the river of Petralia and along this river to the place where it " descends to Licata which divides Girgenti and Butera, and thence "along the sea shore to the river Belice, which is the boundary of " Mazzara, and then along the course of this river to below Corleone "where the boundary commences, excepting Vicari Corleone and "Termini. And I grant the ownership to the same Lord Bishop "Girlandus and other bishops after him of Casale Catta with 100 " villeins." So that in point of fact the diocese occupied nearly the whole of central Sicily with the exception of one or two towns. The inspeximus charter is a much longer document, and is entitled : — "Libellus de successione pontificum Agrigenti et de Institutione " Prebendarum et aliarum ecclesiarum dyocesis sicut ex relatione "cognovimus precendentium Seniorum et ipsi inspeximus in "eodem statu." The recitals and parcels or boundaries, substantially the same as in Count Roger's charter, are followed by particulars of the fourteen prebends and a number of benefices, and then by a list of prelates who occupied the see of Girgenti. The first of them, S. Gerlandus, held the see from 1093 to 1104. He was sent for by his kinsman, Count Roger, from Besan^on in Burgundy, and consecrated by Urban then Bishop of Rome. " Sanctus Gerlandus in sex annis hedificando complevit Episcopium "et Curiam prope Castellum propter timoi'em innumeralium Sarace- 42 CHARTERS OF GIRGENTI "norum habitanciurn in Agrigento quia pauci Xiistiaui ibi usque "ad mortem Regis Guillelmi secuncli." He was succeeded by Drogo, and Albert and Warm. The first two held the see for a year apiece, and the third succeeded in 1105. " Postquem f uit episcopus Gualterius francigena qui in episcopatu " residens de Saracenis multis valde verebatur .... mox " emptis multis bufalis fecit trahi lapides magriOS de civitate veteri " et tribus annis complevit edificium tuiris A piece of vandalism that called down upon his head the wrath of Amari, who observes " ch'ei non riposi in pace." ^ " Huic successit Gentilis tuscus qui fuit cancellariiis Regis " Ungarife .... Post hunc vero fuit Baitholomeus electus "dum asset Constantinoplim legatus et post tei'cium annum fuit "consecratus a germano suo Gualterio Panoimitano archiepiscopo - " .... postea sub Rege Tancredo Panormi archiepiscopus "fuit." He was succeeded by Ursus who had an eventful career, and among other experiences was taken prisoner by the Saracens. He was succeeded by Raymond of Acquaviva, Dean of Girgenti, in the reign of the Emperor Frederick. Nearly all these are good Norman names, and as I have said before, however much attached the conquerors may have been to the Roman see, they filled the bishoprics with their own kinsmen and countrymen. The charter mentions the Greek parish church in Girgenti, a Greek benifice in Caltabellotta, the Greek church at Castronuovo and the monastery of S. George at Trocculi (Triocala) below Calta- bellotta. The latter was evidently the cause of friction. In Girgenti itself there is a basilican church called S. Maria dei Greci, built on the plinth of a Greek temple. Portions of the walls may date from the Byzantine period, but they have been too much restored and pulled about in later times to be capable of certain identification. Another very early church dedicated to S. Biagio stands at the extreme east of the ancient city, also built on the site of a temple. The present building is Norman but may be a restoration of an 1. Amari, vol. III., p. 210, in a footnote. 2. Walter of the Mill (Offamilo), the builder of Palermo Cathedral. St .^ V ^';i,i^\ ^"ii^V ^'X. 22. HERACLEA MINOA. Bastion in the City Walls. Masons marks. The Imver courses of the wall. To face page 4J. THE SOUTH COAST 43 earlier Byzantine church. There are some very early crosses cut in the west and south walls. Another interesting building is the chapel in the Norman castle of Favara, not to be confused with the castle of the same name at Palermo. Favara is a large country town about nine miles north- east of Girgenti. The castle, built by the Chiaramonte family, is in a good state of preservation. The chapel is situated in the keep on the first floor. The nave, a small square chamber, is covered by a dome of high pitch in the usual Palermo style supported on pendentives and pointed arches. The apse, a small square chamber with lateral doors, is lighted by a pointed window flanked by three columns on each side, one of Egyptian porphyry and the others decorated with chevrons in coloured marble ; a photograph of the Norman door of the chapel will be found in the frontispiece and is one of the best of many doorways of the same kind found in different parts of the island. Perhaps the best of them is in the castle of Maniace in Syracuse. The sea coast as well as the country inland westward of Girgenti is very sparsely populated and there are scarcely any towns and no villages. In summer this district must be quite waterless and as hot as an oven. After leaving Porto Empedocle the high road passes through two small places called Siculiana and Montalegro, both dating from the Spanish occupation, and containing nothing of interest. At the latter I was told that a fine " cuba,' the name the peasants give to domed buildings, still stood on a hill at the mouth of the river Platani. This spot now identified as the site of the ancient Greek city of Heraclea Minoa is also known as Capo Bianco from the lime stone clifif of the headland, I visited it through the kindness of the Judge la Mantia and friends at Sciacca who took me in a steam tug. The Cuba proved to be the lower part of a fine circular bastion of the classical city wall, made of large and well preserved masonry with several masons' marks upon it. In conse- quence of the encroachment of the sea the bastion now stands at the edge of the cliff, and is in imminent danger of falling some two hundred feet on to the beach below. I have, therefore, reproduced a photograph of it as well as of the wall and the masons' marks so that some record of these interesting ruins may be preserved. I could find no trace of a church and the ancient city which 44 SCIACCA appears from some fragments I found to have survived till Roman days is now razed to the ground. Nearly all the site is arable land cultivated by peasants living at Heraclea Cattolica situated in the hills inland. Twelve miles west of Heraclea Minoa is Sciacca, the ancient Thermae Selinuntinae, and the hot springs dedicated to S. Calogero on a hill behind the town are still resorted to for their medicinal properties. In Sciacca itself there are no Christian remains of the pre-Saracen period, though the spot is par excellence the scene of the traditional labours of S. Calogero. A Norman church dedicated to S. Nicholas appears to be of very early date. It has a plain nave and aisles terminating in three semi-circular apses ; of these only the central apse is used for service, the south aisle and apse have been turned into a vestry and store-room, and the north aisle and apse incorporated into an adjoining house. The church possesses an ancient wall-painting of S. Nicholas blessing in the Latin way, and a large wooden cross with a painting of the crucifixion in the late Greek style, both of the post Saracen period. Other crosses of a similar kind will be found in the cathedral and the church of S. Catharina at Mazzara. The Sanctuary of S. Calogero has been entirely rebuilt, and nothing remains of the traditional cell. The present church is quite modern. Five miles behind Sciacca perched up among the pinnacles of a short but steep mountain range stands the Norman town of Caltabellotta, the successor of the ancient Triocala. I have observed that, during Norman times at any rate, this ancient diocese was merged in that of Girgenti, and presumably by that time the Christian community had entirely disappeared. At any rate there are no remains of the pre-Saracen period unless a few fragments in a little monastery below the town can be attributed to it. There are, however, a fine Norman abbey church and two small chapels. The former now in process of restoration, contains a holy water stoup of the same age as the older part of the church with two keys and the legend Tu es Petrus carved upon it in early gothic characters. The doorways of the chapels are very handsome. The view from Caltabellotta is exceedingly fine, and extends from the cathedral tower of Girgenti, visible through a gap in the coast range to the promontory near Mazzara. 23. TYNDARTS. The so-called Gvmnastum. ^- y^. Norili aisle on the left ; Siuth aisle on the right. To face page 44. AS A wifcli r , ! . ./ ^tNXXW^b ^te-'oV'^^T oY. ,VV ^^,»^ s-MsX/Jl J-^ ./\'A ,\\\\UV\\VW\'f >A^\ Cs " a =U .•?_V y>.ti\ i-is^ ^^ 24. im^li^-:- TYNDARIS. Apse of the Gymnasium. CEFALU. Ruin of the church built oz'er the ''prehistoric honse\- S.W. angle shoioing the W. front and the doors of the house ; and the S. side. lt."35 7"'S B I "70 To face page 45. CEFALU AND TUSA 45 The Christian remains at Selinunto I have already alluded to. I found some evidence of early Christian churches at Cefalu, Tyndaris, Alaesa (a little village near S. Stefano di Camastra, now known as Tusa), Termini and Carini. The first three were ancient episcopal sees. Upon the so-called temple of Diana, on the rock above Cefalu, are the remains of a small chapel consisting of a square nave, a short chancel and a semi-circular e-mt apse. The semi-dome vault of the apse inside was faced with square stones and springs from a plain bevelled cornice. Otherwise the masonry is of the roughest description, composed of rubble and pieces of brick and cement. The measurements of the nave are, length 7"57 met., breadth 7"15. The chancel 4*05 broad, the apse 3"20 broad, and 1*70 deep. The floor is filled with debris of stones and tiles, and there is not enough left to show where either doors or windows were placed or how the church was roofed. The materials are altogether insufficient to form any idea of the date of the church which has been attributed without justification to the eighth century. About two miles north-east of Tusa there is a farmhouse with a plain rectangular chapel adjoining it, erected in 1481, as a stone tablet in the west gable records. It is supposed to be on the site of the Koman and Byzantine city of Alaesa. The church has a piece of Roman column for a holy water stoup, and two Roman millstones are preserved in it, but with those exceptions there is nothing to connect it with the pre-Norman period. In a field close by, full of Roman tiles and pottery, there is a remarkable Roman sepulchral chamber made of square stones and vaulted in a jDeculiar manner. Similar work will be found in a monument in a corner of the cemetery at Termini. There are many Greek and Roman remains at Tyndaris and the most important of them is the Roman basilica called the gymna- sium. It is more than likely that this building was used as a church during the Byzantine occupation. The Sanctuary of the Madonna and the chapel in it are relatively modern. The ancient diocese of Tyndaris included the Lipari Islands, but I found no trace of an early church on any of them, and in Byzantine days the population was, I have no doubt, much smaller than it is now. There is some doubt whether Carini, a small country town 46 TEEMINI between Palermo and Castelvetrano was an ancient bishopric. The classical town stood on the seashore, and there is nothing of it left above ground. About half way between the sea shore and the railway station I found a Norman pointed arch in a wall, and a small chapel probably also Norman, but with no distinctive archi- tectural feature about it. The mediaeval town is situated about three miles inland on a hill crowned by a fine Norman castle. Termini has several remains of the classical period, at least one ancient church, some Byzantine fragments and inscriptions and a Greek triptych in the museum.^ The triptych is about 2 ft. square, and the principal figures in it are the Blessed Virgin and our Lord surrounded by Angels, Apostles and Bishops, and Doctors of the Church. On the outside of the right panel is a colossal St. Christopher crossing a stream. The Saint is represented with a lamb's head ; he is dressed as a warrior and holds a staff in one hand, and supports our Saviour on the other arm. Our Saviour is clothed with a tunic, His left hand rests on the head, and He is blessing with His right hand in the Greek way ; near the figures are the letters I C X C with abbreviation marks above. Above these figures is a half -figure of St. Nicholas. On the outside of the left panel is a half-figure of Anastasius of Alexandria, and below similar figures of S. Miletus of Antioch, of S. Theodore, and of another saint. In the left top corner there is another similar figure of S. Gregory. The central figures in the chief panel are the Blessed Virgin vested from head to foot in a red robe,^ Our Lord is seated on her arm, and is dressed in a long green tunic. His left hand holds a scroll and His right hand is raised in benediction in the Greek way. He has a cruciform nimbus with the letters o^n. By the Virgin's side are the letters MP ©Y with abbreviation marks above, and H OAHTHTPIA. On either side are two small robed angels with the letters m and r representing the Archangels Michael and Gabriel. On either side of the panel are two figures representing S. John the Baptist and S. John the Evangelist respectively. The former represented with a beard holds an open book in one hand and with 1. — See also Archivio Storico Sieiliano. N.S. Ann. vii. Fas. i.-iv., p. 142-146. Palermo, 1883. 2. Probably in allusion to Count Eoger's vision dm ing the siege of Palermo. 25. CEFAL U. Interior of the church on the ''Prehistoric house ' taken from the N. W. corner of the 7iave looking towards the chancel : the Nor mail citadel is in the background. To face page 46. .?s iho out i"e, auu .11' 0Y mih tseutinig the taiUt; of the pauel are i\ ■■ ^ ^". John theEr - •'onrd holds •^V yi»\ 'i'itA. ^^'X .rt£ .\Yv\U%*^\ ■ xV^'^n'^A^'^ 26. TERMINI. Byzantine triptych of the 12th Century in the Museum. To face page 4^. TEIPTYCH AT TERMINI 47 the other is blessmg in the Greek way. Below the central figures are three bishops S. John Chrysoslom, S. Gregory Nazianzene and S. Basil the Great. Each holds a book in one hand and the other hand is raised in blessing. On the inside panel of the right side are representations of S. Joseph and the Annunciation and below them a figure of S. George on horseback slaying a Dragon. On the corresponding panel on the left side are two half figures of S. Joachim and S. Ann who are being blessed in the Latin way by the Blessed Virgin risen from her seat. The Holy Ghost hovers over her head. Beneath this group is an equestrian figure of S. Demetrius slaying the Captain of the Saracens. This figure is introduced in allusion to the vision of the Saint by Count Eoger during the battle of Cerami in 1063, and from the introduction of it as well as from the general character of the painting, the triptych is probably to be attributed to the twelfth century. I translate the following remarks upon the style of painting from the Article by De Michele : — " The draughtsmanship of the figures is somewhat hard, but the " contours are boldly executed ; the transparent and brilliant " colouring seems to have been applied by a process of distemper, or " else by encaustic, a mode of painting in colours with a wax body, " which was employed by the ancient Greeks and Komans, and which " is believed to have continued to flourish in Sicily even after the " decadence of the aits, and to have lasted until the fifteenth century, " as is confirmed by the famous picture of the Triumph of Death, " painted in encaustic by Antonio Crescervaio, on a wall in the " vestibule of the ancient chief hospital of Palermo. " And lastly, it seems to me well to call attention to the fact that " all the figures and drapery were first chased on the gilt ground, " and the contours were afterAvards traced in black. The flesh was " rendered by a light brown pigment, and the contour traced in the "same colour, somewhat darker; but the eyebrows, eyelids and eyes " are in black. Transparent pigments are applied over textures on "the folds already drawn; and the half tones and flesh lights are " rendered with bold touches and a full body that contrast with and " stand out against a mass of brown colouring." The old chiesa matrice dedicated to S. Giacomo seems to be Norman work on the site of a much earlier building. I noticed 48 TERMINI that one of the steps leading up to it from the street originally- formed part of the architrave of a classical building, and has a triglyph carved on it. There are said to be other classical pieces built into the tower, but I could not find them. The Norman church was a basilica with three semi-circular apses at the east end, the body being divided into a nave of four bays by pillars supporting circular arches and clerestory windows above. A fifth bay was the chancel, and the arches of it were supported by pilasters as well as pillars. The roofs of nave and aisles were made of timber. I have not often seen the original plan of a church so completely altered. The sanctuary and a square apse have been inserted in the place originally occupied by the west end of the church, and the original main apse in the east end has been shorn right off, and the principal door built in its place. The aisle roofs have been raised above the clerestory, and the windows are now blocked up. In point of fact the church has been completely ruined both inside and outside. Another church of very peculiar design, and unfortunately gradually falling into ruins, is situated on the highest point of the town near the public garden. The present fabric, comparatively modern, seems to have been copied from an earlier church, and from the general appearance probably the model was a Byzantine- Arab building. I copied the following incription over the west door : — " A.M.D.G. Divique prcecursoris Dni Jesu Xti vetustissima " exuentes forma ecclesiam istam ab immemorabili conditam jam " jam coUabentem Kev^» Saco^ S. Th. Dr, D. Dom<^"^ occurso beneflis " et Cappel"** ac pii operis sci Mercurii in eadem erecti Dep*"^ "spectabilis D. Nicolaus Marsala U.I.D.D. Joseph Lo Faso ambo " quoq Depti protectoris prsestitiis ab eodem opere maxima ex parte " ex suppetiis ad recentioris formae refectionem ab anno Salutis "MDCCLXXVIIO ad annum usque MDCCLXXXUM impensius " seduloque incubuerunt.'' The nave of this church is an octagon supporting a dome of the same shape. It is enclosed in a square curtain wall so that from outside the church appears to be square with the octagon dome over it. The building is in a deplorable if not unsafe condition ; indeed the custodian only let me in after some demur. The walls inside VARIOUS FRAGMENTS 49 are of rubble and cement covered with whitewash cracked in several places from top to bottom. Beyond the inscription quoted above and a well, there is nothing of interest inside. From the terrace in front there is a magnificent panorama of the sea coast from Solunto to Cefalu. ^ Messina and Catania have so often suffered from earthquake and eruptions that the absence of early buildings is scarcely a matter for surprise. There are no Byzantine remains in the latter, but in the former there is, I believe, a fragment consisting of a doorway of the monastery of S. Salvatore dei Greci. I only know this from hearsay, for the site of this monastery is now occupied by the fort of S. Salvatore at the end of the mole in Messina harbour, and therefore not accessible to foreigners. A little domed church near Gesso, the first station out of Messina on the Palermo line, was destroyed when the railway was built. The site was pointed out to me about 100 yards south of the station, but I was unable to obtain either any satisfactory description or any plans or drawings of it. Murray says that it was dedicated to, and supposed to have been founded by S. Gregory the Great .^ The church of S. Pancras at Taormina occupies, according to tradition, the site of one of the earliest churches in Sicily. But the superstructure of the present building at any rate is comparatively modern. In a garden below S. Pancras, in a south-easterly- direction, there is a little rectangular building with a semi-circular apse, now used as a toolshed ; over a modern door there is a stone slab with the date 1876, but I am inclined to think that this is a little late Byzantine building, possibly a chapel, and the ancient house adjoining it may have been a monastery. 1» Uandbcok/or S. ttxly and Sicily, 9tli Edition, p* 361, published in 1890. 50 SARDINIA. The traveller who is not a sportsman or an antiquary will find little to compensate him for the fatigue and discomfort of a journey into Sardinia ; the means of access leave much to be desired, the steamers are small and have little to recommend them except their cleanliness, and the passages are long and usually very rough. The scenery is not as fine or as accessible as in Corsica, and the hotels in Cagliari and Sassari, conducted entirely as commercial houses, lack comfort though they are tolerably clean and the food is fairly good. The inns in the small towns and villages offer the kind of accommodation frequently advertised on the walls of Sicilian taverns, Si loca, si vende, si fa da 'manchiare,' and the appropriate word for the faring is camjMve. On the other hand the climate in spring is pleasant, the people with their varied and picturesque costumes are attractive, and for the sportsman and antiquary Sardinia offers a first-rate field. The latter will find the island full of antiquities of every age commencing with the pre-historic dolmens and the nurhagi. The churches of Sardinia stand in much the same relation to the Pisans as those of Sicily do to the Normans. Not only does Sardinia owe her best churches to Pisan architects, but they seem to have destroyed the earlier buildings as the Normans did in Sicily. The pre-Pisan churches fall into two groups; the early basilican and the byzantine, and very few of either remain. The cathedral of S. Gavino at Porto Torres, partly rebuilt by Pisan architects in the beginning of the eleventh century, is by far the most interesting and important in the first group. Only four byzantine churches substantially in their original condition have been found up to the present, the largest and most important being the church of S. Saturnino at Cagliari. The other churches built while the Byzantine government actually or nominally ruled Sardinia are S. Giovanni in Sinis near Ik -in \ r SARDINIA. To face page c;o. tlio sm? \yf\,\^7\\:^. luch the Vorto Vecchio BorUfacix) St ret It of "^ Bortifctc/:^ S^ TfteresoA Asirvccrcv/(^ TslandSS SARDINIA 51 Oristano, S. Giovanni d'Assemini, a village near Cagliari, and the chapel of S. Savina at Silanus near Macomer, in the centre of the island. The first is a building of considerable antiquity. Some record of S. Saturnino and S. Giovanni d'Assemini has come down to us, the former in connection with S. Fulgentius,^ bishop of Ruspe in Africa during the Vandal dominion, and the latter in a charter preserved in Genoa dated in the commencement of the twelfth century.'- During this long period Sardinia was governed first by the Vandals, later as a province of the Byzantine Empire, and still later by quasi independent princes or counts owning allegiance to the Emperor at Constantinople. After the fall of Rome Sardinia was ruled by the Vandals and followed the fortunes of their African kingdom. The life of S. Fulgentius shows that in one respect at least they imitated their Roman predecessors, and made it a place of exile for the orthodox bishops of the African Church. History does not record whether the Vandal sovereigns sought to impose the Arian faith upon their Sardinian subjects, but there is evidence that Trasamond treated Sardinia in much the same tolerant way that Calabria was treated by the iconoclast Emperors in the eighth century,^ for S. Fulgentius w^as allowed during his exile to build a monastery near the ancient basilica of S. Saturnino at Cagliari. After the fall of the Vandal kingdom, and the conquest of Africa and Sicily by Belisarius, Sardinia became a province of the Byzantine Empire, and was administered as part of the African exarchate. After the Saracen conquest cf Africa the Byzantine possessions in the Western Mediterranean were reduced to Sardinia, the Balearic Islands and a few sea ports in Spain and Africa. The exarchate of Africa, however, continued to appear in the imperial rescripts, and Sardinia was still nominally adminis- tered as part of it.* But by the tenth century, though all connection with Constantinople was not yet severed, the Byzantine 1. S. Fulgentius, 468-533, was made bishop of Ruspe in 508. He was exiled to Sardinia and, with other exiled African bishops, he was summoned by Trasamond to Carthage to dispute with the Arian bishops. After the conference he was again exiled to Sardinia where he remained till 523. 2. Scano, Storia deli' Atte in Sardegna, p. 30. 3. Later Roman Empire, vol. II., p. 449. 4. Diehl, L'Afrique Byzantine, pp. 110 and 58G. 52 SARDINIA governors acquired independent position?, and as time went on their offices became hereditary and developed into the judgeships of the free States of Cagliari, Torres, Gallura and Arborea.^ It is easy enough to fix the commencement of the Byzantine dominion in Sardinia with Belisarius' conquest of the Vandals in Africa in the reign of Justinian, but difficult to select a precise date or event when it may be said to have concluded. On the whole, perhaps, the most convenient and the least open to objection for the present purpose is 1073, when Gregory VII., of Rome, claimed the allegiance of the Sardinian judges and threatened them with a Norman invasion if they refused it.^ At that moment there was no question of the Byzantine suzerainty which, in fact, must have come to an end many years earlier. From the gradual emancipation of the judges we may conclude that church and state in the island were left in a great measure to take care of themselves.^ From the tenour of the occasional relations with the Roman see it would seem that the sympathies of both were with the East rather than with the West. At one moment the differences between the Sardinian Church and Rome were upon matters of faith and not merely upon questions of church government or ecclesiastical supremacy. Two Sardinian bishops attended the Lateran Council in Rome'* and sided with the Latin Church in condemning the " Type," while not long afterwards the Pontiff refustd to recognise the consecration of some other Sardinian bishops by the Metropolitan of Cagliari because they all sympathised with the Emperor.^ The Sardinian Church was apparently divided into camps and the majority, if not in numbers at least in importance, sided with the Emperor. When a different view of the theological controversy of the day was taken 1. Besta, La Sardegna Medioevale, vol. I., pp. 40 and 47. 2. Besta, vol. I., pp. 79 and 80. 3. Gay, L'ltalie Meridionale, p. 209. 4. Summoned by Martin to consider the " Type " of the Emperor Constans II., Besta, vol. I., p. 26. They were Deodatus of Cagliari and Valentine. During the Byzantine occupation the seven original bishoprics had dwindled down to four : at Cagliari, Torres, Sulcis, and Pausania. Under the present arrangement there are three archbishops : at Cagliari, with suffragans at Galtelli-Nuoro, Iglesias, and Ogliastra; at Oristano, with a suffragan at Ales and Terralba ; and at Sassari, with suffi-agans at Alghero, Ampurias and Tempio, Bisarchio, and Bosa. 5. Besta, vol. I., p. 27, and Diehi, p. 537, on the relation between the Byzantine Governors and Eome ; p. 509, upon the state of the Church. SARDINIA 53 by Constans' successor and son Constantine Pogonatus and the Type was revoked, the Sardinian clergy again sided with the new Emperor and changed their views. The MetropoHtan and the Roman see were reconciled and the consecration of these particular bishops, to whom exception had been taken, was recognised and confirmed.^ As politics obviously played a considerable part on these various occasions, so politics may account for the relations between the Roman see and Sardinia being suspended for about 150 years. There is no record to show what the State did with the Roman patrimony or what side the Church took in the controversy about the images. But during this period the patrimony in Corsica remains, while that in Sardinia disappears from the Roman annals.^ The successful Saracen invasion and ultimate conquest of Sicily affected Sardinia in isolating it more completely from the central government at Constantinople. It is strange but apparently the fact that Sardinia, though frequently threatened and occasionally invaded, was never actually occupied by the Saracens as Sicily was.^ But the Saracen success in Sicily and South Italy even close up to Rome itself had another and indirect effect upon the Church in Sardinia. It brought about an alliance between the Greek Empire and the Roman see under John VIII., and from this alliance ensued among other things a renewal of the relations between the Sardinian princes and the Pontiff. In 873 John YIII. had to recall the princes of Sardinia to Orthodox doctrines because they yielded too easily to the seduction of heresy flowing from Greek learning or from the influence of Greek merchants who carried on the slave trade in the ports of the island."* To appreciate all these words imply a study should be made of John's pontificate, of his relations with the Emperor at Constantinople on the one hand and the states of Amalfi and Naples on the other, and of his endeavour to stem the tide of Saracen advance in Italy. Were 1. Besla, vol. I., p. 27, and Diehl, p. 537, on the relation between the Byzantine Governors and Eome; p. 509, upon the state of the Church. 2. Besta, vol. I., pp. 41 and 42. 3. Amari, vol. III., pp. 5-8. There were several invasions. The more important were apparently in 710, 752, 813, 817, 935, and in 1016 under Musetto. 4. I have translated this passage from Besta, vol "^ p. 43. 64 SAEDINIA the trading communities in Sardinia taking the same attitude towards the Saracens as Amalfi and Naples ; who were the slaves, what was the course of trade, and was Sinis one of the ports used.^ One of the important consequences of the alliance between Eome and the Byzantines at the close of the ninth century w^as the presence of a Greek fleet in the Tyrrhenian Sea, and this fleet may have contributed in some degree to save Sardinia from a permanent Saracen occupation.- The Arab ships are recorded to have been raiding Corsica and Sardinia^ at this time, and I do not suppose that at a moment when the Saracens were masters of practically all the shores of the western Mediterranean, in Africa, Spain and southern France, there was much scope for church building in an exposed and isolated sea port like Sinis. I have gone in some detail into the general history of Sardinia at this moment to see if any assistance could be obtained from it in fixing the date of S. Giovanni in Sinis, usually attributed to the 9th century. I am afraid there is hardly enough even to justify my conjecture that if the church was not built before the end of the 9th century when the island was virtually abandoned by the Empire* it was not built till the beginning of the eleventh century after the defeat of Musetto (Moghedid) in 1016, when the Saracen raids against Sardinia ceased altogether. The general style and appearance of the church justify a preference for the earlier date. The inscriptions in the churches at Assemini and the neighbour- hood have been assigned to the time of the Byzantine revival in the reign of the Emperor Nicephorus Phocas. Upon the whole I think the church of S. Giovanni d'Assemini is not older than the beginning of the eleventh century, and that the same might be said of the existing portions of the church of S. Saturnino. The general character of the buildings and some architectural details in S. Saturnino certainly support this view and that is all I think that can be safely said about them. 1. Gay, p. 129. Also Amari, vol. I., cap. XI., and especially pp. 442-443. Upon this period see the learned and interesting volume I. of L' Europe et le Saint Siege, by A. Lapotre. Paris. 1895. 2. Gay, pp. 162-163. 3. Amari III., p. 5. 4. Id., p. 12. SAKDINIA 55 I have come to the conclusion that these churches at Assemini, CagHari and Sinis are not later than the third quarter of the llt4:i century for another reason. 1 have alluded to the relations between Rome and Sardinia in Constans' reign and during the pontificate of John VIII. The differences between the Greek and and Latin Churches were of long standing in 1073 when Gregory VII. wrote to the Judges of the Sardinian States calling upon them to act as loyal sons towards the Latin Church and to renew that devoted attachment which already existed between the Eoman Curia and the Sardinian race (gente Sarde).^ Those who would not were threatened with licentia invadendi or permission of occupation to the Normans or others of the many applicants who sought permission to conquer the island. The Judges eventually gave way and in the end the Archbishop of Pisa was appointed the Pontiff's legate. The allegiance of the Sardinian princes was claimed primarily on spiritual grounds, and Gregory laments that the worship of Christ was sunk in the greatest neglect, and that the Sardinians (Sardi) more than any other people kept themselves estranged from the Catholic communion.^ The concluding sentence is obviously directed against the Greeks. Besta shows that up to the year 1000 Sardinia was looked upon by the Latins as a den (covo) of heretics.- After that date it is represented as full of religious fervour indicating the separation of the people from the heterodoxy derived from the Greeks and the triumph of Catholicity among them. Gregory's words, like his predecessor's, seem to be addressed to the princes, probably of Greek descent, who still clung to the Greek rite. As I have said they eventually gave way and with the appointment of the Legate the union with the Pioman church was completed. This change synchronises with the introduction of Pisan architecture into Sardinia, and it seems to me unlikely that an important building like S. Saturnino, or indeed any of the Greek buildings I am going to describe, would have been erected after the change had taken place. S. Saturnino however seems to have been in part restored by Pisan architects. 1. Besta, vol. I., pp. 79 and 80. I gather that a distinction is intended to be drawn between the faith of the people and that of the Princes. 2. Besta, vol. 1., p. 74. 56 SAEDINIA The reader will appreciate that my reasons for this conclusion are based upon conjecture which I can only justify on the ground that this early history of Sardinia is and must remain unknown through the absence of all contemporary records about it. The journey from Aranci Bay to Cagliari takes twelve hours for the trains go very slowly. Passing some pretty coast scenery and the sea port of Terranuova, the classical Pausania, with a Pisan church dedicated to S. Simplicio, the railway ascends gradually between two mountain ranges through broken and thinly populated country covered with scrub and " maquis " to the plateau of Chilivani. This piece of the line forms the southern boundary of the ancient state and and modern province of Gallura. This country of some geological interest and different from the rest of Sardinia, includes the Limbara mountain range and the coast in the straits of Bonifacio. The marble from the Limbara was exported by the Romans to Italy and the granite pillars of the cathedral at Pisa were quarried at S. Theresa di Gallura opposite Bonifacio. Though the flora of the district is Corsican rather than Sardinian the climate must be milder for near Terranuova and Aranci Bay I noticed anemones, the pink pyramid orchis and asphodel all in bloom, while the first two were not above ground and the last had scarcely budded even in sheltered spots of Corsica. A magnificent view of the jagged finger-like peaks of the Limbara and of the straits themselves can be had from the Ospedale pass above Porto Vecchio in Corsica, or from the deck of a steamer passing through the straits. At Chilivani the branch to Portotorres leaves the main line and crossing marshes and flat land and the watershed between the Tyrrhenian and Mediterranean, passes first through peculiar limestone hills with flat tops, and then along a river bed through a series of ravines and gorges down to Sassari. Many of these limestone hills have a cap of hard rock overhanging the escarp- ment of soft stone worn away by the action of water. The general appearance of this district will be familiar to geologists and at no .T£ r-\'A-A •'Nl'^'i^^^'A^'^ 27. V.J PORTO TORRES. Abbey church of S. Gaimio. South side of the nave. T'o face page sy. POETOTOERES 57 very remote period the country must have been raised into a series of plateaux. In process of time these were worn away by the weather into deep gorges and canons, leaving peculiar table- top hills and mountains scattered about the landscape. Sassari is the see of an archbishop, possesses a University and is a flourishing provincial town with little or no architectual interest. At the neighbouring villages of Osilo, Sennori and Sorso the peasants wear the handsomest costumes in the island. They are very pretty, brilliant in colour and decorated with peculiar silver buttons similar to those worn by the Bosnian peasants. From Sassari the rauway gradually descends through open country to Portotorres. This little seaport, once an important Eoman station, is now a poor straggling village. It still does a considerable trade in horses, cattle and produce with Corsica and the mainland. There are two weekly steamers to Leghorn, one of them touching Bastia, and a fortnightly steamer to Ajaccio. The latter, a small French boat, does the passage in six hours affording, on the whole, the shortest and most agreeable means of access to Sardinia.^ The cathedral and the Roman bridge are the two lions of Portotorres. The former dedicated to S. Gavino is one of the most interesting churches in Sardinia. It dates from the beginning of the eleventh century and contains a number of fragments including tombs, pillars, and capitals from classical and byzantine buildings. The alterations made during the Spanish dominion are unimportant, and the additions, as well as some adjoining outhouses, have been removed since I first visited the church eleven years ago. The more prominent of these additions consisted of a rough parapet pierced with loop holes built over the whole length of the aisle walls on both sides of the church, showing that it had been fortified. These parapets were a great eyesore and concealed the proportion and details of the nave. The general appearance of S. Gavino and the architectural details leave no real room for doubt as to its age, though local antiquaries have tried to make out from documents of a relatively late date that it was built in the sixth century, but, in fact, three 1. As an acknowledgment of much kindness shewn to me I recommend the little inn at Portotorres known as the Tre Amici. 58 S. GAVINO, PORTOTORRES ? capitals with good Byzantine carving and a stone carved with a cross afford the only evidence in the present church of an earlier christian building. The antiquaries have no doubt confused the church meant in the documents with S. Gavino. The ground plan of S. Gavino is that of a basilica divided into nave and aisles, the former terminating in a semi-circular apse at each end. This arrangement was common in a Roman basilica, but is very rare in a Christian church.^ The exterior is very simple. The roof is quite straight and unbroken, and terminates in a gable at each end. The walls of both the nave and aisles as well as the apses and gables and the square ends of the aisles are quite plain and relieved only by the usual Pisan round-head arcade and flat pilaster. The windows too, of the simplest kind, are merely long narrow slits with round heads. A shallow cornice under the eaves of both the nave and aisles roofs is supported by a single arcade of plain small round -head arches resting alternately on bevelled corbels and flat square pilasters. The windows I have just spoken of are pierced in both nave and aisle walls, the former as clerestory, under every alternate corbel : the church therefore was poorly lighted. The cornice under the eaves of the nave roof is returned across the east and west fronts at the same level and immediately above the apex of the apse roof. The gable above is decorated with the same arches supporting the eaves and resting on pilasters of varying length reaching to the cornice. The whole building is raised on a plinth of three steps after the fashion of a classical temple. The plan of the church made by Dr. Scano shows three doors on each side. The only original one left on the north side towards the west end is of the usual Pisan type with square jambs and lintel ; the latter supported by two carved elbows has a tympanum with an incised ornament, and a round arch above. There is a typical Pisan beast over the side of the door. Another door on the north side is a fourteenth century alteration with an escutcheon as finial and a demicouped angel supporting the arch on each side 1. Among other examples are tlie cathedral at Mainz and the ruined basilicas at Sbei'tla in Tunis and Matifou in Algiers. 28. PORTO TORRES. Church of S. Gainno. North side of the nave. To face page $8. .«£ v?^ *?ift\y4v\«i\ .PS <^l5i^»\«»X'>^ 29. PORTO TORRES. Abbey church of S. Gavtno. Interior of the nave looking East. To face page 5g S. GAYINO, PORTOTOREES 69 of the door. These angels hold an escutcheon with the arms of Portotorres, a battlemented square tower. The main door on the south side also dates from the fourteenth century. Similar angels support the arch and a crucifix as finial which fortunately escaped destruction when the parapet was added, for I have an old photograph showing that it was partly embedded in the masonry. Between the door and the eastern apse there is a block of white marble with a cross carved on it in relief, a piece of byzantine work belonging, I have no doubt, to the earlier church. The nave inside is covered by a roof of plain wood rafters, and the walls with the clerestory are supported on each side of the nave by three pilasters and eleven pillars. The latter with capitals and abaci above support plain round arches. These pilasters divide the nave into three large and one small bay, the first two bays from the east end having four pillars in each bay, the third bay three pillars on each side and the last a short bay of one arch between the pilaster and the west end of the church. Taking these bays in order from the east end I noted the following pillars and their capitals. In the first bay on the north side : first pillar, fluted marble shaft with small corinthian cap, both from a classical building ; second pillai*, plain granite shaft, ionic cap ; third and fourth pillars, same shafts as the last, but with corinthian caps. On the south side of this bay all the pillars have corinthian caps ; the shafts of the first and the last two are plain, while the second is fluted and made of marble. In the second bay all the pillars on both sides have plain shafts and corinthian caps. The third bay has three pillars on each side ; on the south side the first pillar has an ionic cap, the second a byzantine, and the third a corinthian ; on the north side the first and third pillars have byzantine caps, and the middle one a plain cap. All the shafts in this bay are plain and made of granite. The byzantine caps are made of marble and carved with the same design of vase or chalice, doves and acanthus foliage. -P appears on each face, and in one instance this monogram is reversed : this is not an unusual occurence. The device is, of course, a great deal older than the fabric of the present cathedral, and if these are not genuine sixth century capitals belonging to the earlier church of which there is now no trace, they are very 60 S. GAVINO, POKTOTOKRES skilful Pisan imitations. But I think that they belonged to the older church, built probably during Justinian's reign, when an attempt was made to revive the old importance of Portotorres as an outpost of the Roman empire. By a clumsy alteration of late date the high altar, sanctuary, and choir erected on a kind of tribune occupying the third bay, entirely spoil the effect cf the interior viewed from the east end. A crypt, like a long tunnel with barrel vault, runs the whole length of the nave and terminates in a rectangular chapel con- taining the shrine of S. Gavino. It is of no interest and merely contains a few fragments of Roman tombs and a classical has relief of some importance. The plans and arrangement of the two apses are similar. They are separated from the nave by a stone screen supported by a central pointed and two smaller round headed arches on either side resting on pillars with circular shafts and caps. The arrangement is most peculiar and I imagine quite unique ; but these pillars and screens did not form part of the original design and date from the fourteenth century. Since I visited the church ten years ago, these screens have been lowered and the top of the eastern screen has been made into an organ loft. The caps of the pillars in this screen are decorated with a plain running pattern of roses and interlaced basket work. The roof of the apse is groined and has a boss carved with a representation of S. Gavino on horse back. The apse itself is used a kind of store room and a staircase leads up to the organ. The western apse is practically similar but the caps of the pillars have two carved figures bearing the arms of Portotorres and some grotesque men and beasts. The aisles, very narrow in proportion to the nave, as usual in buildings of this style, are covered with cross vaulting resting on the abaci of the pillars and on plain corbels in the walls. The whole length of the nave including the apse is 43"50 met. and the total width including nave and aisles is 17'72. From Chilivani the main line to Cagliari winds up the escarp- ment of the Macomer plateau in long curves and after passing through a forest of small oaks and scrub, across stony country not unlike the County Kerry, reaches Macomer. The village stands in a picturesque spot on the verge of the southern escarpment over- 30. PORTO TORRES. Church of S. Gavino. The apse at the East cud and gable ; a stone with cross on the S. Side of the naj'c^ outside ; E. F. A byzantine Capital in the chancel. .WTS2!*.>T«^/'- To face page 60. A>-i«^44rjOi\V!Wj««r\»ks^jtfl5ti 'i«itVtt«-sVAi V \Y4Sj\^>'^ ^•■I If. ,\«>^ftv\ , v\->\oV. ,^i>f"\5>;"i\ov ■, .\?i ^^tj\ v»i»\ ^'^^ 31. STA SARBANA. St'lanus. West front; South side shorving the broken roof of the South chapel; the central apse is on the right. Elevation East side. Metres Plan; (Dr. Ashhy.J To face page 6i. S. SAEBANA, SILANUS 61 looking the great plain of Oristano. Branch lines run from here westward to the sea port of Borso and eastward through some of the finest scenerj' in Sardinia into the mountains of Nuoro. Near Macomer there are many nurhagi close to the railway. The village of Silanus is the second station on the line from Macomer to Nuoro. In a field about half a mile south of the station and some thirty yards from a conspicuous nurhag stands the little Byzantine chapel known as S. Sarbana, probably a local corruption of S. Sabina. This little building is not only a real archaeological curiosity, but in one respect unique for the dome inside is clearly shaped after the fashion of the neighbouring nurhag. The chapel has a circular nave with a porch and door at the west side and a semi-circular apse at the east side ; it was flanked on the north and south by two rectangular aisles ending in apses. Only the north aisle remains. Though very simple and lacking in any ornament the general outside effect of this little chapel is very pleasing and the proportion as good as the plan is novel. The nave is a plain drum of well cut blocks of limestone without decoration of any kind beyond two bands of black basalt running round the upper half of it. The roof is a flat cone covered with tiles and the eaves are supported by a bevelled band of stones projecting beyond the drum of the nave. This same bevelling appears under all the eaves throughout the building. The north aisle is a plain rectangular room with a door in the west front, now blocked up, a window in the middle of the north side, and a semi- circular east apse, also with a window in it. The door, placed a little on one side and not in the centre of the west end of this aisle, has two large monolith jambs and a round head arch with a narrow cornice arch over. The same arrangement of arching will be seen in the porch. All the windows have round heads and are built in the same way with a third inner course bevelled to the aperture. The porch projects from the west side of the nave on a plinth or step which apparently ran the whole width of the chapel. The jambs of the porch are made of large blocks of basalt and the upper fronts and the arches are of limestone. Ascending three steps we come to a plain square door and through it into the nave. G2 S. SAEBANA, SILANUS The interior is a circular chamber six metres in diameter covered with a conical or beehive vault made of well dressed stones and unbroken by cornice or ornament of any kind. The arrangement inside is very simple, and in fact a reproduction of the chambers of the nurhag. The altar raised two steps above the floor of the nave stands in the apse and the front or chancel step is carried right across the nave covering about a third of it. In the plan I have shewn the portions of the nave where a stone seat runs at the foot of the wall. The interior, now quite dark, must have been poorly lighted even before the apse window was blocked. I found it impossible to take a photograph inside ; but one of the ruined S. aisle will show a piece of the roof and the aisle arch connecting the nave and aisle. Both aisles had waggon vaults of good masonry. In the apse of the N. aisle there is the original stone slab on a block of masonry which served as the prothesis. I failed to find any dedication crosses, marks or inscriptions on it or in any part of the building. Nothing is at present known of the history of this little chapel but it is certainly a byzantine building and, from the general appearance, I should judge it to be of the tenth century or later. It has evidently been much restored but nothing short of wanton destruction could have ruined the massive masonry of the south aisle. Leaving Macomer in a southerly direction the line gradually descends the escarpment into the plain ; then crossing some low lying and uninteresting country cut up with Ficus Indica hedges and dotted about here and there with nurhagi and a few poor looking villages the train reaches Oristano. Oristano is a dull cathedral town the see of an archbisop with no buildings of particular interest. The cathedral is a large ugly romanesque building of the type usual in the eighteenth century. A small piece of arcading in the wall of the choir aisle shows that the earlier building was Pisan, and some interesting byzantine has reliefs have been found showing that there was a still earlier church. The detached campanile is an octagonal structure covered with a pomme and cone roof, and under the eaves are enormous grotesque heads of Titans made of stone or brick and painted red. 32. S/JV/S. Church of S. Gtoi'anfit. W. and S. sides. E.F. E. To face page 62. u. L_ n Inch'cft ki( A.I ASO j.'i S5Ufi\^"ift\«^'C 33. S/JV/S. Church of S. Gtova7im. East end and apse. To face page 6j. ORISTANO 63 This feature is quite novel to me but in shape and window ornament the campanile resembles that of Lerida cathedral and the Miguelfcte at Valencia in Spain. A handsome Pisan church dedicated to S. Justa in the south suburb, and a bastion and one of the old town gates are the only other ancient buildings Oristano possesses. In a square in front of the Court house stands a fine modern statue of Eleanora the 14th century soldier lawgiver and ruler of this part of Sardinia while it was administered by the Consuls or Judices of the republic of Arborea. Sinis occupies the site of the ancient Tharros a Phcenician or Carthagenian seaport mentioned by Ptolemy which still existed in Roman times and was also an early Christian episcopal see. Nothing is known of the date or circumstances of its destruction. The town was situated on a sandy isthmus connecting a low range of coast hills with a bluff forming the northern promontory of the lagoon of Oristano. The only way to visit it is by a carriage drive of fourteen miles from Oristano through the village of Cabras and across the marshes on the north shore of Oristano Bay. Sinis is frequently inaccessible in winter owing to floods, and in summer the whole district is full of malaria. The latter no doubt led to the abandon- ment of Tharros and the removal of the population to Oristano, though in point of healthiness the new town is not much better. The natives of Oristano call Cabras the tomb of the foreigner, and La Marmora says that even Sardinians cannot live there in summer and autumn unless they are born in the district. Beside the church of S. Giovanni there are some early rock-cut tombs, a few huts and sheds for cattle, goats and sheep. We passed several large flocks tended by shepherds wearing the picturesque but sombre national costume of Cabras. The shepherds were fine strong looking men and belied the reputation of the district for unhealthiness. I found their dialect very difficult to understand, and some of them looked more like Spaniards than Italians. The district of Tharros, now most forlorn and woebegone, reminded me, as we drove across the marsh, of the Turk's remark about the roads of Asia Minor, "In summer all the country is a road, and who wants to travel in winter? " The roaii came to an 64 S. GIOVANNI IN SINIS abrupt end just beyond Cabras, and for the remaining eight miles my driver had to pick his way across the marsh through pools and bogs as best he could. We eventually reached our destination and found the church of S. Giovanni built of a coarse grained stone quarried in the neigh- bouring promontory. Though there is no frost here and the site chosen is sheltered from the disintegrating effects of the sirocco, the stone has perished a great deal and the church owes its preservation to the great thickness of the walls, and to the solid roof and dome. Excluding the central apse at the east end it is 19 met. long by 17^ broad, the ground plan being consequently almost square. The vaulting, however, is arranged so as to divide the interior into a nave and aisles of three bays, a domed intersection with a shallow chancel, a single apse on the east and rudimentary transepts on the north and south. Taking the outside of the church first. The west end has a plain unbroken surface across the width of the building. The only ornament on it is the stone cornice or eaves, following the contour of the nave and aisle vaults, projecting a few inches and giving a simple finish to the facade.^ There is a plain square headed door with massive jambs and lintel and a little octagonal window above it, but no discharging arch. The surface of the lintel is slightly chiselled away at the sides and top so as to leave a projecting surface to correspond with the width of the jambs. Apart from the jambs of the chancel and transept windows, this is the only attempt at decoration I could find anywhere on the building.^ The north and south sides of the nave are propped up by massive buttresses with sloping tops. The arch of a door now blocked up occupied the western bay between the first and second buttresses on the south side. There is a small door still used between the third buttress and the transept on the north side. The rest of the north side is hidden by outbuildings. The faces of the chancel and the transepts outside are carried up to the roof and finished off to follow the contour of the vault in 1. Compare the Cathedral at Sebenico in Dalmatia. 2. Compare this decoration with that on the door of the so-called prehietoric house at Cefalu in Sicily. 34. 5. GIOVANNI IN SINIS. Arches of the nave and one of the vents on the roof. To face page 64. M. so lUI to dv yyrx?. Y.v-\Y»?nHJi.o\5^ i?.- vidth of tht coration SS'iSS^-^Z* p .Bfc .^(s v^\\ v.^v^\ oT 35. S. GIOVANNI IN SIN IS. North side showing the vents in the roof and part of the north transept. To face page 65. S. GIOVANNI IN SINIS 65 the same way as the west front of the nave. Each face is pierced by a double round headed light with massive jambs at the sides and a monolith partition in the centre. The latter appears to have been ornamented on the outside with a little half-detached pillar, but the stone is too worn to make the detail of the caps and bases recognisable. The eastern apse built of square blocks is semi- circular and roofed with a semi dome covered with cement ; it occupies the whole width of the chancel end and reaches up to the cill of the window above. There are no side apses, and the east walls of the transepts where apses might have been are pierced by two round headed windows near the ground. They are now blocked up. Two square holes in the roof of the nave on the north side will be seen in the photograph. We had to climb on to the roof of the little wagonette and then scramble up a buttress on to the aisle vault to inspect them. The driver is preparing for this manoeuvre, and he described it later in the day to his friends as " bisognava arrampicarse come simie." There were three holes originally, of which one is blocked up. They seem to have been windows or vents for the nave and were cut slantwise through the vault. The construction is most peculiar and quite novel to me. I was unable to make a satisfactory examination of them outside owing to their being choked with rubble, and there is only a faint trace of them visible inside the church. Entering now by the west door, two steps lead down to the stone floor of the nave. The interior is not as dark as might be expected owing to light from the windows in the transepts falling in such a way as to accentuate the details of construction and ornament, and for a building of this size they are unusually massive. The vault and walls are green with moisture, and after an hour spent inside, the leaves of my sketch book were too wet to draw upon. The nave and aisles are divided into three bays by two heavy square piers and three round arches supporting the vault. The latter is made of fifteen rows of square cut stones springing from a fine bold cornice 4'48 met. from the floor. This cornice is continued on both sides of the nave from the west wall to the west chancel arch supporting the dome, and it also occurs on the east 66 S. GIOVANNI IN SINIS side of the north transept and the return face of the apse arch. The cornice in the rest of the building is of a different and much smaller pattern. Plain square stone seats are placed at the foot of each pillar and on either side of the west door as well as on either side of the chancel. The curious vents or windows I have spoken of were blocked up inside, and a mark or stain visible just above the cornice in the photograph indicates their position. There can be no doubt, I think, that the nave and the rest of the church were built at different times, for the chancel arch and piers and the transept vaults are of smaller span than the vault of the nave. The chancel arch is still unfinished, and at the place where it joins the nave vault there is a break. The nave cornice, too, stops abruptly at the chancel arch, and the cornice on the chancel arch is not only of a smaller pattern but at a different level. The odd thing about that is that a piece of cornice exactly like the nave cornice is repeated in the north transept.^ If the nave and its cornice are of very early date, so also is the dome supported on plain pedentives, for, like all early domes, it is very shallow and inclines at almost the same angle as the pedentives. It is composed of square cut stones laid in thirteen rows and finished up to the crown, and I could find no indication of the dome crown being open as La Marmora seems to imply ; but as he says he was never inside the church his suggestion cannot be relied on. The piers supporting the dome are recessed to admit of an ornamental marble pillar and this decorative device of African origin will be found at S. Saturnine, in the church at Sohag and in most of the Norman-Byzantine buildings in Sicily. It also occurs in the Byzantine church at El Kef in Tunis. But there is nothing to show whether these recesses are part of the original design or were cut later. The recess in the south-west pilaster has now been filled up. The transepts are alike save in the detail of the cornice I have already noticed. The windows in the east sides, visible from out- side, have been blocked up and the arched recesses in the east wall are now used as chapels. Some alterations in the north-west corner of the north transept close to the cill of the window, show 1. The nave cornice is exactly like one in a very early church at Hierapolis, near Laodicea, in Asia Minor, usually ascribed to the 4th century. S/N/S. Church of 5. Gnwanni. Interior showtufi;^ the chancel arch, dotne and apse. To face page 66. .i^L , \'.u\\v jo^xX^ i2»,']\|» ^-MnvK^ 1 marble decorative devici .?>^ •i>;ft^>^w\'iN^ S. GIOVANNI IN SINIS 67 that the cornices do not fit and that the transepts have been restored or altered apparently at a later date. The chancel apse is plain and small in proportion to the size of the church and calls for no comment. The altars are raised two steps above the floor and appear to be comparatively modern. It remains to be said that there is no trace of either wall paintings, mosaics or marble, nor could I find any dedication crosses or masons' marks on the walls either inside or out. The whole of this interesting church has the appearance of great antiquity. The nave arches, the cornice and the seats seem to be the older parts, and the nave generally resembles S. Phocas at Priolo, but the break between the nave and chancel and the rough unfinished arch of the latter need to be accounted for, and Dr. Scano considers that the domed intersection is the oldest part of the church. About a mile north of Sinis in a spot called the agro di Cabras there is a large compound about two acres square, surrounded by huts provided for pilgrims who come here twice a year. In the centre of this kind of village green there is a small chapel of late date. Underneath it a gallery or crypt leads to a cruciform rock cut chapel or baptistry, about ten feet below the earth level, with a well in the middle of it. It has been suggested that this is a very early Christian remain of the period of the perscutions, and that the plan of S. Giovanni was copied from it. I was unable to make any examination of the chapel as the floor was knee deep in water. The train going southward crosses the Campidano or plain of Oristano and between S. Gavino and San Luri passes almost imperceptibly over the watershed into the plain of Cagliari. The journey is not interesting and the flat and ugly landscape on either side of the line is bounded by ranges of rugged hills from six to eight miles distant. Assemini is the last station but one about nine miles from Cagliari, and a little beyond it the line crosses some marshes and skirting large lagoons enters the city on the west side. Cagliari is situated facing the sea on a low limestone hill surrounded by lagoons and marshes. Owing to the impregnable 68 CAGLIABI position a settlement existed from the earliest times. The mediaeval city,* known as Calaris, is now represented by the upper town containing the citadel, the public buildings, and the cathedral. The modern town is built on the southern slopes terminating in a fine marina facing the harbour. It has the reputation of being the hottest spot in the kingdom. The cathedral, originally a Pisan building, has been refaced and defaced in barroque style, and is covered within and without with plaster and whitewash. Some attempt has been made to restore the west front by removing the plaster, with the result that the building presents a most delapidated appearance. The interior is unusually clean and pleasing, and contains some of the furniture of the original cathedral including two ambones or pulpits, finely carved, and crouching lions in marble on either side of the chancel steps, all in the usual Pisan style. The see is an archbishopric with the titular rank of the primacy of Sardinia and Corsica, and, so far as the Archbishop's rank permits, he is dependent on the Archbishop of Pisa. The Primacy of Corsica must, in fact, for at least a century have been merely titular, for at the French Kevolution the ancient dioceses of Corsica were suppressed and united in the Bishopric of Ajaccio now dependent on the Metropolitan of Aix in Provence. The Church of S. Saturnino is the largest byzantine building I have seen in this part of Europe. It has been much mutilated and restored, and only the choir and domed lantern are used as a church. The adjoining walls on the west side may have formed part of a basilica mentioned in the life of S. Fulgentius. But it would seem that whatever be their date or their original purpose, they afterwards formed the outer aisle walls of the nave of a later church. S. Fulgentius was exiled to Sardinia with other African bishops by Trasamond in the sixth century, and obtained leave to found a monastery at Cagliari near the basilica of S. Saturnino. The basilica then in existence must have been built, therefore, in the earliest style of Christian architecture, and it remains to be proved whether any part of the existing remains belong to it. At a much later date S. Saturnino was assigned to the guild of doctors and apothecaries of Cagliari, and the invocation was changed to Ss. Cosmo and Damiano, and by those names it is now known. -Tf. \>^ V^Yi^^ Titi\ i>^^ 4Vt Ai6-io'^ .l>f. 39. CAGLIARI. Abbey church of S. Satiirm'no apse and N. side ; part of the north wall with a dedication cross; an ancient tombstone^ with a cross, now in the wall of the nave, and part of the base of one of the nave pillars. m- ?\ Wmm^r Corbels in the squinches of the dome. To face page Ji. S. SATUENINO, CAGLIARI 71 among them is a small stone slab with a rude carving of a Maltese cross, ^ a spear and what may be the reed and sponge. This was probably an early Christian grave-stone and indicates that these walls do not date from the earliest Christian period. I have already given my reason for thinking that the present building in the main is not later than the first quarter of the eleventh century. But probably it was not finished when the Pisans came to be masters of Cagliari, for though the choir nave and aisles retain the Byzantine form, they have been refaced and the usual arcading of the Pisan period will be seen in good preservation under the eaves on the west side of the choir.- The arches in the lower part of the north transept apparently lead to a crypt, where there may be traces of an earlier building, for there can be, I think, no doubt that the present church has been built at various times and largely with old materials ; while the Maltese cross on the walls and the crosses on the corbels in the ambulatories appear to belong respectively to the earliest Christian and to the Justinian periods the choir has certainly been restored as late as the Pisan occupation. But in character the church is certainly neither Pisan nor Lombard, and while I think it is later than might be expected, the plan and construction of the dome and the similarity to S. Sophia at Salonica leave, as it seems to me, no room for doubt that it was built by a Greek architect and I should ascribe it to the first half of the eleventh century. Assemini, a poor village in a rich agricultural district, possesses a large fourteenth century church dedicated to S. Peter and a small Byzantine chapel dedicated to S. John. The latter, an interesting building, probably built in the tenth century, resembles in some details the S. Croce in Camerina chapels in Sicily. There must have been a considerable byzantine settlement here- abouts for Greek inscriptions have been found in the churches of several neighbouring villages, at Villasor and Decimoputzu, at S. Antioco in Sulcis on the west coast near Iglesias and at 1. Compare this cross with a group of three in S. Sofia at Constantinople. Antoniades, p. 236, vol. II., see below. 2. The church was reconsecrated in 1119. The elevation should be compared with that of the Mone tou Li bos at Constantinople ; this Church was built in 900-950 and pulled down in 1904. Figured on p. 325 of Byzantinai Ekklesiai by Paspatis, 1877. 72 ASSEMINI Maracalagonis behind Cagliari. Unfortunately all these churches have been either destroyed or remodelled beyond recognition, and S. Giovanni at Assemini is the only one that has survived. Fortunately the inside was not affected by alterations made to the church about forty years ago when the outside was completely changed. The original ground plan is a Greek cross with nave, chancel and transepts of equal length, and intersection covered by a high pitched dome supported on round arches and squinches in the four angles. The rest of the building is covered by a stone waggon vault springing from a cornice. There is a small semi-circular apse at the east end and the transepts are closed by flat walls. Whatever claim to merit the outside elevation of this little chapel originally possessed has been completely destroyed by the insertion of aisles in the four corners of the cross so that the ground plan of the present building is a square. The west end facing on to a little close in the village has a plain door and a square window above it, and, as at S. Giovanni in Sinis, the vault is carried forward making a semi-circular cornice at the top of the facade ; this way of carrying out the vault has also been adopted at the end of the transepts and chancel. Entering the west door and descending two steps we come to the floor now raised half a metre above the original level ; the intervening space filled in some places with bones was used for burial. The walls of the nave, chancel and transepts have been pierced with plain arches to give access to the modern aisles ; on the whole this has improved the general appear- ance inside, but even now it is of the plainest description, and whatever decoration may have existed has been covered with white- wash. The dome inside is of the same diameter as the square of the intersection and consequently fits on to the front of the four arches supporting it. In that respect it differs from S. Saturnine and the S. Croce chapels where the domes are wider than the square. Outside it resembles S. Saturnino in having half the height concealed by the square superstructure of the intersection carried up to counterweigh the thrust of the dome. The following measurements will show how small this little chapel is. The total length (including the apse, 1*50 met.) is 10 met. 95 cent. The square under the dome is 2 met. Each ASSEMINL Church of S. Giovanni. Interior looking East; exterior from S. W. ; Crosses on the wall in the church of S. Pietro. Plan of S. Gvwanni (Scano); the heavy line shnvs the original church. To face page J2. .Of V a stoii Hi * a pk in d in »iria, thevft S',i,V\f\ VS\N\_'.i'*^ ASSEMINI 73 transept is 3 met. 58" long. The width of the nave and of each transept is 2 met. The cornice or spring of the vault is 2 met. 79" from the present floor ; the cornice itself is 22 centimetres thick and the height of the vault 1 met. 70- from the cornice. This gives a total height from the floor to the crown of the vault of 4 met. 71'. The walls are uniformly 60 centimetres thick. The following inscription on a stone found in this church is rendered by Besta^ thus : — JK[up;]e 0o)'idc jou aou SouXtv rcopKOTopqov cipXovTO^ 5'cf/3S>;J7c[s'] K\^a]i its ^OV\Lv K(i}pv(f>e(ov 'AiroaToXwv JJeTpov Ka\ TIavXov kuI tov M7/01' '/'uavvof Tou BairirvcrTov, koX tt}?) IlapdevofxdpTvpo^ Bapl3dpa vol. I., p. 25. 2. Gay, cap. 5. 62 CALABRIA Squillace, the last Byzantine fortress in Calabria, and the garrison finding further resistance useless surrendered and embarked for Constantinople.^ With their departure the Greek dominion in Calabria came to an end and passed into the hands of the Normans. There remained the Church and clergy who were attached to the Patriarchate of Constantinople and the monks of the. Basilian order who were also Greeks. Here I must go back for a moment to two events that preceded the surrender of Squillace, and taken together directly affected the Calabrian Church : the rupture between the Greek and Latin Churches in the reign of Constantine Monomachos in 1054, and the treaty of Melfi between the Normans and the Roman see in 1059. The Greek and Latin Churches having failed to agree upon the questions in dispute, not the least important being that of the Roman supremacy, the Latins determined to utilise the Normans to regain the patrimony and the spiritual jurisdiction that had been taken away from them by the Emperor Leo III. in the 8th century. The bargain with the Normans was struck at the Council of Melfi between the Pontiff Nicholas II. and Robert Guiscard, the latter undertaking to place all the churches in his dominions under the jurisdiction of Rome and the former conferring upon Guiscard the duchies of Calabria, Apulia and Sicily.^ According to this arrangement, within 30 years of the Norman conquest, the metropolitan sees of Reggio and Sta. Severina and the suffragan sees of Squillace, Cassano, Bisignano, Cerenzia, Umbria- tico, Isola di capo Rizzuto, Nicastro, and Tropea were all Latinised by the appointment of Latin bishops and the introduction of the Latin liturgy. Where the Greek community was numerous or important as at Rossano, Gerace, Cotrone, Bova and Stilo, the promise made at the treaty of Melfi was not kept and the Greek bishops and liturgy were retained. The change would have been effected more rapidly and completely but for an incident in the history of Rossano in 1093, ' when the Greek archbishop died, and the citizens revolted against Roger the Duke of Calabria, Guiscard's son, because he appointed a 1. Chalandon, vol. I., p. 174. 2. Gay, p. 518. At the council of Melfi in 1089, Urban II. conferred the duchies of Apulia and Calabria on Guiscard's son, the Duke Roger. Chalandon, vol. I., p. 297. CALABHIA SB Latin to the vacant see. This revolt would no doubt, in other circumstances, have been suppressed and the Latin appointment confirmed. But Rossano was an important fortress, the key of the road between Apulia and Taranto and Western Calabria, and the revolt happened to synchronize with a rebellion of discontented Calabrian barons, among them notably the Duke's brother-in-law, William of Grantmesnil. Having seized the surrounding country Grantmesnil offered his services to the citizens who accepted them, and with Rossano in hand he bid his kinsmen defiance. Many of the barons followed Grantmesnil's lead, and the whole of the Norman territory was soon in a condition of anarchy. By this rebellion the suzerainty of the Hauteville family was put in peril, and the circumstances were considered sufficiently serious to bring to Rossano the Duke, his brother Bohemond, and their uncle Count Roger, of Sicily, with a numerous army. The combination of circumstances was fortunate for the Greeks and they gained their point. Upon Roger offering to allow them to elect their own bishop they surrendered the city and Grantmesnil's rebellion collapsed. By the arrangement then made a succession of Greek bishops held the see till the year 1364 and the holder became the metropolitan for the Greek community in S. Italy. The experience at Rossano was not thrown away. In the principal Greek centres the bishops and clergy were retained and the Norman law, enabling religious communities to hold property, was applied to the many monasteries that sj)rang up all over Calabria. This was a most important change, and it brought much wealth and prosperity to the Basilian houses in Calabria. Gerace kept its Greek bishopric till 1497, the last bishop and one of his predecessors being Constantinopolitans. Cotrone retained its Greek bishopric till 1261, two of the prelates being sent on missions from Rome to the Court of the Byzantine Emperors. Oppido retained its Greek bishopric till 1349, and there appears to have been a Greek bishop at Stilo for Some time after the Norman conquest. At Bova the cathedral with its chapter and clergy retained the Greek rite till the end of the 16th century. The policy of conciliating the Greeks that these princes were prudent enough to follow had been initiated by their kinsman Robert Guiscavd. He seems to have been inspired by a sense of $4 CALABEIA imagination, conceiving himself to be the legitimate successor of the Byzantine Emperor : under his rule Calabria was still called a theme, a strategos or an exarch ruled the cities, spatharios, turmarque, and other imperial titles are appended to names in Norman charters, and in his person he affected the dress and copied the seal of the Emperor.^ The passage from Martroye given below is interesting for the comparison it suggests between Theodoric, of whom it is written, and Gaiscard to whom it may be applied. I may convieniently conclude these historical notes with a passing reference to the Latin occupation of Constantinople (1204-1237), to the development of the schism between the Greek and Latin Churches that ensued, to the end of the Norman kings by the defeat and death of Manfred (1266), and to the ecclesiastical policy of the Angevins who succeeded them in Sicily and S. Italy. These events in the general history of S. Europe, each in turn, contributed to a change in the fortunes of the Greeks and their Church in southern Italy. After the death of Manfred what I may term the modus vivendi maintained by the Norman sovereigns for the two Churches in Calabria since the reign of Constantino Monomachos, came gradu- ally to an end, and the Greeks were presented with the alternative of conforming to the Latin Church or accepting the position of schismatics and all the disabilities that followed from it,'^ By the middle of the sixteenth century the Greeks who had not emigrated or amalgamated with the Italian population were reduced to small communities in one or two isolated spots, like Bova, in the extremity of the Calabrian peninsula. The traveller who goes to Calabria must be prepared for the simplest lodging, the plainest fare, and a good deal of discomfort. The best months to travel in are from December to April, and 1. Chalandon, vol. I., p. 260. Compare with Martroye, p. 73. * II ^vita avec soin tout ce qui aurait pu humilier ou exasperer les vaincus. ' II temoigna la plus vive admiration pour leur civilisation et leurs monu- ' ments, leur laissa leur lois et leurs institutions, ne les inquieta point dans ' leurs croyances, affocta d'agir, de se vetir, de parler comme un Romain. II ' esperait faire illusion a ses sujets, au point de leur persuader que c'etait • toujour* TEmpire rotnain qui les gouvernait ' 2. Battifol, p. xxxvi. y^\ CALABRIA. y/if / i I, To face page 84. v% -^^ft^ Tjcugorie^ro S. Chzrico Gtclf of Medi te rraneart SecL ^V, \jSybccris ^ Core^Lwcrwy issano Cosenzcc S. Euphemic^ StaSeverina^ Catcuvzaro^ • ^TRoceUettcL -. J 0tx.lf of CS c^quUlace StalctttK SquMcuce StUc ^ * * ^ QercLce Locris ^Monasterojce I ortian Sea Bora CSpartivento \ STALETTI 85 though they may be wet and cold there are frequent intervals of brilliant sunshine, when the temperature is like that of a warm spring day in England and very pleasant. In other seasons the climate is very hot, and in the autumn the country is full of mosquitos and malaria. A carriage attached to the evening post train from Naples to Reggio, labelled Cosenza and usually filled with passengers, affords the quickest means of reaching S. Calabria. It leaves the main line at Sta. Eufemia, and crossing the Appenines joins the south coast line (from Reggio to Taranto) at Catanzaro Marina. The modest buffet and rooms over the station there will provide the traveller with food and lodging for a week, and may be used as a convenient centre to visit the churches I am going to describe. Those who are not pressed for time will wisely prefer to go from Naples to Reggio, or better still Messina, and make one or other of those cities their headquarters. They should provide themselves with food and bedding, and proceed to Catanzaro by the railway round Cape Spartivento. I had perhaps better begin with the Rocelletta and Staletti, as they are within a few kilometres of Catanzaro Marina, and easily accessible from the turnpike road leading along the sea shore to Reggio. The hospice of S. Gregory Thaumaturgos at Staletti claims to have been founded by Cassiodorus, the secretary to Theodoric, in the early part of the 6th century. The same claim is made for the Rocelletta. There is, however, no evidence to support one claim or the other. All that can be said is that Cassiodorus was a native of Squillace near by, and that somewhere in the neighbourhood he founded a religious house where the inmates devoted their spare time to agriculture and the copying of manuscripts. I should add that no part of the present buildings above ground date from the 6th century ; the church at Stalletti is comparatively modern, and though some doubt may exist as to the precise date of the Rocelletta, the form of chancel was intended to accomodate a ritual that did not exist till long after Cassiodorus' days. On the 11th October, 1551,^ the Papal Commissioners sent to inspect the Basilian monasteries in Calabria visited Staletti and the 1. In the reign of Charles V. Calabria was then part of the Spanish dominions. 86 CALABRIA Rocelletta. In their report^ the first name is spelt in three different ways, Staldati, Stalatti and Staltatei. ' On the 11th day of October we reached the monastery of the * Holy (Divi) Gregory of Staldati, where we found the abbott, a ' Latin priest, and one deacon monk : the tomb of the Holy (Divi) * Gregory is there and many miracles are performed. There are * moreover many relics. The abbot of this abbey was commenda- ' tarius Tiberius Canossa and he had Latin priests to celebrate the 'office until Greeks could be obtained.' * On the 11th day of October we came to S. Mary of Old ' Squillace, which is an abbey and not a parish church, and is also 'called Episcopatus Squillacensis. Near S. Basil of Camardi 'there is an ancient abbey, which was near the sea, but the abbott 'Mark Anthony Armogica, for fear of the Turks, made a grangia 'below the village of Stalatti and called the monastery after ' S. (Sanctus) Gregory.' ' On the same day we came to the aforesaid monastery of Old ' Squillace which is only a short distance from the sea shore. We ' saw there a bare altar without covering ; the walls of the church ' were painted with different saints ; it is without door and badly * used since it is uninhabitable on account of pirates. In the said ' church three masses are celebrated daily. The abbott is a man ' of good life wont to recite his office daily : he is also a priest and ' celebrates mass from time to time.' On the following day the Commissioners visited another mon- astery, near by, called Magliotis. They found it in a neglected condition, and ordered the abbott to have the office celebrated by Latin priests until Greek monks could be sent. On the same day they visited another monastery, S. Angelus of Maid a on the Tyrrhenian side near Nicastro. There they found the church ' fabricata ad usum Graecorum,' and they reprimanded the abbott because he had a tin (stagno) chalice and ordered him to get a silver one. On the 15th May they found the monastery of S. Mary of Carra abandoned for fear of brigands. The church had been built ' more Graecorum.' These entries are interesting in several ways. In the first place they show that Staletti and the Rocelletta were still reckoned as Basilian monasteries in the middle of the 1. Translated from Battifol, p. 109. STALETTI 87 sixteenth century. The Hospice of S, Gregory had been removed from the sea-shore up to the hills near Staletti, and the Rocelletta had also been abandoned on account of the brigands and pirates ; and by ' Turks ' we are no doubt to understand Tunisian Corsairs who infested the coast and caused the towns to be moved from the shore to the high ground.^ In the next place they show that there was a deficiency of Greek clergy to perform the service. The entry in the Commissioners' report immediately preceding that of the 11th October is dated in May of the same year, and relates to a nunnery, where the abbess had, for this reason, petitioned the Roman see for leave to change the service from Greek to Latin. Lastly they show that even at that date the Rocelletta had begun to fall into ruin. The Hospice of S. Gregory is built on the crest of a high hill about five miles south of the Catanzaro Marina station and from the railway platform it can be seen standing out prominently on the sky line. It now consists of a small domed cruciform chapel, with a single apse, a refectory, cells, offices and a little garden covering altogether about an acre of ground. The chapel was rebuilt in the eighteenth century and had been recently repainted when I saw it. Portions of the refectory wall, a doorway, and some fragments were pointed out to me as belonging to an earlier building, and may have been part of the original foundation. In consideration, I suppose, of the traditional connexion with Cassiodorus, the Hospice ia still licensed by the State and main- tained by a small endowment, and a monk and lay brother minister to the wants of the old and infirm of the village. If Cassiodorus really built a monastery here the site was well chosen, for the garden stands on the edge of the headland some 900 feet high, commanding an extensive and beautiful view of the coast from Cape Stilo to Cotrone, and overlooking Squillace bay, the ' Scyllaceum nauvi- fragum ' of Ovid and Virgil.^ 1. It is difficult to realise that these Tunisian pirates were permitted to exist till well into the nineteenth century. I do not think the circumstance is men- tioned in Dean Church's life but he told me that when he was a child the Tuscan soldiers would not allow his nurse to take him walking on the mole of Leghorn harbour when a corsair ship from Tunis came into port. 2. This district retains its stormy reputation. The local saying is, Un' amico vho e cosi raro come un' giorno senza vento a Catanzaro. 88 CALABRIA The Eoccelletta del Vescovo di Squillace, to give the rum its official name, stands in an olive wood at the foot of a little hill about 200 yards north of the high road to Squillace and Staletti and a mile from the Catanzaro Marina railway station. It can be easily seen from the train and must not be confused with Rocella, a pretty place further down the coast between Stilo and Gerace. The earliest detailed description I have been able to find is in a German book published in 1773 by J. H. EiedeseP and the earliest picture of it is in a French book published in 1783 by the abb6 de Saint Nom.^ Riedesel describes it as follows : — ' A building ' of bricks which had been described to me as a Greek Temple ' stands below Catanzaro which next to Cosenza is the greatest ' city of Calabria and the capital of Calabria Ultra. Its form how- ' ever is such that I look upon it rather as a Gothic or a Norman * building. For though it is a parallelogram yet from the square ' towers at its corners one must need conclude that it could not be * built by the Greeks, especially as the towers plainly appear to be * of equal antiquity with the rest of the building and are not, as ' may perhaps be said of the narrow vaulted windows, built in later ' times.' The abbe de Saint Nom merely records the existence of the ruin, calling it the Rochetta, but says nothing about the architecture.^ He gives an engraving from a drawing by Chatelet, showing the surrounding country and a distant view of the ruin. The latter is too small to be of any practical use but it shows that the two towers described by Riedesel were situated at the extremities of the transepts. The ground plan is a T-shaped basilica consisting of a nave without aisles, and a chancel with three apses at the east end.'* The two parts of the church were joined by a cross intersection covered with cross vaults and supported on either side by two towers. 1. Travels through Sicily and Magna Grecia and Egypt, translated by J. H. Forster, F.R.S., from J. H. von Eiedesel, pub. C. C. Dilly, London, 1772, p. 159. 2. Voyage Pittoresque Naple et la Sidle, Abbe de Saint Nom, Paris, 1783, drawings by Chatelet, vol. III., p. 110. 3. He mentions the discovery of stone cannon balls in the precincts, and says that his journey was interrupted by the great earthquake of the 5th July, 1783. 4. The dimensions are on the plan which I have copied from J. Strzygowski's Kleinasien ein Neuland der Kunstgeschichte, Leipsic, 1903. It is taken from Caviglia. .i.l lun/' S, ji-f .Q***) ^■^Si\^"^!Si\< ■• ^^% ^: 'f ^ ? „k««!»3Sw«e»» N. and E. sides of the church. K.F. To face page go. u •V\-%\ '\ ehnreh : f f'^-"- ' fV of the church there are a farm yar* 1 in the 1 ore conapicioaB garments were a white sqa ■ ' ' > - •■ ^^^ ^^^ ^^ t.Q 'V^O^T^^'X ovjiioG and i >»t«}ietti. .z^ ':\ 45. THE ROCELLETTA View in the interior frrmi the nave looking tozvards the chancel; in the foreground the debris of the fallen transept. To face page Qf. ROCCELLETTA 91 The age of this building has been much discussed. Lenormant fixes it in the fifth century, Caviglia between 550 and 600, Bertaux during the Norman occupation, and I. Grceschel in the last decade of the eleventh century. As to the earlier dates I would observe that it seems difficult to believe that a building of this size and importance, and in this unprotected and undefendable situation could have been built, or if built have survived during the Arab invasions that continued from the end of the seventh century till the beginning of the eleventh. Those who are familiar with the churches of Constantinople will be struck by the remarkable similarity in the general appearance of the Roccelletta with that of certain buildings in the Aiwan Serai quarter near the palace of the Blachernae. The treatment of the south front of the nave, and especially the windows and the round arches over them, is almost exactly like that of the facade of part of the palace called the Tekfour Serai. ^ The upper part of this facade is usually ascribed to the early sovereigns of the dynasty of the Komneni in the eleventh century. The decoration of the apses with niches occurs in several churches in Constantinople, as for instance in the church of the ' Pantocrator,' now the Zeirek djami, built by the Empress Irene the wife of John Komnenos (1118-1143), in the church of the monastery ' tou Pantepoptou,' now the Eski Imaret djami, built by Anna, mother of the Emperor Alexius Komnenos (1081-1118), and in a chapel dedicated to S. Thecla also converted into a mosque, and called the Toklou D6d6 mesjidi. According to Anna Komnena, this chapel, situated in the precincts of the palace of the Blachernae, was built by the Emperor Isaac Komnenos (1057-1059). Two other instances will be found in a chapel dedicated to S. Nicholas, in a locality known as the Bogdan Serai ^ in the same quarter, and in the church of Sta Theodosia now known as the Gul djami. 1 . There are some excellent photographs of this building in the interesting volume Byzantine Constantinople, by A. Van Millingen ; J. Murray, 1899. 2. These churches are illustrated by Paspatis in his book Byzantinai Meletai, published at Constantinople, 1877. Not a few of them, including the church of the Mone tou Libos (like S. Saturnino at Cagliari, p. 71), are now destroyed. Zeirek, a Turkish word, means earthquake. Eski Imaret means tlie old alms- house. Toklou is obviously a corruption of Thecla. Gul djami means flower mosque. Constantinople was taken on the dedication festival day of the church, and the Turks found it decorated with flowers. 92 CALABRIA In his notes upon the Roccelletta M. Strzygowski does not allude to these buildings in Constantinople, but in considering the form and ground plan he says, * Die Roccella von Squillace ist vielleicht der alteste erhaltene und * bekannt gewordene altromanische Bau dieser Art. Seine Bedeu- * tung fur die Entwicklungsgeschichte der Romanischen Bauforiuen * ware unermesslich.' If this suggestion is well founded, then the Roccelletta is a building of the greatest archaeological importance and antiquarian interest. The use of it as a model will account for the frequent presence of three apses in Italian and Sicilian churches of the Norman period, built for the Latin or the Gallican rite, where the Communion service did not require more than one, and where the triple apse was introduced either for decorative effect or because the guilds and architects employed were accustomed to build the sanctuary of a church in that way. The three apses were, however, essential for the due performance of the Greek rite, according to the liturgy of S. Chrysostom, in use by the Greeks at this time. They appear, therefore, as a matter of course in the Roccelletta, where the Greek liturgy was to be said, and as I have already pointed out, the Roccelletta remained a Greek monastery till the sixteenth century. I have included the cathedral of Gerace among the churches described in this chapter, for it would, presumably, be one of the first churches copied from the Roccelletta, since it is the nearest, indeed the only, cathedral of the eleventh century extant in Calabria. The ground plan is almost exactly the same as that of the Roccelletta, and as Gerace was one of the few cities in S. Italy, where the Normans allowed the Greeks to retain their bishops, and the Greek rite continued in use till the fourteenth century, the three apses would presumably occur there too, as a matter of course. It is to be hoped that some more information about this great building may be brought to light in the examination of the archives of the Greek monasteries now being made by the Italian Government. But for the present, the actual date and the circum- stances of the foundation can only be matters of speculation. 46 THE ROCELLETTA. Chancel and prothesis showing recess for a pillar in the angle of the pilaster ; View of the nave from the chancel. To face page g2. d*^ -Q y^tt^ ^"i^Y «i\ . I \ \ 'A A AA3 CiSbe W^\ iied ill ii /i.i.. >>rT«atioT' GEEACE 98 Gerace represents Locris, the ancient capital of the Locri Epizephyrii. Nothing remains of the classical city but some un- explored ruins near a martello tower about a mile south of Gerace Marina. This small modern town has sprung up round the railway station, and the traveller will find accommodation, but no food, at the albergo Locri near by. The present city stands on the top of a kind of bastion of rock with precipitous sides, about six miles from the railway. A fine road leads to it, first across the plain, and then winding gradually up the hill enters the city by the east gate. Near the south gate, on a mural tablet from the ancient city, is the following inscription : — " Jovi ultimo maximo diis deabusque inmortalibus et Romae " feternae Locrenses." The view from this spot, looking over the plain, the sea, and the coast from Eoccella to Cape Spartivento, is very fine. The bishopric was known first as Locres, later as 'Agia Kyriaki, and finally as Gerace, the Latinised form of the Greek name. The removal from the classical site probably took place in the beginning of the eighth century, when the Arab raids made the coast uninhabitable The cathedral, dedicated to the Blessed Virgin, is a T-shaped basilica, consisting of a nave with two aisles, and a chancel with three apses at the east end. The two parts of the church are joined by a cross intersection covered with a dome and cross vaults. The ground plan is, therefore, like that of the Eoccelletta, and the two churches are about the same size.^ The chancel and the apses are raised above the floor of the rest of the church, and beneath them is a crypt like those of S. Nicholas at Bari, and the cathedral of Otranto. This crypt, the nave, and the south-east apse appear to be the only parts of the original cathedral still visible ; the rest has been covered up with later additions, including restorations after the great earthquake of 1783, when the church was severely shaken. The nave inside is divided into two equal lengths or bays by square pilasters on each side. Each bay has five columns bearing 1. The ground plan was almost exactly the same till the later central and S. apses were built out. Schulz gives a plan of the crypt, and the dimensions of the cathedral according to Caviglia, 282 feet long and 88 feet wide. Denkmaeler dcr Kunst des Mittelalters in Unteritalien, Dresden, 1860. Vol. II, p. 351. 94 CALABBIA round head arches, and supporting the nave wall. In the clerestory this wall is pierced with six large roundhead windows. These columns, taken from the ancient Locri, are very interesting. I noted the following on the north side : — The first pillar is red marble, the next two white marble fluted, the fourth brescia, the fifth fluted white marble with a pretty Corinthian cap. The sixth, seventh, eighth, ninth, and tenth are of granite, the eighth having a red marble top. On the south side the first column is verde- antique with a white marble Corinthian cap. The second and third have similar caps and white marble fluted shafts ; the fourth is plain white marble, the fifth cippolino, the sixth brescia in two blocks ; the seventh granite, the eighth white fluted in two blocks, and the ninth and tenth granite. The caps are of different kinds, chiefly Lombard or Corinthian, with abaci above them. The arches throughout are round and made of large stones. The general appearance of the nave will recall that of the early basilican churches of Rome and S. Gavino at Porto Torres. The roof is made of wood rafters and a ceiling of boards covered with paper. It was in a most dilapidated condition when I saw it. The east end inside has been so much altered that no part of the original work is now visible. A broad flight of steps in the north transept leads down to the crypt. The roof is cross vaulted, supported on pillars with capitals of various designs, and among them examples of the Lombard type. The crypt is therefore much more ornate than that of the Roccelletta, and the plan given by Schulz shows that it resembles more closely those at Otranto and Bari. The north-east apse still retains the original decoration on the outside. It has one window, and the eaves are supported by the small arcading usually found in Lombard buildings of this date. The south apse has been rebuilt, and a large gateway with a pointed arch, leading to the cathedral close, has been built up against it. The central apse was refaced in the 17th century. According to Schulz the cathedral was founded in 1045, that is to say while Gerace was still ruled by the Byzantines, and 14 years before the Council of Melfi and the Norman occupation. The city is often mentioned in Norman times, and was shared in equal parts by the brothers Robert Guiscard and Roger, the Count of Sicily, .T^ J \\\<-\ \vw\^ i\\vnVv»vi'A'.»\ :ul .?,Q ^<^.n\ vs^Y <^\ 47. STfLO. La Cnttolica. South side. Elevations and plan hv Stg. Ahatino from ^ Napoli Nobilissima.'' To face page gS- STILO 95 when they divided the spoils of war after the conquest of Palermo. In the course of the Xorman conquest the citizens, who were mainly Greeks, obtained ceitain privileges from the Hauteville family, including the right to appoint their own Greek bishops and to retain the Greek liturgy. Though the cathedral must have been built to accommodate the Greek rite, which continued in use at Gerace till the 14th century, the construction and details are Lombard. The explanation of this probably is that a Lombard guild of architects and workmen were employed to do the work. I should note in passing that with the exception of Siponto and Mater a this is the oldest cathedral in S. Italy. I now come to the two little chapels at Stilo and Rossano, the former known as the ' Cattolica,' and the latter dedicated to S. Mark. They are purely Byzantine in plan, design and construc- tion, and, so far as I know, they are the only examples of their kind west of the Adriatic. I have already alluded to the importance of Rossano as a Byzantine fortress and the seat of the local adminis- tration. Stilo is occasionally mentioned in the late Byzantine and early Norman history. There is some doubt as to the exact locality of the battle of Stilo, an important engagement in the German campaign organised to conquer Sicily, when the Emperor Otto was defeated, and the Arab general was killed and his army afterwards dispersed. The name of the station, ' Monastevace,' is derived from an abbey of S. John patronised by King Roger and his mother. Stilo, like Gerace, is situated about six miles from the sea, on high ground formed by the talus of a rocky spur of the Appenines. The scenery in the mountains at the back of the town is very fine and wild. On the occasion of my visit this district had for some time been the headquarters of the famous bandit Musolino, and under martial law. Of these circumstances I was unaware until I found myself escorted for the day by carabinieri from Monasterace. The Cattolica is built on a ledge of rock at the foot of the cliffs behind the town. The ground plan is square, covered by a roof divided into nine compartments. The squares in the centre and at the four angles are covered with pepper-pot shaped domes and the intervening spaces with barrel vaulting made of brick. The nave and 96 CALABRIA aisles terminate in three semi-circular apses covered by semi-domes with flat tiled roofs. The domes have flat tops, covered with tiles supported on three rows of narrow bricks under the eaves. The central dome, larger than the others, is decorated with rows of large square tiles, arranged diamond wise round the centre of the drum. It has four small windows at the cardinal points : these are divided into two lights by a small twisted column. The other four domes are also decorated with tiles in the same way lighted by single round headed windows at the cardinal points. A narrow cornice or string course of brick is carried round the drums of the domes and in contour over the windows. Each apse has a single round headed light and string course treated in the same way. The main door is on the south side : the jambs and lintel are square and made of stone : there is a discharging arch over it and a tympanum pierced with a trefoil aperture : above it is a little round headed light. The entire fabric outside is made of bricks and tiles set in mortar and rather roughly finished. Passing now to the inside : the central dome is supported on four marble pillars, all obviously taken from an older building. As two of the shafts are not long enough, ancient capitals, one of them being of the Corinthian order, have been inverted and used as bases : the caps are plain of the usual simple Byzantine shape. Two of the shafts are of white marble, the third is of granite, and the fourth of cippolino. On the pillar nearest the door on the right a dedication cross is incised.^ The domes are supported on pendentives and the arches throughout the chapel are round. Beside the incised cross there was, according to Schulz, a marble panel with scroll design surrounding a cross, but I did not notice it. The wall paintings he speaks of have almost disappeared. In the centre apse Our Saviour is represented seated on a rainbow in the act of benediction with two winged saints or cherubs on either side, the whole in an oval panel. There are paintings on the North wall near the prothesis supposed to represent S. Nicholas, in the diaconicon, and on the West wall. 1. The cross is reproduced on page 102. It is on the S.E. pillar. The accom- panying sketch was made by me in 1898. The inverted capital is shown in the foreground ; another classical fragment is lying on the floor, and on the right is the modern holy water bason. 48. ■Jfc.«"«ii:i:..-»«. «av. STILO. La Catiolica The interior looking East^ showing the inverted capital ; the holv water basin on the right. E.F. La CaUolica — Veduta delle absidi (lalo (orientale). Elevation by Sig. Abatino, from ^ Napoli Nobilissima.'^ To face page g6. .8^ hree et wise roi' iinal poi \.>\ s and lintel are square a ^and a tyi^ he;.... , .)f.i>- iff^iffifttt vm^ - •»«*• <»*o al fragtni :^^ ./-^A t [ ■^ *'\:- A vi,^... .Q^ rAV ?;^:0'A Siw-A Sift's. ■r of .i-v\^\L .'A \v> i«»\^\ .0^/VK\A ill iare mtid .^Q •t5,»\^'i»\ «f^ 49. ROSS A NO. Church of S. Mark. East End showing- three apses. Interior from N. aisle looking into the prothesis. lJ Plan of S. Mark. COREGLIANO. Church of the Patire, showing three apses at the East end. To face page p/. ROSSANO 97 The dimensions of this chapel are : ground plan of the square 6'15 met., the square under the central dome is 2'20 met., the height from the floor to the base or spring of the central dome 5'80 met., diameter of the central dome 1'60 met., and the pillars and their caps are 2*78 met. high. The little church dedicated to S. Mark at Rossano is situated at the west end of the city on a spur of rock. It consists of two parts, the church proper and a large square hall or narthex at the west end. The church itself has a square ground plan, and like the Cattolica, is divided into nine compartments. Of these the central and angle compartments are covered with high pitched pepper pot domes, the central dome being rather larger and higher than the other four, and the other compartments are covered with barrel- vaults. At the east end there are three semicircular apses, covered with semidomes, to do duty for the altar the prothesis and diaconicon respectively. These semidomes and the five domes are covered with coarse tiles. The domes have been smeared over with whitewash. The entrance to the church is by a door on the north side of the narthex. This door appears to be modern. The narthex walls are made of stone and the timber roof is hidden inside by a plain rafter ceiling. There is a stone floor and a rough bench or seat runs round the three sides of the narthex. The first sketch is taken from the narthex and shows the three arches at the west end of the church corresponding to the nave and the two aisles, the floor being raised a step above the narthex. The second sketch is taken from the south-west corner of the church. The photograph is taken from the north aisle looking up to the roof to show the pendentive. For the purpose of showing the structure of the church I have omitted the altar in the second sketch. It now stands directly under the central dome and is raised upon two stone steps. The niche on the pilaster shown in the photograph is modern ; there is a similar niche to correspond on the other side. The church was originally poorly lighted, the two square windows on the north and south sides being modern. The central dome has four single light round head windows at the cardinal points ; the smaller domes have one similar window apiece facing the outside 98 CALABRIA of the church. If the central apse had a window it is now blocked up. Each lateral apse has a single double light window, the lights being divided by a small pillar with cap and base of the usual style. On the N.W. angle of the church, at the spot where the second sketch is taken, there is a door leading out into a small triangular court and a priest's house. The dimensions of the church are : length 7'50 met., width 7"15 met., width of arches throughout 1*46 met. ; the spring of the barrel vault is 5 met. from the ground. It is therefore a trifle larger than the Cattolica, but in ground plan they are almost exactly similar, the main difference being the heavy pilasters supporting the central dome in place of the pillars. The church had been renovated and whitewashed shortly before my visit and any frescoes or paintings there may have been were consequently covered up. I understand that the restorations were made at the expense of the parish priest and this interesting little building may now be looked upon as saved for some time to come. The general appearance of these two churches, built within a short time of one another and probably by the same guild, will be familiar to anyone acquainted with the churches in Greece and Constantinople. They are not older than the Norman conquest, and would appear to belong to the 11th century or later. I was told that another chapel at Rossano of the same kind had only recently fallen down ; it stood just below S. Mark, and has been replaced by a modern square edifice. It was dedicated to S. Nicholas. The archbishop of Rossano, who kindly interested himself in my research, told me that S. Mark and S. Nicholas were the only Byzantine churches that he knew of in his diocese, or indeed in this part of Italy. Before passing from Rossano I should mention the ruin of the monastery of the Patire at Coregliano near by. This famous Basilian house, associated with the life of S. Nil, continued to exist till the beginning of the nineteenth century, when it was suppressed in the general dissolution of the monasteries by Joseph Bonaparte, King of Naples. The church, recently restored, was a basilica so. ROSSANO. Church of S. Mark. The interior seen from the Narthex. To face pa^e q8. og id that 1 for R< • ol ..... and pn' . Mark, ioeiated / \ \ I' Iv y^^^u .ov. : obtain a drn irico (no 4o! ng extir nte Ra; : 4 l.'WLH.- , *l n^esso '• >g oave with of .;^ ^^t(^\-i»\ \T .I'?. v. \xAtCK 54. TEBESSA. ^ Trefoil chapel ; staircase leading from the Basilica. Photograph by my father. Plan (M. Clariiwal) X, Baptistry; y and z, chambers. 7 o face page loj. 103 TUNIS. HENCHIR MAATRIA and SIDI MOHAMMED EL GEBIOUI. The history of the Church in North Africa after the proclamation of Christianity may be divided into three periods. The first, or Roman period, commencing in 314 and ending with the conquest of Carthage by Genseric in 437 ; the second, or Vandal period, ending with the capture of Gelimer by Belisarius in 534 ; and the third, or Byzantine period, ending after Carthage was taken by the Arabs, when the Christians were given the option of becoming Mahometans or leaving the country. From that time Christianity disappeared from North Africa, and by degrees the country, like Morocco of to-day, lapsed into a condition of anarchy and deso- lation. During the four hundred years these combined periods lasted at least two important emigrations of Christian clergy from North Africa into Sicily took place. The first occurred during the persecutions of the orthodox communities by Hunneric, the son of Genseric, in 484, and the second while the Arab invasion was in progress between 647 and 699, and again in 717 when the Christians were expelled. The Sicilian trefoil chapels at Malvagna and Maccari suggested to me that the plan and details of construction were derived from buildings erected by the Arabs in Tunis during the progress of their conquest, and some features in the chapels at Camerina indicated that they too might be of African origin but of much earlier date than Malvagna and Maccari. And it seemed likely that these plans were introduced into Sicily by the clergy who had emigrated during the Vandal and Arab persecutions. The following notes are the result of several journeys into the interior of Tunis to see if I could obtain any evidence of the origin or date of these Sicilian buildings. There appear to be only five examples of trefoil chapels in North Africa, three in Tunis, and two in Algiers. The four described in 104 TUNIS this chapter are : the trefoil chapel attached to the basilica at Tebessa : a similar but smaller chapel attached to the basilica at Carthage, called the Damus el Karita : a trefoil chapel in the plains south of Kairouan, near Sidi Amor bou Hadjla : a chapel like the last at Henchir Maatria, near Teboursouk. There is also a square chapel in the district of Sbeitla, called Haouch Khima, which may be conveniently classed with these trefoil buildings. I had better perhaps begin with the chapel at Tebessa, because it has been carefully explored and there is some evidence of the date of its erection. It has been described by the first explorers, Commandant Seriziat in 1868,^ and Commandant Clarinval in 1870; 2 also by M. Daprat in 1895-6,^ and M. Gsell in 1901.^ The accompanying plans are taken from Commandant Clarinval's and M. Duprat's articles ; to the former I have added the letter X, to indicate the position of a baptismal font which had not then been discovered, and Y and Z to two chambers on either side of a staircase; to the latter I have, added A to mark a little church of later date. The commencement of the exploration in 1868 is described by Commandant Seriziat, and it was taken up two years later by Commandant Clarinval. These explorations show that the trefoil chapel was enclosed in a rectangular structure adjoining the south side of the basilica. It was approached by a flight of steps from the floor of the basilica, the latter being raised 3 metres above ground level. The square nave of the chapel was apparently roofed by a cross vault ^ (Voute d'aretes), and covered inside with mosaics and outside with tiles." The style of vault and the general 1. Publication of the Archaeological Society of Constantine, vol. II. of the 2nd Series (vol. XII. of the Collection), 1868. 2. Same, vol. IV. of the 2nd Series (vol. XIV. of the Collection), 1870. 3. Same, vol. IX. of the 3rd Series (vol. XXX. of the Collection), 1897. 4. Les Monuments Antiques de L'Algerie, by S. Gsell, Paris, Fontemoing, 1901. 5. Gsell, p. 271. I use the term ' cross vault/ and add the French equivalent Voute d'aretes, to make it clear that I mean a plain vault and not one supported by the crossed arches found in Norman and later mediaeval architecture. My description of the vault at Camerina, on page 6, and the use of the word 'rib' is in that respect misleading. The Camerina central vaults, like those in these African chapels, were Voutes d'aretes. 6. Commandant Clarinval's article. 55. TEBESSA. Trefoil chapel. West apse. PI. XI .C£A ^^ 5ysiAS mmw s.aisiAim LaiAno&ia OfJfa^/tKf A hofs tomb found in the floor ; from the Constnntinc Archceological Society^ s publication. Photograph by my father. To face page lo^. .?.?. i^VT' ,Lli ir;-! v., ^ .dS .7'0\ V^ft\Viti\.Q\ 56. TEBESSA. Trefoil chapel. East and south apses. Plan (M. Duprat). ri ., ^^ pi — '-^ 5 -> o To face page 105. TEBESSA 105 appearance of the building inside when perfect will be best under- stood from the illustrations of the chapel El Gebioui at Hadjla. On three sides of the nave there are semi-circular apses covered by semi-domes. The walls of these apses were apparently veneered with marble and the semi-domes inside covered with mosaics. In the four angles of the building are small rectangular chambers, those adjoining the staircase being divided in two stories. This staircase occupied the fourth side of the nave on the north side. The walls of the chambers adjoining the staircase abut on, or are applied to the wall of the basilica ; they show that the trefoil chapel and the rectangular building containing it were not built at the same time as the basilica. They also show that the floors of the upper stories were supported on cross vaulting (Voute d'aretes) and the springs of the arches, made of a soft stone, still remain in. the corners. Commandants Seriziat and Clarinval found : — On the staircase, a sarcophagus of the early Christian period with the Constantinian monogram X P and three rudely carved allegorical figures. This sarcophagus is now the altar of the French church at Tebessa. The floor of the chapel was covered with earth, debris of tiles, pieces of mosaic^ and cinders, showing that the building had been destroyed by fire. At the entrance of each apse and also at the foot of the staircase stood two cippolino columns, one on either side. The bases of all and portions of the shafts of some of these pillars are still in site. The floor of the chapel was almost entirely covered with mosaic. At the spot marked D on M. Clarinval's plan, the tomb of a child was found. It was covered by a mosaic, representing a boy girt with a sword.- Beneath the mosaic they found a wood coffin with metal fittings and the bones of a child from 8 to 10 years old. M. Gsell is of opinion that the date given at the end of the 1. Gsell, p. 271. Le Carre (the nave) presentait des motifs ornementaux ; I'abside du fond (central apse) des series de calices, d'ou sortaient des ceps de vigne s'enroulant les uns dans les autres ; I'abside de gauche, des lozanges, des cercles, des croix gammees : celle de droite des oiseaux et, au milieu, un cerf . 2. Illustrated opposite. 106 TUNIS inscription falls in the year 508 and that the King's name in the last line but two should read (Thrasa) MVNDI.i In the course of excavating this tomb, the Officers discovered the original mosaic floor at a depth of 1-20. At the spot marked E-H on M. Clarinval's plan, a square pit was found faced with square stones on three sides, and rubble on the fourth. This pit was choked with debris. At 1'20 below the surface the original mosaic floor was again found, and below it two funeral urns and a small pot. Close to this spot were also found a slab, four small pillars, supposed to be part of an altar, and a small, but very beautiful piece of mosaic, recently stolen from the Museum at Tebessa, where it had been deposited. M. M. Seriziat and Clarinval express no opinion and M. M. Duprat and Gsell differ as to the age of this chapel and the purpose it was built to serve. They all agree, however, that it was not built at the same time as the basilica, and this view seems to me to be right, for the reason already given. I might mention also, that the varied and curious mason's marks found all over the masonry of the basilica do not appear on the chapel. These marks are very interesting and seem to be Berber characters.^ They exist, no doubt, elsewhere in North Africa, but I did not find them on any building I visited. M. Duprat considers that the basilica and the adjacent buildings were built for secular purposes and in- cluded a forum, market, and so on, and that this trefoil edifice I have been describing as a chapel was the public bath. It is quite true that this trefoil plan is frequently found in public baths. Examples, for instance, will be found at Thelepta and Ain Tounga. I prefer M. Gsell' s opinion that the basilica was built for a church and that the chapel may have been the original baptistry.^ The adoption for a baptistry of a trefoil plan associated with a Bath seems reasonable and probable enough. 1. Gsell, p. 273. 2. The illustration is from drawings I made. They will be also found in the publications of the Archaeological Society of Constantine, p. 74 of vol. XXX., 9th vol. of the 3rd series, 1897. Also vol. XXII., 1st vol. of 3rd series, 1883 ; and these marks should be compared with the Characters given in the Instructions du Comite des Travaux Historiques et Scientifiques Eecherche des Antiquites dans le Nord de I'Afrique. Conseils aux Archeologues et aux Voyageurs. Paris, Leroux, 1890, pp. 48-50. 3. Photographs of some Byzantine fonts will be found opposite p. 126. 57. BAPTISMAL FOI^tS. '"' S3 52 54 55 The numbers refer io page 126. To face page 106. .T2 ''.V\^\t-.?v to the >' m is fr< ■s I .: .?Vo\ ^■^liA^ '>'ivs\ o'X .'sf A 'v«^'i^'\ -.s Nii\ o\ X-i'^svV?. >i^ TW\^^wo;j:>ft o\. 58. CARTHAGE. Plan of 'the trefoil chapel attached to the christian basilica called the Damns el Karita ; and general plan. Plan (Delattre). In this plan of the Damns el Karita the little trefoil chapel is at the extremity of the semi circular cloister. To face page loj. CARTHAGE 107 The trefoil chapel attached to the basilica, known as the Damus el Karita, is much smaller than the chapel at Tebessa. After much trouble and delay the site was purchased from native proprietors by Father Delattre of the Peres Blancs. The accompanying general plan, taken from his book ^ shows the result of his excava- tions, still in progress, and the relative positions of the chapel and the basilica. The whole of this large building, or group of buildings, has been razed to the ground and nothing but the foundations and a foot or two of the walls remain. They show that the church stood north-east and south-west. On the north- east side there was a large enclosure surrounded by a cloister, and terminating at the east end in a semi-circle. The trefoil chapel stood in the middle of this semi-circle, facing the east side of the church. In shape, in general appearance, and in the position of the chapel, this enclosure must have resembled the Roman nymphaeum at Zaghouan. As the accompanying photograph shows, this source of the water supplied in Roman times to Carthage and now to Tunis, is still almost perfect. The square cella or temple occupies the centre of the semi-circle, as it were the keystone of an arch, and contained the statues and altar of the divinities. The temple has a peculiar form of cross vault (Voute d'aretes), and I shall have occasion to refer to it again. In describing the excavations M. Delattre calls the enclosure, atrium, the well or fountain, nymphaeum, and the trefoil chapel, cella trichora.^ The excavation of the chapel resulted in graves and traces of sarcophagi being found under the floor of the apses. Only a metre and a half of the walls still remain ; they were composed of a lower course of large square stones with rubble above, held together at intervals by square blocks placed vertically. The inside had been veneered with marble. The apse arches were supported by detached pillars as at Tebessa ; they and their caps have disappeared leaving only the foundations composed of large square stones to receive the bases of the pillars.^ The roof, presumably a voute d'aretes, was covered with coloured mosaics. The conclusion arrived at by M. Delattre and Professor 1. Un Pelerinage aux Buines de Carthage. Pub, Poncet, Lyons, 1906. 2. The small plan is taken from measurements made by me in 1910. 3. These blocks are marked A on ray plan. 108 TUNIS de Rossi in regard to the purpose of the trefoil chapel seems to be that it was a martyr's shrine. There is no evidence to prove the date either of the chapel or the other parts of this group of buildings. The excavations of the basilica uncovered some cisterns, a Roman Columbarium in the shape of a trefoil built before the foundation of the church (on the north side) and some additions made after it was completed. A baptismal font and a chapel connected with it were found on the south-west side. The former is of a common type and no guide to the date of the church. It is remarkable that these buildings have been so completely destroyed, whereas the mosaic floors, parts of pillars and their caps, the baptistry and even chancel screens, have been preserved in the church on the south side of the city. The explanation seems to be that the floor and subsoil of the Damus el Karita were ransacked for tombs and epitaphs, and a large mass of materials consisting of the more interesting decorative details, like caps and mouldings, were taken to the Seminary garden at Carthage. If these materials were carved for the basilica and not taken from older buildings, they show that the church was built in the fourth or early fifth century, before the Roman patterns had been modified or replaced by the native Berber or Byzantine designs. As the buildings described above are in the middle of large and important cities, some indication of their age may be obtained from their position, surroundings, and the result of excavation. Some idea of their general appearance may be derived from the chapels at Maatria and Hadjla. These chapels, unconnected with any great buildings, are in remote places in the interior of the country and owe their fortunate preservation to that circumstance. Henchir Maatria, an Arab village situated in a valley about 6 kilom. north of Teboursouk, on the bridle road leading to Pont de Trajan, stands on the site of a large Roman country town. The Mahometan Zaouia or chapel of the village occupies a site in the ancient forum, and the walls appear to be part of a small Roman temple that preceded it. A large stone with figures of cupids and 59. HENCHIR MA ATRIA. The chapel looking West. On the right the N. apse a7id the N. side of the lantern showing long and short work in the angle. The east wall and all but a few stones of the East arch have fallen down disclosing the interior and S. W. angle of the lantern where traces of the cross vault are visible. On the extreme left is the S. apse. » i O I 2 3^5 } ■* Metres. Plan (E.H.F.) To face page 1 08. .QZ •\\\ / / ^ ♦ t i \ o >■ I ■ i ■ I.I V .^\ ^ -x 1 "hi)\ \. til iigUTi iP. '--T!« i ^i"^ >^ .Od "\0 Ai>i'A'i\«\ \^ov vito\ViV>\?iX 60. HENCHIR MA ATRIA. Interior of the chapel looking West, showing the interior of the lantern and the slab in the angle to support a spring of the voute d' aretes. To face page log. MAATRIA 109 garlands of flowers, carved in the style of the best Roman period, lies in front of the plinth of the temple. This chapel was noticed by M. Diehl,^ but it has evidently not been excavated nor, so far as I am aware, has it been described by the French antiquaries. The plan shows that in shape and general design it resembled the chapels at Tebessa and Carthage. The main entrance faces N.E. and the central apse S.W. The chief architectural features in it are : — A square nave with semi-circular apses on three sides and a narthex or porch on the fourth side making a cruciform ground plan. A square super-structure over the nave, lighted by four windows over the apse arches and covered by a voute d'aretes with a flat roof outside. The heads of the windows were segments of circles in the form usually found in Roman buildings. The semi-circular apses were covered with semi-domes inside. The material is chiefly rubble bound together with a very strong cement and finished off at the angles with squared stones in long and short work. The walls are thick, ranging from 1 to 1| metres. The springs of the vault rest on flat stones inserted into the angles of the nave. The apse arches rest on square pilasters made of blocks of stone, some of them dressed and taken from Roman buildings. The outside roofs of the nave and apses were covered with cement and appear to have been flat. There are no traces of ornaments, mosaics, or carved marbles. The illustration shows that the narthex is gone, the South apse has been almost entirely destroyed, while the West apse has been split into fragments, apparently by an earthquake. The interesting features about this chapel are : the small portions of the flat roof that remain : the method of supporting the vault by stones or slabs inserted in the angles of the nave ; upon these stones the pendentive and voute d'aretes rested : the long and short work in the upper part of the nave wall shown in the photograph : the use of old materials taken from Roman buildings. 1. L'Afrique Byzantine, p. 422. 110 TUNIS The interior is filled with rubble from the fallen roof and walls, and the d6bris have been allowed to remain where they fell undisturbed. The ground is consequently raised considerably above the original floor level and, judging by the piers of the nave, the floor must be about 1 met. 50 below the present surface. Until the interior has been cleared and excavated, and the arable land round the outside explored, it is of course impossible to say with certainty for what purpose this chapel was built, whether as a village church, a baptistry or a memorial chapel. The long and short work and the presence of dressed stones from older buildings indicate that it was not built in the 4th century, directly after Christianity was proclaimed. I noticed masonry of the same kind, near the forum, appearing to belong to a Byzantine block house or fort. These military block houses are common in this part of the country, especially in spots of strategic importance like Maatria. I should attribute this building to the same date as the fort, that is to say, the first years after the Byzantine conquest in the middle of the 6th century. The chapel of El Gebioui near Sidi Amor bou Hadjla is officially known as the Zaouia of Sidi Mohammed el Gebioui, pronounced Ke-bee-you. As the name implies, it has been appropriated for the burial of a Mahometan saint. In the older military map^ it is marked under this name, but in the later one^ as Ksar el Beo. I understand the words el K6-bee-you to mean " the domes " in local Bedouin dialect. I had no end of difficulty in locating this interesting building, situated in longitude 89' 60" E, latitude 39' 20" N. near the old caravan road between Kairouan and Gabes by Skirra. This read leaves Kairouan by the south gate and crossing the river Zeroud at a ford, passes by the village of Sidi Amor el Kenani. Thence it crosses the railway about 2 kilometres West of the station and continues in a more or less direct line to the two Zaouia and group of seven or eight Arab houses, constituting the village of Sidi Amor bou Hadjla at 34 kilometres from Kairouan. The chapel is 12 1. Au 200,000 erne (1881-1887). 2. Au 100,000 erne (1896-1899). 61. SID I MOHAMMED. EL GEBIOUI. N. W. view of the chapel. 12. 3 -f 5" t il I I I — ^ /Metres Plan (E.H.F.) To face page no. Id .c\^A\[Ar.kVvO\r. \viv<. ii Cluq.i liie k> frmn nl.i t in the 4th c t«ed mason rv of ti ^'sf^^*>^^ i,>vxvja>itt»\*^^ AA\ V;.Ji\ yiv\ «jX \ V 1 ; ;i^t;ia each faoe of innot have had much claim 11. : . : ' live oatside foil 1 • 1 1 • .^W V; t*«atistrr»tr -Tss. j««{^4:iysof l:u%»e drew' idtogetber superior to that of mm, and a shallow narthex or po general appearance of the building can Taacha T have already roferr lid, fron .; 01 til'..' ' .VV\ 'H"V^to\ V\\i\ i>'\ Bi^ertcL TahcLrc • ^^Jf^ TtcniSi MaatricL ^Ain TcungcL •TeboLcrsoufc^ f HamnuLmet Tibar* •J^ougga y p •EC Kef * Fnfida + ^ ^Sousse Tebessa "'^Jv "^^Diei Kasrirh • ^ticLcach Hhumt [Fl Rosfe/v Sfaxt ThinncL jSbeitia. QebiouC "^ ^Ras CapoucUco mFericLThou Menchir debeul ^^afscv Oabes Tunisia i 1 >j > TUNIS 115 far failed to identify any particular church with this early period in North Africa. From the writings of the African Fathers it is certain that soon after 314 many churches were built. At first they were built, as the ruins show, in the suburbs of the cities, in order, it is said, not to offend the pagans, but, as seems more probable, to be near the Christian cemeteries. It is also certain that directly after the Arab conquests in the end of the 7th century when the Christian clergy were dispersed and exiled, the native Berber population in a body became Mahometans, and church building came to an abrupt end. Unlike their successors, the Turks, the Arabs do not seem to have converted the churches into mosques. From the cinders found in the debris, many churches appear to have been burnt out, but the shell or fabric was often, if not usually, left to perish away gradually and finally fall into a heap of ruins. If the proclamation of Christianity affords a convenient terminus a quo for the study and dating of Christian architecture in North Africa, the Arab conquest with the proscription of Christianity and the dispersal of the Christians that ensued, supplies a correspond- ingly convenient terminus ad quem. To fix the date of any particular church within this period is a very difficult matter, as the architecture, the basilican plan, the timber and tile roofing, the position of the altar, the choir or tribunes, the semi- circular apse with clergy seats in tiers, and the bishops' throne in the centre, were the same throughout the whole period. Enough is left to show that the larger churches resemble the Roman basilicas of S. Clemente or S. Lorenzo in general aspect and internal arrangement. A few churches in sea-port towns on the East coast of Tunis were apparently decorated with capitals, consoles, and other carvings designed from Constantinople patterns of the Justinian period.'^ But in the Interior the churches were decorated either with poor copies of Roman models or native designs derived from them, and they afford no evidence of date, nor can any evidence be obtained from the fonts or baptistries, for they were all constructed with slight modifications on one pattern. Again the triapsidal arrangement occurs in the earliest as well as the latest buildings. In Constanti- 1. The capitals in the mosque at Kairouan and some fragments at Sfax, 116 TUNIS nople this arrangement affords an unfailing test of the age of a church, because the three apses were provided to suit an elaborate ritual, introduced either during, or immediately after the Justinian period, and still practised in the Greek Church of to-day. In some African churches a single side apse occurs. In one case, where the church has been burned, broken fragments of the glass Communion vessels were found on the floor of the side apse. It has been sur- mised that these side chapels were used as vestries and for the offerings of the faithful ; if that was so, then they may be regarded in a way as the predecessors of the Byzantine prothesis and diaconicon. The ceremonial offering of bread and wine by the laity is a custom of great antiquity. It is still retained in the liturgy of S. Ambrose, followed in Milan Cathedral, and, I believe, nowhere else.^ I do not know if it was an incident of the African Church service. A conjecture that the later and more elaborate Byzantine ritual I have alluded to was copied from a use of the African Church presents itself temptingly, but there is no evidence to justify it, or indeed an assumption that these side chapels in the early African churches, whether basilican or trefoil, had any ritual significance. There is a strong probability that they had, and that is all that can safely be said. Their presence in an early church like the basilica at Tebessa came to me as a surprise, and shows that this architectural peculiarity cannot be used, as in Constan- tinople, to fix the date of a church, for it occurs also in later churches like El Kef, which were certainly built after the Byzantine conquest. I mentioned just now that stones from a Roman building had been largely used in the chapel at Maatria. The use of Roman materials in quantities in either forts or churches may generally be taken to indicate that those buildings were put up after the Byzan- tine conquest, for in the military treatise of Justinian's period, called the Nea Taktika, the engineers were expressly advised to use materials from ancient buildings in constructing their forts, and to 1. Visitors to Milan cathedral who attend High Mass will see, when the ser- vice commences, a little procession of six persons, two vergers, two old men and two old women, approach the threshold of the choir where they are received by the celebrant and present to him offerings of bread and wine to be used at the celebration. TUNIS 117 select some site near an ancient town where such materials could be readily obtained. There is ample evidence that this advice was almost invariably adopted. The task that Justinian set himself in reclaiming North Africa for the later Eoman Empire involved much more than the overthrow of the Vandal Kingdom. The country devastated during the century of Vandal dominion had to be re-settled and defended against the Berber tribes and clans, whose subjection neither the Romans nor the Vandals had succeeded in accomplishing. Un- fortunately for Justinian, one of Genseric's first acts when he conquered Africa was to destroy the Eoman fortifications. So soon therefore as Gelimer had been taken prisoner and the Vandal forces were broken up, the Byzantines set to work to build forts and block houses all over the country to keep order and protect the colonists. A great number of these forts still exist in a more or less complete condition, some few indeed are almost as perfect as the day they were put up. They are all built with stones of large size taken from Roman buildings, obviously roughly and hurriedly put together, and they show that the advice in the Nea Taktika was almost invariably adopted. In the ruined towns that I visited in the interior, I noticed that as a rule the churches were built, like the forts, with Roman materials. At Sbeitla and at Uppena, charches had been rebuilt, in part at least, with Roman materials on the sites and over the debris of earlier churches. The Byzantine basilicas at El Kef, at Announa in Algiers, and the recently discovered church in the out- skirts of Dougga, all contain Roman materials. So also do the garrison chapels built in the Byzantine forts at Haidra and Sbeitla. And to this short list I might add many more examples. The presence of Roman materials may be taken then to fix the approximate age of the chapel at Maatria between the Byzantine conquest in the middle of the 6th century and the Arab conquest in the end of the 7th. The plan, the windows, the use of tubes of pottery to build the roofs, and the form of vault, were all in turn copied from Roman buildings. And in these respects El Gebioui is like Maatria and belongs no doubt to the same period. Unfor- tunately these two buildings are devoid of any kind of decorative ornament, though near El Gebioui I found some tiles which belong 118 TUNIS to the Byzantine period and to some extent confirm my view as to the age of that chapel. I may conveniently conclude this chapter by comparing these two African chapels with those at Malvagna, Maccari, Sta. Theresa and Sta. Oroce in Camerina in Sicily. In the African chapels, built on the trefoil plan, the central chamber is covered with a cross vault, the semi-circular apses with semi-domes, and the tops of the windows are made in the shape of a segment of a circle. In the chapels at Camerina, built on a cruciform plan, the central chambers were originally covered with cross vaults but now with domes ; the chancel and transepts are covered with barrel vaults and have square ends and the windows were merely narrow slits. In the chapel at Sta. Theresa, built on a trefoil plan, the central chamber is covered with a very flat dome supported on pendentives and the semi-circular apses are covered with semi-domes. The narthex had a barrel vault, and the only door has a plain semi- circular head. In the chapels at Malvagna and Maccari, also built on the trefoil plan, the central chamber is covered with a flat dome supported on squinches, and the semi-circular apses are covered with semi- domes; the windows have rounds heads. The substantial difference between the African chapels and the Vigna and Bagno at Camerina on the one hand, and the three Byzantine Sicilian chapels on the other is in the method of roofing the central chamber or nave. In the former the cross vault was used, in the latter the dome. In North Africa this change from the vault to the dome occurred after the Arab conquest when a new architecture was introduced and the Eoman and Byzantine plans, cross vaults, and ornaments entirely disappeared. Arab mausolea, tombs, and shrines, of the same size as the Maatria and El Gebioui chapels, exist by the hundred all over the country. They show that the trefoil ground plan was replaced by a plain square one with a single rectangular ^^'v^nssau^mim^^^j^'s^a^ji^^' :m •w' % .lA/\"v\ ,Zd ,(^\\ •i'^V^fY'^">V>\^l\ 65. Interior of the Coffee house called the Koubba, showing the squinches supporting the dome. MA ATRIA. The Arab zaoiiia. To face page Jig. TUNIS 119 chamber covered by a high pitched dome or cupola supported on squinches. These whitewashed cupolas are conspicuous objects dotted about the brown landscape, and cannot fail to attract the traveller's notice. I give photographs of the exterior of the Zaouia at Maatria, near the chapel, and of the interior of an Arab domed building at Sousse, called the Koubba, built for a Moorish bath, but now used as a coffee house. These are typical examples of the exterior elevation, and the internal arrangement for supporting the dome by squinches, found in all these Arab buildings. In conclusion, I should like to point out that the architectural defect in the cross vault as a means of covering, even the moderate span of these African chapels, is at once apparent at El Gebioui. As the vault rises from the springs so the masonry gradually becomes thinner in order to reduce the weight, till on the crown it is a mere shell. So soon as the crown, which acted like the key to an arch, was broken the flatter parts of the vault fell in. This happened both at Maatria and El Gebioui, while in the others the vaults have disappeared altogether. On the other hand it is rare to see a broken cupola, except in cases where the sub-structure has given way. In practice, at any rate, the dome proved to be the more enduring form of roof, and that accounts, I suppose, for the Arabs generally adopting it to cover their buildings, and for the substitution of domes in place of the cross vaults the Camerina chapels were first provided with. The trefoil plan for building small chapels of this kind was adopted over a very wide area ; it occurs in the well-known chapel of Sainte Croix, near Aries, below the monastery of Montmajeur, and in a number of buildings, of the same kind, on the shores of lake St. Seban in Armenia. 120 TUNIS. EL KEF AND HAIDEA. The church at El Kef is built on the basilican plan with a nave and side aisles. It has a semi- circular apse with rectangular chambers on either side of it, all enclosed in a square wall at the east end ; and at the west end a narthex with three doors leading into the nave and aisles and corresponding doors into a forecourt. There was also access to the church by doors in the aisles, two on the north side and two on the south, made with the large stones, being part of a cornice taken from a Koman building. The roof of the nave, made of wood rafters and covered with tiles, was supported on marble pillars in pairs, also taken from Eoman buildings. The caps with one or two exceptions have been taken away, but several shafts and all the white marble bases are in site. The pillars marked on the accompanying plan 1, 3, 6 and 7 are of cippolino, 2 is of grey granite, and 4 of white marble fluted. The aisles and rectangular chambers on each side of the central apse were roofed with cross vaults, and some of the springs still remiiin in the walls. The central apse is made of dressed stones covered by a semi- dome. On the front of the keystone of the main arch there is a Greek cross in a circle carved in relief; the corners of the pilasters supporting the arch are cut away to take ornamental pillars one on each side. These pillars have gone but their Corinthian caps remain. The vaulting of the roof and the arrangement of the apse are the interesting features in this church. The cross shading on the small plan indicates the masonry of the east end, and the horizontal shading a narrow dado 1*75 met. above the ground. Upon this dado stood small short pillars, corresponding with and intended to appear as if they supported the ribs of the vault, forming a kind of collonade. Between these pillars the face of the wall is scooped away into shallow niches or recesses, the cornice 66i EL KEF. The East and West end of the basilica^ interior. ~1 □ □ ] r ] r CZl CZl □ The apse. The basilica. To face page 120. .dd ji«tf-t\%ii!5i raltfr!| And COV' j£ and in this the rn .Oi. \ ^'^si\ Vi vi\ ^ ) \ !'' X6 W \x\ S'^^^^ •.A 'i'M ^'^ 67. .fiT'^r-^ftSyf^.tf? t HAIDRA. Chapel in the Byzantine fortress. View looking down into the apse showing springs of the ribbed semi dome. One of the pillars supporting the chancel arch. To face page 121. EL KEF 121 above them forming a kind of rudimentary canopy. Probably these indicate seats for the clergy, for a block of masonry against the wall in the centre of the apse appear to mark the position of the bishop's throne. This arrangement occurs in the Byzantine churches at Dougga, Announa and elsewhere. It will be found in S. Eirene at Constantinople, rebuilt in the 8th Century, in the church at Torcello, also a Byzantine building, and in the ancient basilican churches of Rome. The semi-dome is decorated with six ribs springing from the cornice and converging to the centre of the vault. These arrangements can be best understood from the accompanying photograph, where fragments of the small pillars and one of the capitals with eagles at the angles will also be seen standing on the dado. Much the same scheme for decorating this apse, and also the Byzantine chapel in the fort at Haidra, seems to have been copied from the kind of Roman decoration found in a building at Henchir Gebeul.^ The Arabs in their turn adopted ribs in building domes like, for example, the coffee-house in Sousse. The recessing of angles in pilasters to take ornamental pillars has been noticed in the Norman chapels at Palermo, in S. Saturnino at Cagliari, and S. Giovanni at Sinis in Sardinia, and in the Roccelletta at Squillace in Calabria. The discovery of this device for decorating the pilasters of a church which must have been built before the end of 7th Century came as a surprise. The entrances to semi-circular apse chambers were often ornamented in this way with decorative pillars, both in Roman buildings as well as in early Christian and Byzantine churches ; the baths at Ain Tounga, and the trefoil chapels at Tebessa and Carthage, and the churches at Announa and Uppena, for instance, are decorated in this way. But in all these early examples the pillars stand clear of the angle of the pilasters. I have already pointed out that this recessing of angles is not found in any churches in Constantinople, and it would seem from these African buildings to have been an adaptation of a Roman idea. 1. Henchir Gebeul, situated about 12 kilotn. west of Feriana^ on one of the old roads leading from the south to Tebessa, was apparently a place of some importance, probably a small fortress at the entrance of a mountain pass leading from a lower to a higher jjlateau. Many of these Eoman towns, like Haidra, Kasserine, Thelepta, Haouch Khima, to name only a few, are situated in strategic positions to command the roads leading up to Tebessa. 122 HAIDRA The garrison chapel at Haidra stands in the middle of the west wall of the Byzantine fort. This great enclosure of about ten acres is surrounded by a high wall with bastions at the corners and at intervals along the sides. The chapel stands at right angles to the wall, and the chancel is built up against one of the bastions. It seems to have been a little basilica with nave and aisles and a single apse. The site has not been excavated and the debris lie where they fell. The base of a tower at the angle, a portion of a pilaster, and the spring of one of the nave arches all on the north side, alone remain. The lower part of the apse, built against the bastion, has been preserved up to the spring of the semi-dome that covered it. The chancel arch was supported by two cippolino columns, one on either side. A few stones of the arch remain resting on the abacus of the pillar. The abacus, if I may for convenience use the term, was decorated, as the photograph shows, with conventional sunflowers or perhaps daisies. The corres- ponding pillar, also of cippolino, has fallen out of its place and is standing inclined at an angle half buried in the ground. The apse, full of the debris of the semi-dome and of the chancel arch, built and decorated on the same principle as El Kef, is the interesting feature in this church. The African churches in general are built on the basilican principle found at El Kef ; the divisions of the nave into choir, altar space, tribunes, and so on, can generally be traced by the patterns in the mosaic flooring and by fragments of the marble or stone screens^ that separated the different parts. I may conveniently close this chapter by describing some examples of Roman, Native, and Byzantine decorative art found in various places in the Regency. The capitals, 1 in the El Djem amphitheatre, 2 at Sbeitla, 3 in the garden of the hospice of S. Joseph of Tibar, are examples of the common Roman-Corinthian style, distinguished by the ornament resembling a torch handle always occuring with marked prominence. The date of the example at El Djem can be fixed in the middle of the 3rd Century, for the amphitheatre was built in 1. Asj for instance at S. Cleinente in Eonie. 68. ^,_____^ H^ CAPITALS. 1. Ji/ Djem, Theatfe. 2. Sbeitla. 3. Ttbar, Monastery Garden. 4. Ban Ficha. Factors yard. 5. Henchir Gehenl. 6. Carthage^ small basilica. 7. Algiers^ Musettm. 8. Algiers.! Musenni. 9. G lie I ma., Public Garden. 10. Algiers, Museum. 1 1 «//(!e of h. iixlthian sty he amp' I .^s.\ •^"^v^\ v>i^ o\ iK« 'vkX^ \vv. ^^ L^m•t•^^\f\ ,\xm'v\\AiA vAXmX^v'^ J.I "i^v^^ -iuiX vA Sy»^t t^ •rw'J^\wo"j;jri o\ Capitals on Plate 6g to face page I2J^ contmued. The following Capitals ivill also he fonnd at: — i8. S. Sofa, Constantinople, ig. Parenzo, Istria. -m 20. S. Sofia, Sa Ionic a. 21. Church of the Wisdom of God, Lower Kings-wood, Surrey ; from the church of S. Johti Studiicm, Constantinople. 22. Kutchuk Agia Sofia, Constantinople ; S. Demetrius, Salonica, and S. Vitale at Ravenna. 23. Museum at Alexandria and S. Vitale at Ravenna. 23a. Capella Palatina, Palermo, and the mosque of Tulun, Cairo. 24 and 27. Parcnzo Cathedral. 25. Mosque of Amr, Cairo and S. Vitale at Ravenna. To accompany fly sheet to face page 12J. .v>d V,\v^\\\\vJ m\\ 'tows* ^ltV'\V«\TS.'fjBl .^Li ^^»\ ^^A! ""^^ 69. CAPITALS. 1 2a. Sheitla. 15. Gnelma. 16. Algiers Museum. 17. Is illustrated in Vol. ii. Ija. Giielma. 18 to 2j. Kairouaii^ Great Mosque. 24. Is illustrated opposite pp. 102. 25. Sbeitla^ garrison chapel. 26. Is omitted. 27. Kairouan^ mosque of the Barber and Parenzo Cathedral. 44. Sfax^ in the outer wall of the principal mosque. I2a^ 75, 16, ly and iJa are native designs. 18 to 2j are Byzantine and the Capitals are probably imported from abroad. 24^ 25 (2 views) and 27 have eagles at the angles. Duplicates of some of these Capitals will be found in Constantinople .1 Cairo^ Salonica, Alexandria., Parenzo and elsewhere. A list is given on the annexed sheet. To face page 123. CARVED CAPITALS 123 honour of the election of the pro-consul Gordian as Emperor in 238. Like so many other Roman and Byzantine capitals, the carving on it has never been completed. The capitals, 4 in the garden of the factor's house at Bou Ficha, near Enfida ville, 5 at Henchir Gebeul, and 6 in the smaller basilica on the south side of Carthage, are examples of the larger and more ornate form of the Roman Corinthian style ; 5 and 6 are without the torch handle ornament. The example at Henchir Gebeul is a beautiful specimen of fine carving. The capitals, 7 and 8 at the museum Algiers, and 9 in the public garden at Guelma, are copies of the old Ionic style rarely seen in Africa ; the best known local examples are on the Punic tomb at Dougga. I am unable to give either the date or the provenance of these capitals. The capital at Algiers 7 with the monogram X P is of course Christian, and presumably later than the beginning of the 4th Century. The capitals in the next two groups are designed by native Berber artists ; 10 at Algiers resembles the style of 4, 5 and 6, but has X P inserted in the acanthus foliage ; 11 at Algiers, 12 at Feriana, 12a at Sbeitla ; 13 at the Algiers Museum and 14 in the great Mosque at Kairouan represent palm trees ; 15 in the public garden at Guelma has wheat ears ; 16 at the Algiers Museum has fish at the corners, a peculiar design I do not recollect to have seen elsewhere ; 17 at the Town Hall, Kairouan, has a little bust of Satan with horns on his head and panels containing arrow heads ; 17a is at Guelma. The next examples are Byzantine ; 18 is the basket pattern, 19 chevrons, 20 the blown acanthus, 21 cornucopia, 22 a lobed shaped cap, and 23 a peculiar looking pattern in a basket work border, possibly a conventional palm tree ; 18 to 22 are found in Constantinople, and 23 at S. Vitale at Ravenna ; 23a is somewhat like the last ; all these are in the great Mosque at Kairouan. The provenance of these capitals is unknown, but believed to be from Byzantine buildings situated on the coast of the Gulf of Hammamet. The following caps are all decorated with eagles at the corners : 24 in the Bardo Museum at Tunis, 25 lying in the temples at 124 CAEVED CAPITALS Sbeitla, 26 in the facade of the Mosque of the Three Gates at Kairouan, 27 in the Mosque of the Barber also at Kairouan. The group of pillars 28 and 29, with Eoman-Corinthian caps, stand in the ruins of Thelepta. They are made of soft stone, much weather worn, and show the use of consoles or brackets for supporting the roof of a building.^ They are prettily carved in the Eoman style, but inferior in point of execution to some consoles lying in the ruins of Henchir Gebeul, the Eoman city mentioned above. The consoles 30, 31 and 32 in the basilica at Tebessa probably date from the later part of the 5th Century ; the ornamentation is copied from Eoman designs more freely treated but less finely executed than the originals. The consoles 33 and 34 are in a building in the N.W. corner of the City of Sbeitla. Console 35, a roughly executed design of peacocks drinking from a cup, is lying in the baptistry adjoining the Byzantine basilica at Sbeitla. 36, a console obtained at Thelepta in the place d'armes at Feriana, is another example of crude work in very low relief, more suitable to wood than to stone carving. Beside the two peacocks drinking from a cup there are a dove, a fish, a lion, and a stag, and on the extremity is an eagle holding a wreath in each claw. The carved stones 37 and 38 are in the courtyard of the Museum at Algiers and at Feriana respectively. The following from Thelepta are now in the place d'armes at Feriana : 39 is the lower part of the shaft of a pillar probably connected with the chancel screen of a church. The photograph does not show the sides grooved to take the panels of a screen. These grooved shafts are frequently met with on the sites of ruined churches, as for instance in the detached baptistry and the Byzantine basilica at Sbeitla, and the basilica on the south side of Carthage. 40, the shaft of a pillar. 41 and 42, fragments of other decorative work are at Feriana. 43 is a Byzantine carving of peacocks drinking from a vase, with a Greek incription on the wall of the great Mosque at Sfax. 44, a Byzantine cap in the wall of another mosque also at Sfax. 1. There is no evidence that it was a church. 70. Marble panel over the door of the Mosque. Inscription^ GYOPASIANIS • CYNAYTH CYcDPOCYNHN • TAC KOCMOYCAC TON AeroN cenTON coyaomoy Cross carved on a marble pillar from Ruspe\ now in the French church at Sfax. Terra cotta plaques of the 6th century from the district South of Kairouan, tiow in the church of the Wisdom of God, Lower Kingswood, Surrey. < Representing the Byzantine Imperial Eagle, Adam and Eve, and S Theodore on horse- back spearing a crocodile. To face page 124. .0^ the US* uA- :i ^AOYOMOO)! ^ATi>tHWy5oq\ oT^ ^Nf- h-.^>^^>^^>^ ^^-^^ ii-u»uV' •iwv* y.v v't .1.^ VI. IT i.»^ 'i;jw\. <^ 71 CONSOLES AND PILLARS. JJ and J4. From a churchy in the N. W. quarter of Sheiila. This church is described in Vol. it. JS- Console in the baptistery of the basilica of Bishop Bellator at Sbeitla. This church is described in Vol. it. j6. Two views of the front and side of a Console now in the Public Square at Fcriana. j8. Fragment in the Square at Feriana. jg. Lower part of the shaft of a pillar probably for a pergula or screen; now in the Square at Feriana. 40. Shaft of a pillar in the Square at Feriana. No. J7 is omitted. Some of the other numbers not illustrated here will be found in Vol. ii. ■mnii^m^ To face page 125. DEDICATION CROSSES 126 45, sketch of a cross carved in relief on a white marble pillar, surmounted by a Corinthian cap, now in a side chapel of the French church at Sfax, used as a pedestal for a modern statue of the Blessed Virgin. This pillar and the cross on it are interesting, because the cross on an orb, like this one, occurs frequently in S. Sophia at Constantinople, and because of the alleged provenance of the pillar from Ruspe (the see of Bishop Fulgentius, exiled during the Vandal dominion) . From information supplied to me by one who had assisted in transporting the pillar, I came to the conclusion that it had been brought from Sidi Maklouf, the site of the ancient Inch ilia, 31 kilometres north of Sfax, of which nothing remains but a ruined mosque and a large Zaouia. In the courtyard of the latter I found seven white marble pillars, similar to that at Sfax, and a few debris of no importance obtained from the ruined mosque, considered by M. Guerin to have perhaps replaced a Christian church.^ Ruspe has usually been located some distance further up the coast and much nearer to Ras Capoudia, the Caput Vada of Procopius. I should feel more disposed to fix it at El Rosfeh (the name itself suggests Ruspe), about 4 kilometres north of Sidi Maklouf, where the remains are more important than at any other spot on the coast between Sfax and Capoudia, not excluding Ras Boutria, a large and important Roman city of which nothing but an amphitheatre remains. It frequently happens that Arab names corrupted from the Roman afford quite good evidence of an ancient site, as for instance in the case of Capoudia, and El Rosfeh may well be the corruption of Ruspe. At El Rosfeh there are remains of a Byzantine block house, a great mound, and a confused mass of debris, including a bath or a baptistry with lobed recesses like those found in the Christian basilica of Uppena and other sites in the Regency. Of the Byzantine city called Justinianopolis, founded after Belisarius' conquest near the Ras Capoudia, nothing remains. 46 and 47 crosses on the wall of the house at Bou Ficha, occupied by the factor of the Enfida domain, are both Byzantine. 48 is a cross on a keystone in the basilica at El Kef. 49, the cross on this stone at Sbeitla, now lying in the detached baptistry, is similar to those on Armenian tombstones at Nozal, a village on 1. Voyage dam la Begenee de Tunis, vol. I., p. 153. 126 UPPENA MOSAICS the road over the Mamisson Pass, on the north slope of the Caucasus, 50,^ and to two in Guildford Castle Chapel. 51, a series of mosaic tombs discovered in the basilicas at Uppena and Sidi Abich are now on the walls of the French church at Enfidaville. Of all these ruined Christian churches by far the most interesting and important for the study of early Christian and Byzantine archaelogy is at Uppena, where two basilicas were found built on the same site : an earlier basilica built before the Vandal occupation, probably in the end of the 4th or beginning of the 5th Century, contained a mosaic tomb with an inscription recording the names of a number of Christian martyrs of the early persecutions, and a later basilica built after the Byzantine conquest contained a mosaic inscription in which the earlier inscription was repeated and decorated with an ornamental border and cross in the Byzantine style. It would seem that the earlier church had been destroyed during the Vandal dominion and rebuilt by the Greeks on a larger scale partly on the old site and with marble pillars from a Eoman building. The font of the earlier church was a plain square pool, while that of the Byzantine church, built on a higher level, was ornamented with the lobed recesses frequently found in fonts of this period all over the Regency. The fonts referred to on p. 106 are 52, at Sidi Abich, 53, two fonts at Uppena, 54 at Sbeitla, and 55 in the basilica on the south side of Carthage. The mosaic of the priest is in the museum at Sfax. It came from the ancient Thinna near by. The tiles illustrated on the plate opposite represent 1, Adam and Eve ; 2, S. Theodore and the serpent ; and 3, the Byzantine eagle. The first comes from the collection of the Bardo Museum at Tunis, and was presented to me by the Director of the Depart- ment of Antiquities. The second and third were presented to me by the ex-Mayor of Kairouan. There is a large collection of over fifty of these tiles in the museum, including a variety of 1. Upon the different forms of cross and monogfram X P see article l)y M. Duprat, in the publication of the ArchfEological Society of Constantine, vol. IX. of the 3rd Series, vol. XXX. of the Collection, 1897; also Gsell, vol. II., p. 115. 72. I MOSAICS. No. SI- Tombs from Uppena and Sidi Abich now in the modern church at Enjidavillc ; and the tomb of a priest.^ from Thmna, now in the Museum at Sfax. To face page 126. .£T A!. .rt.i#'i-MK f»\vA :^\ii^*i i*^<^^\*\ wo Vi'iXvi'^iNiwWv z\ %\ .oV^ ^_ \ >'i\v\ >,^\u o\ »-.\vY 73. C/^OSSES. The ntimhers refer to pages 125 and 126. lOja^^.jT^^mi^'fm^'^jmA^^j^'. -,^^-.- \^">>^ 46 50 49 SO. These photographs came into my hands by accident ; I found them, loose in a second-hand book purchased in Paris^ and the board they are gummed to is inscribed Nazal. The ornaments show that they all belong to Circassian Christians. They are illustrated here to shoiv the head crosses. 47 No. 48 is illustrated on Plate 66 to face page 120. To face page I2y. TILE PLAQUES 127 sacred and secular subjects. Among the former are : Adam and Eve, our Lord and the woman of Samaria, S. Peter receivmg the keys, Abraham, Isaac and the ram in the thicket, Daniel (written AANiEj,) in the lions' den, the miracle of the loaves and fishes, S. Theodore and the serpent with the legend 'Serpente,'^ and S. George and the Dragon. By far the most interesting is the representation of the Blessed Virgin seated on a throne.- She holds our Lord, standing between between her knees, the position usually adopted in Byzantine representations of this subject. The surface is too worn to show whether our Lord is blessing in the Greek way, and that unfortunately is the case with every other tile I have seen in which our Lord is represented. Among the secular subjects are : peacocks drinking out of a chalice, an octopus, a bull, a stag tossing a hound, a lion and a palm tree, a man on horseback and various hunting scenes. Some of them are stained with red paint. It appears that these tiles, used to cover the walls of churches, have been found in various places in the centre of Tunis, chiefly in the neighbourhood of Hadjeb el Aioun,^ about 60 miles in the interior west of Sousse. I mentioned finding fragments at El Gebioui, and there are a number from the district round about it hanging up in the office of the Controlleur Civil at Kairouan.^ The conclusion arrived at seems to be that they date from the 5th or 6th centuries, just before the Byzantine occupation. Greek letters appear, as for instance in the D and L in the name Daniel referred to above, but the inscriptions are always in Latin. Many more of these tiles will no doubt be found, but in this part of Tunis, with one or two exceptions the ancient Christian sites have scarcely been explored, and they offer, as I hope to show hereafter, a wide and interesting field for future archaeological research. 1. Bulletin Archeologique, November, 1909. Eeport by M. Merlin. Plates XIV., XV. and XVI. on pp. 149 and 150. 2. Missions Scientifiques tome XV., 4th fascicule, p. 548. Article by M. Gauckler, illustrated. 3. I have some doubt as to the correctness of the spelling of this name, but give it here as it appears in the map and on the railway. 4. They came from a place 30 kilometres south of Kairouan, 2 kilometres north-east of Henchir Trade. This would be about 12 kilometres north of El Gebioui. 128 NOTES ON SOME COINS.' As Heraclius, Constans II., and Constantine Pogonatus are frequently referred to in these pages, and as they were all closely connected with Africa and Sicily, I have reproduced some photographs of their coins in our collection.'^ Six Emperors are represented in the first group, their combined reigns extending over nearly the whole of the seventh century. They are, Heraclius : his son Heraclius Constantinus who only reigned for a few months and does not appear to have struck any coins : his grandson Constans II. (Constantine III.) : his great grandson Constantine Pogonatus (Constantine IV.) : and his great great grandson Justinian II. The events of these reigns are recorded by Professor Bury in the Later Roman Empire, Vol. II., Book V. According to Mr. Wroth, the coin No. 1, of Heraclius, c. 613, may be a portrait. About 17 years later the coin No. 2, of Heraclius, was struck, representing him with a long flowing beard. The same large beard also appears on most of the coins of Constans II., including those figured here. Constantine ' Pogonatus,' who should have had a big beard, is usually repre- sented, as he is on this coin, with a short beard and moustache. The name ' Pogonatus ' applied to him by the historians originated probably from a mistake as to the Sovereign represented. The historians probably knew Constans' coin with a long beard, and as on the legend he is called Constantine, they perhaps confused the father (Constantine III.) and the son (Constantine IV.). Unless we assume that Heraclius and Constans both had these long beards and that coin 3 is not merely a copy of coin 2, it would seem that Heraclius is the Sovereign who should have been called ' Pogonatus.' Notice the family likeness in the faces of the 2nd, 3rd and 5th coins, and the difference between them and Leo III., No. 11, the founder of the succeeding dynasty. 1. Pages 25 and 77. 2. Supplemented by photographs of a coin in the Museum at Sousse, and of casts of some coins in the British Museum. COINS 120 The coins in the second group are of Constantius the son of Conslantine the Great, Eudocia the wife of Theodosius IT., Leo I., the Great, Leo III., the Isaurian, and Leo YL, the Wise. Constantine IX., Monomachos, reigned when the schism between the Greek and Latin Churches came to a head, and John Komnenos was the contemporary of the Norman Kings of Sicily. The little coin, xvi., is Norman- Sicilian struck in the reign of William II., the builder of the Cuba at Palermo. On the reverse there is an Arabic inscription giving the date 1166-1189. On the obverse a cross with ic xc nika. The student of Byzantine history should not fail to study Mr. Wroth's interesting volumes. I have not in every case attempted in these notes to reproduce the characters of the legends. To Dr. Grueber I am indebted for the casts of coins belonging to the British Museum reproduced here. i. Gold coin. Heraclivs. On the obverse : legend : — DD NN HERACLIUS ET HERAC CONST bust of the sovereign, on the left ; he wears short beard : side hair arranged to curve inwards and not point outwards as on most coins. Smaller bust of youthful Heraclius Constantine, on the right. Both busts facing ; each wears a crown with globus surmounted by a cross : they wear paludamentum and cuirass. Above, a cross. On the reverse : legend : — VICTORIA AVGV a cross potent on three steps with CONOB below, and at the end of the inscrijDtion, Greek D. Date about 613-614. Constantinople mint. This coin may be a portrait; a similar example is in Brit: Mus : Cat : Vol : I., Plate XXIII., No. 5. And described Vol : L, p. 186, No. 17. My wife's collection. 130 COINS ii. Gold coin. Heraclius. On the obverse : legend : — DD NN HERACIvIU.S ET HERA CONST PPAV' bust of the sovereign, on the left ; wears long beard and pointed moustache. Smaller bust of Heraclius Constantine, with short beard, whiskers and moustache. Both busts facing ; each wears low crown with globus surmounted by cross, paludamentum and cuirass. Above in field, a cross. On the reverse : legend : — VICTORIA AVGV a cross potent on three steps with CONOB below, and at the end of the inscription, K. Date about 630. Constantinople mint. This coin appears to be the original or model from which the coins of Constans II. (see below) were copied. It is a better finished coin than the copies. The crowns in Constans' coins are ornamented with feathers or plumes, one on either side of the head. This photograph is taken from a cast made for me by the Brit : Mus: It is in Brit: Mus : Cat: Vol: I., Plate XXIII., No. 9. And described Vol : L, p. 189, No. 41, where the legend is copied.^ Brit : Mus : collection. iii. Gold coin. Constans II. {Constantine III.) On the obverse : legend : — effaced bust of the sovereign, on the left ; wears a long beard and pointed moustache. Smaller bust of Constantine IV., beardless. Both busts facing, each wears low crown with globus surmounted by a cross, and on the sovereign's crown a plume on each side. On the reverse : legend : — VICTORIA A a cross potent on three steps. On the left Heraclius, on COINS 131 the right shorter figure of Tiberius, the younger sons of Constans.^ Each is beardless and stands facing. They wear long robes and crowns with crosses ; they hold in right hands globus crossed. CONOB below. Date 659-668. Constaiitinople mint. The date may be fixed from the fact that Constantino IV., Pogonatus, was created Augustus from 654, and Heraclius and Tiberius were created Caesars from 659. This coin is apparently copied from that of Heraclius (No. ii. above). A similar coin is in Brit: Mus : Cat: Vol: I., Plate XXX., No. 20. And described Vol : I., p. 261, No. 59. My iciJVs collection. 1. For an account of these princes see Later Roman Empire, vol. II. iv. Gold coin. Constans II. {Constantine III.) On the obverse : legend : — . ONCTA bust of the sovereign, on the left; wears a long beard and pointed moustache. Smaller bust of Constantine IV., beardless. Both busts facing, each wears crown with globus crossed. Constans holds globus crossed in his right hand. On the reverse : legend : — VICTO . . lA cross potent on three steps, and CONOB 1 below, and P in the field, right. Date about 654-659. Carthage mint. This little coin is one of many obtained near Tunis, in one ' find ' and purchased by us there. A similar coin is in Brit : Mus : Cat : Vol : I , Plate XXXIL, No. 21. And described in Vol : I., p. 289, No. 272. My wife's collectio)i. 1. Conob, the legend of the Constantinople mint, appears on this coin and on ,\os. V, and VII., though these coins were struck in provincial mints. 132 COINS V. Gold coin. Constantine IV. On the obverse : legend : — {see heloiv)^ bust of the sovereign, full face ; wears short beard and long moustache and hair. Wears armour, a helmet with globus crossed and plume ; in his right hand globus crossed, and in left a shield with knight on horseback. On the reverse : legend : — {see below) ^ a cross potent on three steps, and below CONOB Date about 670-685. Carthage mint. This coin of Constantine ' Pogonatus ' is in Brit : Mus : Cat : Vol : II., Plate XXXVIL, No. 12. And described Vol : II., p. 321, No. 54, where the legends^ are given. On the name Pogonatus see Introductory Note. Photograph from a cast supplied to me by the British Museum. Brit : Mus : collection. vi. Gold coin. Justinian II. On the obverse : legend : — DIUSTINIA NUS PE AV bust of Justinian II., facing, bearded ; wears crown with globus crossed, mantle and robe; in right, globus crossed. On the reverse : legend : — VICTORIA AVGV cross potent on three steps, and below CONOB Date 685-695. Constantinople mint. This emperor was the son of Constantine IV., Pogonatus. The photograph of this coin was taken from a cast supplied to me by the British Museum. It is in Brit : Mus : Cat : Vol : II., Plate XXXVIIL, No. 14. And described in Vol: II., p. 331, No. 10. Brit: Mus: collection. COINS 133 vii. Gold coin. Justinian II. On the obverse : legend : — . . USTI NIANUS PP bust of the sovereign, facing, bearded, long hair; wears crown surmounted by a cross, mantle and robe ; in right hand, globus and cross ; linear border. On the reverse : legend : — VICTORIA AVGV cross potent on three steps, and below CONOB Date 685-695. Carthage mint. This is a very well preserved specimen now in the museum at Sousse. A similar coin is in Brit : Mus : Cat : Vol : II., Plate XXXIX., No. 6. And described in Vol : II., p. 337, No 35. Sousse Museum. viii. Gold coin. Constantius. On the obverse : legend : — DN CONSTAN TIVS PE AVG bust of the sovereign, side face ; a diadem of two rows of stones with medallion over forehead, knotted at the back of head with two pendant pearls. On the reverse : legend : — GLORIA REI PVBLICAE two seated figures holding a medallion with, VOT xxxx and below, anti iDate about 360. AntiocJi mint. Constantius was b. 317, Caesar in 326, d. 361. One of the sons and co-heirs of the great Constantine, survived his brothers Constantino and Constans, and eventually became sole ruler. For a time resided at Antioch [360] where this coin was minted. The lamp at Selinunto probably belongs to this period. My wife's collection. 134 COINS ix. Gold coin. Eudocia, wife of l^heodosius II. On the obverse : legend : — AEI, EVDO CIA AAVG bust of the empress, side face ; wears a diadem with three tassels and a cap, hair dressed in curls on the forehead, earrings and mantle. On the reverse : legend : — CONOB a cross in a laurel wreath tied with ribbon and tassels, and a medallion above, a star in the field. Date about 4:'2,2. Constayitinojjle mint. The empress Avas married in 421 and crowned Augusta in the following year. My wife's collection. X. Gold coin. Leo I. On the obverse : legend : — dnlp:op petav bust of the sovereign, full face ; wears a crested crown and a cuirass ; holds in the right hand a javelin, in the left a shield with a knight on horseback. On the reverse : legend : — VICTORI AAV a winged Victory passant left, profile face, holds a long cross in the right hand, a star in the field at her back, and beneath V CONOB Date about 457. Leo reigned from 457 to 474. He was named the Great and the ' butcher.' During his reign the western line of Roman emperors came to an end, and the administration of Italy passed into the hands of the German chieftain Odoacer. The great event of his reign was the unsuccessful expedition he organised against the Yandals in North Africa. COINS 135 The bust on this com is not a portrait but only a conventional representation used by several sovereigns from Valentinian III. to Justinian II., including Leo's son-in-law the Emperor Zeno. My ivife's collection. xi. Gold coin. Leo III. On the obverse : legend : — D LEO NPEAV bust of the sovereign, full face ; wears a crown surmounted by a globus and cross. His robe has a lozenge pattern ; he holds a mappa in his upraised right hand, and in his left a globus with a cross on it. A border of dots round the coin on both sides. On the reverse : legend : — VICTORIA AV a cross potent on three steps, below CONOB and s in the field. Date about 715-720. Constantinople mint. Leo reigned in Constantinople from 717 to 741. He was named Isaurian. The principal events in his reign so far as Ital}', Sicily and Africa are concerned, were the confiscation of the Patrimony in Sicily and Calabria, and the union of the Sicilian and Calabrian Churches with the Constantinople Patriarchate. The Christians were expelled from North Africa within two years of his accession. This may be a portrait of Leo ; there is a certain likeness between him and our Henry VIII. ; and by a strange coincidence they treated the Roman See in much the same way. An example of this coin is in the Brit : Mus : Cat : Vol : I., Plate XLIL, No. 7. And described in Vol : II., p. 365, as No. 1. My wife's collection. xii. Gold coin. Leo VI. On the obverse : legend : — LEO BASILEUS ROM^N 1 136 COINS bust of Leo VI., with long beard, facing; wears crown with globus crossed and imperial robes ; in right, globus surmounted by patriarchal cross. On the reverse : legend : — + MARIA + bust of the Virgin, facing, orans ; she wears veil, tunic and mantle ; •:• on drapery ; on left, m^ ; on right, ©v Date, end of 9th century. Constantinople mint. Leo VI., surnamed 'the Wise.' For the full legends^ and date of this coin see Brit : Mus : Cat : Vol : II., p. 444, No. 1. And Plate LL, No. 8. Photograph from a cast supplied to me by the British Museum. Brit: Mus: collection. xiii. Gold coin. Constantine IX. On the obverse : legend : — CON . TANT .NO bust of the sovereign, bearded ; wears a crown surmounted by a cross, a jewelled robe. He holds a labarum in his right hand with five stars, in his left globus with a cross. On the reverse : legend : — IHS . . REXREGNANTIUM bust of the Saviour, facing; wears tunic mantle and nimbus cross. His right hand raised in blessing in the Greek way. His left hand holds a book of the Gospels ornamented with one central and many other pellets. A border of dots on both sides of the coin. Date about 1042-1055. Constantinople mint. Constantine IX. reigned in Constantinople from 1042 to 1055. He was named Monomachos. The principal events in his reign, so far as Sicily and Italy were concerned, related to the schism between the Eastern and the Western Churches in 1054. An example of this coin is in the Brit : Mus : Cat : Vol : L, Plate LVIII., No. 8. And described in Vol : II., p. 500, as No. 5. My wife's collection. COINS 187 xiv. Silver coin. Constantino IX. On the obverse : legend : — h[blAX]e[p NITIJICA bust of the Virgin (Panagia Blachernitissa), facing, orans; wears nimbus, and mantle and veil ornamented with four pellets in front crosswise. In the field ivi^r ©v ; two linear borders. On the reverse : legend : — ©KER,0, three lines above these letters partly effaced KfiNCTAN TINf^AEC nOTHTH MONOMA and lines and dots below these letters partly effaced ; two linear borders. A similar com is in Brit: Mus : Cat: Vol: 11. , Plate LIX., No. 5. And described in Vol : II., p. 503, as No. 18. My wife's collection. XV. Gold coin. Jolin II., Komnenos. On the obverse : legend : — illeg : and in the field mp ©v two full length figures of the emperor (on the left) and the Blessed Virgin. The latter holds her left hand up in blessing. The right hand is extended over the side of the emperor's head to indicate the act of coronation. The emperor is dressed in a long mantle, has a diadem on his head, a scroll in his left hand ; the object in the right hand is effaced. On the reverse : legend : — ic xc Our Lord, wearing tunic, mantle and nimbus crossed, is seated on a throne ; holds a book of the Gospels on His knee with one hand, and the other (effaced) is extended in blessing. Date about 1118-1143. 138 COINS This emperor was the son of Alexius Komnenos and the brother of Anna Komnena ; b. 1088, cr: 1118, d. about 1143. This coin was sent during the Eusso-Turkish war in a bag full of other coins from Constantinople to London to be melted down as part of the ^^ar loan. My father purchased it from the Bank of England. An example of this coin is in the Brit : Mus : Cat : in Vol : I., Plate LXYII., No. 9 And described in Vol : K., p. 559. My wife's collection. xvi. Gold coin. Williaiii II., of Sicily. On the obverse : legend : — IC XC NIKA and a cross. On the reverse: Arabic characters with the date. Date about 11(56-1189. These little coins were struck by the Norman rulers of Sicily at a time when, in some parts of Sicily at any rate, the population was chiefly Saracen and spoke Arabic. See the passage in the Charter of Girgenti on p. 42 above, in the first and second lines. My wife's collection. INDEX INDEX. ABATING, Sig: (Plates) 95, 96. Africa. 19, 54, 68, 78, 103 to 128. Agia Kyriaki, 80, 81, 93, 99. Aglra or Argiro, 20. Agro di Cabras, 67. Aguglia, 2. Ain Toimga, 106, 121. Aix en Provence, 68. Ajaccio, 57, 68. Alaesa, 45. Alarir, 19, 20. Ales, 52. Alexandria, 19, 80, 101. Alghero, 52. Algiers, 103, 117, 123, 124. Anialafrida, 21. Anialasuntha, 21, 22. Arnalfi, 53, 54. Amantea, 78, 80. Ampurias, 52. Angevines. The, 84. Annovina, 117, 121. Antioch, 19, 26, 80, 101, 103. Apse, triple, use of, 31, 92, 101, 115, 116. Apulia, 77, 78, 82, 83. Arabia, 111. Arabs, see Saracens. Aranci Bav, 56. Arborea. 52, 63. Archbishoprics — in Sardinia, 80. in .Sicily, 1, 25, 29, 30. Ardfert, 23. Arians, 17, 18, 21, 51. Aries, 17, 119. Armenian patriarch, 30. chapels, 119. tombs. 125. Asia Minor, 63, 66, 114. Asproinonte, 81. Assemini, 51, 54, 55, 67, 71, 73, 74. Athalaric, 21. BAGHDAD, 36. Baghoria, 36. Balearic Islands, 51, 75. Bari. 78, 93, 94. Basil, juonks of Saint, 26, 37, 81, 82. 83, 85, 86, 98. Bastia, 57. Battifol, M.. 75, 79, 86, 101. Bedouin, The, 110, 111. Belcastro, 80. Belice, 41- Belisarius, 16, 20, 22, 23, 24. 27, 51, 52, 103, 125. Berbers, 106, 108, 115, 117, 123. Besancon, 41. Bin bir Kilessi, 114, Bisarchio. 52. Bishops. Calabiian, 79, 80, 81, 82, 83, 84. Norman and French in Sicily, 26 27, 42. Sardinian, 52, 68. Sicilian, 1, 17, 18, 25, 26, 27, 28, 29, .30, 41, 42, 80. Agatho, of Rome, 26. Albert, of Oirgenti, 42. Autrustine of Hippo. 17 Auxentius, 18. Basil, of Ijipari, 80. Caius, 18. Conon, of Borne, 26. Constantine, of Lontini, 87. Constantine, Pat. of Antioch, 26. Crestus, of Syracuse, 17. Deodatus, of Cagliari, 52. Drogo, of Girgenti, 42. Fulgentius, of Ruspe, 19, 51, 68, 70, 125. Gaudioso, of Messina, 80. Gelasius, of Rome. 21. Gentilis, of Girgenti, 42. George, of Syracuse, 25. Germanos, of Syracuse, IS. Girlandiis. of Girgenti. 41. Gregory Asbesta, 26. Gregory the Great, of Rome, 24, 35, 77. Gregory VII., of Rome. 52. 55. Gregory, of Girgenti, 25. Hadrian I., of Rome, 80. Humbert, of Sicily, 30. .lohn VIIT., of Rome, 53, 55, 80. John, of Taormina, 80. .lohn. of Tiiocala, 80. Leo II., of Rome. 26. I-eo TX., of Rome, 30. Martin, of Rome, 25, 79. Melo, Patriarch of Constanti- nople, 26. Nicholas I., of Rome, 80. Nicodemus, Greek Ajchbishop of Palermo, 30. 11. INDEX Bishops — Photius, Patriarch of Constanti- nople, 28, 80- Raymond, of Giry:enti, 42. Rufinianus, of Africa, 1, IP. Sergius, of Rome, 26. Stephen, of Syracuse, 80. Stephen IV., of Rome, 26. Theodore, of Catania, 80. Theodore, of Palermo, 80. Theofanes, Patriarch of Antioch, 26. Theofano, of r.ilybeo, 80. Theodosius, of Syracuse, 25. Urban II., of Rome, 25, 41. Ursus, of Girt^enti, 42. Valentine, of Sardinia, 52. Walter of the Mill, of Palermo, 30, 36, 43. Warin, of Girsrenti, 42. Zosinms, of Syracuse, 16. 25. Bisignano, 81, 82. Bohemond. 83. Bonagia, 39. Bonifacio, 56. Bosa. 52. Bou Ficha, 123, 125. Bova, 82, 83, 84. Bricia, 23. Brutii, 78. Burgos, 39. Bury. Professor, 17, 23, 24, 29, 128. Butera, 4 I.- Byzantine Architecture and Art in Africa. 106, 108, 110, 116, 117, 118, 119, 121, 122, 123, 124, 125, 126, 127. Architecture and Art in Calabria, 77, 86, 90, 91, 95, 98, 99, 100. Architecture and Art in Sardinia, 50, 51, 55, 66. 68, 69. 71. Architecture and Art in Sicily, 14, 15. 22, 23, 24, 27, 31, 32, 3.3, 36, 37, 38, 39, 42, 43. 48. 49. Conquest of Africa, 103, 110, 111, 116, 117, 126, 127. Conquest of Sardinia, 52. Conquest of Sicily, 22, 23. Emperors in relation to Calabria, 77, 78, 79. 80, 81, 82. 83. Emperors in relation to Sardinia, 52, 53, .54, 74, 75. Emperors in relation to Sicily, 21, 25, 26, 28, 29. Occupation, provinces, and c-overnment of Africa, 116, 117. Occupation, provinces and government of Calabria, 77, 78, 79, 82, 94. Byzantine — Occupation, provinces and irovernment of Sardinia, 51, 52, 53, 54, 74, 75. Occupation. provinces and government of Sicily, 24, 25, 45. Princes : Heraclius Constantine, 128, 129, 130; Heraclius and Tiberius, sons of Constans II., 130, 131. Titles, 74, 75, 79, 84. CABRAS, 63, 67. Caoliari, 17, 32, 50, 52, 55, 56, 60, 67, 68, 70, 71, 72, 73, 75, 121. Cairo, 30. Calabria, 18, 27, 29, .32, 51, 76 to 102, 121. Access to, 84. Bishops in, 79, 80, 81, 82, 83, 84, 92. Byzantine administration of, 77, 78. The Church in, 51, 78, 79, 80, 81, 82, 83, 84, Definition of, 78. Duchy of, 78, 82. Lituruy in, 78, 86, 100, 101. Monks in, 28, 29, 81, 82, 85, 87, 87. Greeks in, 77, 81, 82, 83, 84, 86, 94, 95. Normans in, 77, 82, 83, 92, 94, 95. Saracens in, 86, 87. Calaris, 68. Caltabellotta, 42, 44. Camarana, 23. Camardi, 86. Camastra, 45. Camerina. Santa Croce in, 1. 5, 6, 7, 8, 15, 18, 20, 22, 69, 71, 72, 103, 104, 118, 119. Canada, Greek rite in, 27. Capella Palatina, Palermo, 36. Capo Bianco, 43. Capo Rizzutto, 81, 82. Caput Vada, 125. Cargese. 27. Carini, 45, 46. Carra, 86. Carthage, 19, 51, 103, 107, 108, 109. 121, 123, 124, 126, 131, 132, 133. Casale Catta, 41. Casar Giafr, 34. Cassano, 82. Cassihile, 12. Cassiodorus, 81, 85, 87. INDEX 111. Castello di Maniace, 43. Castello di Mare Dolce, 33. Castelvetrano, 11, 14, 37. Castigliono, 8. Castrogiovanni, 81. Castronuovo, 42. Catania, 12, 49. Catanzaro, 8o, 87, 88. Cattolica, la Stilo, 95, 97, 98, 100. Caucana, 22, 23. Caucasus, The, 126. Cefala, 41. Cefalu, 12, 45, 49, 64. Cerami, 47, Cerenzia, 82. Charters of Girgenti, Norman, 40, 41, 42, 138. Charters of Girgenti, Saracen, 31. Chiaranionti Family, 43, Chilivani, 56, 60. Cicero, 11. Clarinval, Commandant, 104, 105, 106. . Coins, see page 129. Comnmnion vessels, of glass, 86. Con.stantinople, 15, 20, .53, 69, 71, 77, 80, 82, 83, 84, 91, 92, 98, 100, 112, 115, 116, 121, 123, 125, 129, 130, 131, 132, 134, 135, 136, 138. Coregliano,. 98, Corleone, 41, Corsairs, 87. Corsica, 27. 53, 54, 57, 68. Cosenza, 81, 85, 88. Cotrone (Croton), 77, 79, 80, 81, 82, 83, 87. Councils, Aries, 17. Constantinople, 6th (Ecumeni- cal, 79 : Svnod, 80. Lateran (64 9\ 25, 52, 79. Nicea (325), 18. 7th (Ecumenical, 80. Rimini, 18. Eome, 24, 79, SO. Sardica, 18. Trent, 26. Crati, 78. Crosses, dedication and other carved, Assemini, S. Peter, 74. Bou Ficha, 125. Cagliari, S. Saturnino, 71. ElKef, 125. Feriana, 124. Girgenti, S. Biagio, 43. Guildford Castle, Surrev, 126. On Capitals of Pillars, 123, 124, 125, 126. Porto Torres, S. Gavino, 58. Crosses — Priolo, 4. Sfax, 125. Stilo, 96. Of wood in Churches at Maz- zara and Sciacca, 44. Cuba, Palermo, 34, 35. Cuba, near Syracuse, 1, 12, 13, 14, 15, 24. Cubola, Palermo, 35, 36. DALMATIA, 14, 64. Damns el Karita, 104, 107, 108. Decimoputzu, 71. Delattre, Revd. Pere, 107. Delia, Chapel of the Trinity, 11, 14, 37, 38. Demone, vol. 29. Deodatus, Bishop of Cagliari, 52, De Rossi, Professor. 108. Diehl, M., 109. Dioceses in Sicily, 1. Dioceses in Sardinia, 52. Dioceses in Calabria, 80, 81, see also Bishops. Domes, conical, 62. Domes on pendentives, 13, 14, 97. Domes on sqiiinches, 6, 7, 8, 9, 10, 11, 12, 14, 32, 35, 37, 38, 39, 69, 70, 72, 118, 119. Domes, Pepper-pot shape, 95, 96, 97, 100. Donatists, 17, 21. Dougga, 117, 121, 123. Dsimmi. 29. Duprat, M., 104, 106. EGYPT, 19. Eleanora of Arborea, 63. El Djem, 111, 122. El Gebioui, see Gebioui, Elias, of Syracuse, 25, El Kef, 66,' 90, 116, 117, 120, 123, 122, 125, Emperors, see also List of Coins pp, 128 to 138. Alexius Komnenos, 91, 138. Ba.sil, the Macedonian, 75. Charles V. of Spain, 85. Constantine the Great, 15, 17, 23, 133. Constans I., 133. Constans II. or Constantine III., 16, 24, 25. 26, 27, 52, 53, 78, 79, 81, 128, 130, 131. Constantine II., 133. Constantine IV., or Pogonatvis, 24, .53, 79, 81, 128, 131, 32. Constantine IX. Monomachos, 82, 84. 136, 137. Constantius, 18, 129, 133. Eudocia, 134. Frederick, 42, IV. INDEX Emperors — Gordian, 123. Helena, 23. Henry V., 23. Heraclius, 24, 128, 129, 130, 131. Irene, 80, 81. Isaac Konmenos, 91. John Konmenos, 91, 137, 138. Justinian, 15, 19, 21, 22, 24, 52, 78, 116, 117. Justinian II., 128, 132, 133, 135. Leo I., 134, 135. Leo III., 25, 26, 28, 77, 78, 79, 82, 129, 135. Leo VI., 28, 77, 80, 81. Nicephoros Phocas, 54, 74, 75. Otto, 11, 95. Valentinian III., 135. William II. of Sicilv, 138. Zeno, 135. Enlidavillo, 123, 125, 126. Ephesus, 14, 19. Epiphanios, a deacon, 80. Eremiti, Palermo, 35. FAVARA, Castle near Palermo, 31, 33, 34, 35, 36, 100. Favara, Castle near Girgenti, 43. Feriana, 121, 123, 124. Fonts, Baptismal, 104, 106, 126. Francavilla, 9. Fulgontius, S., 19, 51, 68, 70, 125. GABES, 110. Gaeta, 99, 100. Galatone, a priest, 80. Gallura, 52, 56. Galtelli-Nuoro, 52. Garufi, Professor, 27, 40. Gauckler, M., 127. Gaul, 19, 27, 92. Gavino, ch. of St., 50, 57, 60, 67. Gebioui, Chapel of El, 104, 105, 110, 111, 112, 113, 114, 117, 118, 119. 127. Gelimer, 21, 103, 116, Genoa, 51, 73. Genseric, 19, 20, 21, 103, 116. Gesso, 49. Gerace, 77, 80, 81, 82, 88, 92, 93, 94, 100. Giafar, Emir, 34. Girgenti, 1, 25, 40, 41, 42, 43, 44. Goths in Sicily, 21, 22. Grantmesnil, William of, 83. Greeks in Calabria, 77, 81. 82, 83, 84, 86, 94, 95. In Sardinia, .50, 51, 52, 53. 54, 55, 59, 60, 69, 71, 73, 74. 75. In Sicily, 21, 24, 25. Greeks — Rite in Africa, 115. 116. Rite in Calabria, 27, 77, 78, 87, 92, 95. Rite in Corsica, 27. Rite in Canada, 27. Rite in Sicily, 24, 25, 26, 27. Gregorovius, 16. Gsell, M., 104, 105, 106. Guelma, 123. Guerin, M., 125. Guiamar, 74. Guildford, Crosses in the Castle at, 126. Guiscard, 31, 34, 82, 83, 84, 94. Gunda,mund, 21. HADJEB EL AIOUN. 113, 127. Hadjla, 104, 105, 108, 110, 111. Haidra, 117, 121, 122. Hanunamet, Gulf of, 123. Haouch Khima, 104, 114, 121. Haouch Taacha. 114. Hauteville, The family of, 83, 95. Henchir Gebeul, 121, 123, 124. Henchir Maatria, see Maatria. Henchir Trade, 127. Heraclea Cattolica, 44. Heraclea Minoa, 43, 44. Herzegovina, 6. Hierapolis, 66. Hilderic, 21. Hipparis, 23. Hunneric, 21, 103. Hybla, 2. ICHANA, 10. Iglesias, 52, 71. Images, controversy respecting the, 26, 28, 29, 80. Ispahan, 33. JATO, 41. Jerusalem, 23, 80. .Judges of Sardinia, 52, 55, 75. .Justinian, see Emperors. Justinianopolis, 125. KAIROUAN, 33, 104, 110, 113, 114, 115, 123, 124, 126, 127. Kasserine, 114, 121. Koubba, at Sousse, 119. Ksar el Beo, see Gebioul. Kings, Chieftains and Princes. Alaric, 19. Athalaric, 21. Bohemond, 83. Gelimer, 21, 103, 117. Genseric, 19, 20, 21, 103, 117. Guiamar, 74. Gundamund, 21. INDEX V. Kings, Chieftains and Princes. Henrv VIII., 135. Hilderic, 21. Hunnerin, 21, 103. Joseph Bonaparte, of Naples, 98. Mahomet II., the Sultan, 29. Odoacer, 134. Roger, see Roger. Tancred, 42. Theodoric, 21, 22, 84. Trasamond, 21, 51, 68, 106. Kings wood, Tiles in the Church at Lower, Surrey, 113. Kyriaki, see Agia Kyriaki. LA CAVA, 75. Lagoiiegro. 99. La Marmora, 63, 66. Laodicea, 66. Lateran Council, see Coimcils. Lecce, 99, 100. Leghorn, 57, 87. Lentini, 1, 25. Lerida, 63. Levant, The, 29, 78. Lilybeo, 1, 25, 40, 80. Limbara, Mts., 56. IJpari, 1, 25, 45. Liturgies, 26, 101, 116. Locris, 77, 79, 81, 93, 94. Lombardy, 77, 78, 81. Longarino, 12. Liiri, S., 67. MAATRIA, 104, 108, 109, 110, 112, 114, 116, 117, 118, 119. Maccari, 1, 8. 9, 10, 11, 12, 15, 24, 38, 103, 118. Machara , 1 1 . Macomer, 51, 60, 61, 62, 75. Magliotis, 86. Masrnisi, 1. Mahomet II., the Sultan, 29. Mainz, 78. Malta. 1. Malvagna, 8, 9, 10, 15, 24, 103, 118. Maniace, 43. Maracalagonis. 72. Marsala, 20, 40. Martorana, 38. Matera, 95. Matifoii, 58. Mazzara, 28, 20, 38, 39, 41, 44. Melfi, 19, 24. Mellili, 1. Merlin, M., 113, 127. Messina, 1, 2, 5, 28, 31, 36, 49, 81, 85, 101. Mio;rations of Christians, 19, 20, 29, 78, 81, 84, 103, 115. NAPLES, 53, 54, 85, 98. Nea Taktika, the treatise, 116. Nestorians. the. 16. Nicastro, 81, 82, 86. Nicea, 18, 80. Nicephoros Phocas, the Emperor, 54, 74, 75. Nicephoros Phocas, the general, 81. Nicotera, 80, 81. Nona, in Dalmatia, 14. NORMAN Bishops in Calabria, g9 go QA Bishops 'in Sicily, 26, 40, 41, 42. Coins in Sicily, 129. Conquest of Calabria, 77, 81, 94. Conquest of Sicily, 26, 27, 28. Conquest of Palermo, 29. Noto, 1, 10, 12, 19, 29. Nozal, Armenia, 125. Nvioro, 61. Nurhagi, 50, 61, 75. OANIS, 23. Odoacer, 20, 21. Oglet el Lebbia, 111. Ogliastra, 52. Oil Presses in Africa, 113. Oppido, 83. Oristano, 51, 52, 61, 63, 67. Orsi, Professor, 5, 8, 13. Osilo, 57. Otranto, 80, 93. 94, 97. Ouled Fargalla Tribe, 111. Ovid, 86. PACHINO, 1, 10, 19. Palermo, 1, 14, 25. 29, 30, 31, 33, 69, 95, 100, 121, 129. Pallazuolo, 19. ! Pantalica, 10. Passaro, 10. Patriarchs. Alexandria, 30. Antioch, 26. Constantinople, 24, 28, 30, 77, 78, 79, 80, 82, Photius, of Constantinople, 28, 80. Photius, of Seleucia, 30. Melo, 26. Patrimony of the Roman See in Africa, 79, 80. In Calabria, 78, 79, 80, 82, 135. In Corsica, 53. In Sardinia, 53. In Sicily, 26, 28, 135. Pausania, Terra Nuova, 52, 56. Peres Blancs, The, 106. Petralia, 41. Phoenicians, in Sicily, 40. Pirri, 41. VI. INDEX Pisa, 50, 55, 56, 68. Platani River, 48. Pliny, 11. Pont de Trajan, Tunis, 108. Porta Augusta, 1. Porto Empedocle, 43. Longobardo, 23. Torres, 50, 56, 57, 60, 94. Vecrhio, Corsica, 56. Pottery Tubes, 111, Pottery Tiles, 113. Priolo,' 1, 2, 15, 18, 22, 67. Procopius, 21, 23, 125. Ptolemy, 63. Queen Amalaf reda, 2 1 . Queen Amalasuntha, 21, 22. Queen Eleanora of Arborea, 63. RANDAZZO, 8, 9. Rasacambro, 23. Ras Boutria, 125. Ras Capoudia. 125. Ravenna, 79, 112, 123. Reggio, 20, 78, 79, 80, 81, 85. Roccella. 88, 93. Roccelletta, 32, 85, 86, 87, 88, 89, 90, 91, 92, 93, 100, 121. Roger, the Count, 25, 26, 31, 34, 40, 41, 46, 47, 83, 94. Roger, the Duke of Calabria, 82, 83. Roger, the King of 8icilv, 34, 35, 95. Romans in Africa, 108, 109, 112, 113, 114, 115, 116, 117, 120, 121, 125. Rome, 1, 19, 20, 24, 25, 26, 52, 70. 78, 79, 94, 101, 113, 121, 122. Romotta, 28. Rosieh, El, 125. Rossano, 77, 78, 79, 81, 82. 83, 95, 96, 97, 98, 101. Rossolini, 19. Ruspe, 51, 125. Saints and Dedications. Agatho, Bishop of Rome, 26. Ambrose of Milan, 116. Anastasius, 46. Andrew, Church at Trani, 99. Angelo, Monasteries, in Sicily, 30, in Calabria, 99. Angel us of Mai da Church, 86. Ann, 47. Antioco, Church at Sulcis, 71, 73. Augustine. Bacchus, Church at Cagliari, 14. Barbara, 75. Basil, 47. Order of Monks (see Basil). Biagio, Church at Girgenti, 42. Calogero, 20, 44. Saints and Dedications — Cataldo, Church at Palermo, 36. Cataido, Church at Jjecce, 99. Catharina, Church at Mazzara, 38, U. Ciriaco, Church at Palermo, 29. Ciriaco, Church in Calabria, 30 (see also S. Chirico). Chirico, Church, 99. Christopher, 46. Clemente, Church at Rome, 115, 121. S. Croce (see Camerina). Cosmo, Church at Cagliari, 68. Croix, Church near Aries, 116. Damiano, Church at Cagliari, 68. Demetrius, 47. Egidio, Church at Maz/ara, 10, 14, "38, 39. Eirene, 121. Elia, of Castrogiovanni, 81. Elia, Spelota, iOl. Fantin, 101. Fulgentius, Bishop of Ruspe, 19, 51, 68, 70, 125. Gavino, Church at Porto Torres, 58, 59, 60, 94. George, 47. George, Church at Triocala, 42. Germano, Monte Cassino, 99, 100. Giacomo, Church at Termini, 47. Giovanni, degli Eremiti, Church at Palermo, 37. Giovanni, del l^epprosi. Church at Palermo, 31, 32, 35, 38. Giovanni, in Sinis, Church, 50, 54, 64, 67, 72, 121. Giovanni, d'Assemini, Church, 51, 54, 72, 73, 74. Giovanni, Syracuse, Church, 1. Giuliano Monte, 14, 38, 39, 40. Gregory, the Great of Rome, 24, 46, 77. Gregory, Asbesta, 26. Gregory, Nazianzene. 47. Gregory, Thaumaturuos, 85, 86, 87, Helena, The Empress, 23. Hilarion, 19, 20. James, liturgy of, 100. .Joachim, 47. John Chrysostom, 26, 47, 92. John, Church at Syracuse, 16. John, Church at Trapani, 14, 39. John, Cliurch at Assemini, 71. John, Church at Lilybeo, 40. John in TruUo, Church at Con- stantinople, 100. •John Baptist, Termini, 46. John Evangelist, Termini, 46. INDEX Vll. Saints and Dedications — Joseph, 47, 121. Joseph, the hymn writer, 26. Joseph, Church at Gaeta, 99. Justa, Ch\irch at Oristano, 63. Lawrence of Scicli, 23. Lawrence of Genoa, 73. Lorenzo of Rome, 115. Lucia, Church at Syracuse, 1. Marcian, Church at Syracuse, 12, 15, 16, 18. Maria dei Greci. Church at Girgenti, 42. Maria of Vicari Monastery, 30. Maria, Church at Monte Cassino, 99. Mark, Church at Rossano, 77, 95, 97, 98, 100. Martin of Rome, 52, 79. Mary of Carra, Church, 86. Mary of Squillace, Church, 86. Nicholas, Church at Sciacca, 44, 46. Nicholas, Church at Constanti- nople, 91. Nicholas, Church at Bari, 93. Nicholas, Church at Rossano, 98. Nicholas, Church at Jjccce, 99. Nil, of Rossano, 81, 98. Orsola, Church at Palermo, 35, 36, 38. Pancras, Church at Taormina, 49. Paul, Apostle, 16. Peter, Church at Assemini, 71, 73, 74. Peter, Church at Otranto, 99. Phillip and James, Chapel at Palermo, 34. Phillip d'Agira, 20, 23. Phillip of Scicli, 23. Phillip Demona, Monastery, 30. Phocas, Church at Priolo, 2, 15, 19, 20, 67. Rufinianus, 19. Sabina, or Sarbana, Chapel in Sardinia, 51, 61, 75. Salvatore del Greci, Messina, 49. Saturnine, Church at Cagliari, 50, 51, 54, 55, 66, 68, 72, 91, 121. Sergius, Church at Constantinople, 14. Simplicius, Church at Terranuova, Sardinia, 56. Sophia, Constantinople, 69, 71, 73, 125. Sophia, Church at Villasor, 71, 73, 74. Thecla, Church at Constantinople, 91, 100. Saints and Dedications — Theodore, 46. Theresa, Chapel near Syracuse, 1, 12, 13, 14, 15,24, 118.' Theresa di Gallura, Sardinia, 56. Theodosia, Church at Constanti- nople, 91. Trinita at Delia, 11, 14, 37, 38. Vitale, Ravenna, 112, 123. Saladin; M., Ill, 114. Salinas, Professor, 17, 40. Salonica, 69, 71. Salso, River, 41. Saracens, Invasion and Occu- pation of Sicily, 24, 25, 26, 27, 28, 29, 40. Position of Christians under, 28, 29, 30, 31. Charters to Christian Monas- tery, 31. Inscriptions : on Coins, 129, 138 ; on the Cuba at Palermo, 36 ; on the Capital of a Pillar, 32. Invasions of Sardinia, 51, 53, 54. Invasions of Calabria, 77, 81. In Africa, 103, 115, 119. Sardinia, pages 50 to 76. Access to, 50. Assemini, 71. Bishoprics in, 52. Byzantine Administration and Conquest of, 50, 51. Byzantine Architecture in, 55. Cagliari, 68. Cathedrals in, 62, 68. Churches in, 50, 51. Judges of the States in, 51, 55, 75. Nurhagi, 75. Pisans in, 50, 55. Porto Torres, 57. Relations with the Roman See, 52, 55. Saracens in, 53. Sarbana, S., 61. Sinis, 64. States of, 52, 75. The Church in, 52, 68. Vandals in, 51. Sassari, 50, 56, 57. Sbeitla, 58, 104, 117, 121, 123, 124, 125, 126. Scano, Dr., 58, 67, 69, 70, 73, 74. Sciacca, 43, 44. Seals, Episcopal, 26. Sebenico, 64. Vlll. INDEX Selinunto, 17, 44, 45. Sennori, 57. Seriziat, Commandant, 104, 105, 106. Severina, Santa, 80, 81, 82, 99. Sfax, 113, 115, 124, 125, 126. Sicily, pages 1 to 49. Bishoprics in, 1, 17, 18, 26, 29, 30, 41, 43, 56, 80. Byzantine Conquest and Ad- ministration of, 22, 23. Byzantine Architecture, 14, 24, 31. Belisarivis in. 22, 23, 24. The Church in, 1, 15 to 27. Constans II. in, 24, 25, 78, 79. Goths in, 21, 22. Liturgies in, 100, 101, 102. Monks in, 19, 20, 29, 30, 31, 34, 77, 81. Normans in, 16, 23, 28, 29 to 36, 40, 41, 42, 44, 47, 48. Saracens in, 24, 25, 26, 27, 28, 29, 30, 31, 34, 40. Vandals in, 20, 21. Siculiana, 43. Sidi Abich, 126. Sidi Amor el Kenani, 110. Sidi Amor bou Hadjla, 103, 110. Sidi Mohammed el Gebioui, 109. Sidi Maklouf, 125. Silanus, 51, 61, 75. Sinis, 5, 32, 50, 54, 55, 63, 67, 121. Sinope, 2. Siponto, 95. Skirra, 110. Sohag, 66. Solunto, 36, 49. Sorso, 57. Sousse, 113, 119, 121, 127, 133. Spaccaforno, 19. Spain, 39, 51, 54, 85. Spartivento, 85, 93. Spatharii, 74. Squillace, 32, 77, 79, 81, 82, 121, 85, 86, 87, 88. Staletti, 77, 85, 86, 87, 90. States of Sardinia, 75. Stilo, 77, 8 , 83, 87, 88, 95, 96, 97, 99. Sulcis, 52, 71, 74. Sultan Mahomet II., 29. Sybaris, 77. Synods, see Councils. Syracuse, 1, 2, 12, 13, 15, 16, 18, 19, 24, 25, 43, 78, 79, 81, Syria, 19. TAORMINA, 1, 25, 49, 80, 81. Taranto, 78, 83. Tauriana, 79, 80, 81. Tebessa, 104, 105, 106, 107, 109, 114, 116, 121. Teboursouk, 104, 108. Tempio, 52. Termini, 41, 45, 48. Terralba, 52. Terra Nuova, 56. Tharros, 63. Thelepta, 106, 112, 114, 121. Theodoric, 21, 22, 84. Thmna, 126. Tibar, 122. Tile Plaques, 113, 137. Torcello, 121. Torchitorios, Sardinian Family, 73, 75. Torres, 52. Totila, 21, 78. Trapani, 39. Trasamond, 21, 51, 68, 106. Triocala, 1, 25, 44. Troina, 1. Tropea, 79, 80, 81, 82. Tunis, 66, 87, 103. Turks, 29, 31, 63, 86, 115. Tusa, 1, 45. Tyche, 1. Tyndaris, 1, 25, 45. Type, The Edict called the, 25, 52, 58, 79. UNIATE, Greek rite, 27. Uppena, 117, 121, 125, 126. VALENTIA, 63. Valentine, a Sardinian Bishop, 52. Val d'Ispica, 19. Val Demone, 29, Val di Mazzara, 29. Val di Noto, 29. Vandals, in Africa, 103, 117. 125, 126, 134. Vandals, in Sardinia, 51, 52. Vandals, in Sicily, 20, 21. Verres, 11. Vicari, 41. Villasor, 71, 73, 74. Vindicari, 10. Vittoria, 19. Voute d' Aretes, Examples of 104 105, 107, 109, 111. ZAGHOUAN, The, 107, 113. Zaouia, 108, 110, 113, 118. Zaragossa, 32. Zeroud, River, 110. Zinneth, 41. Zisa, 34, 35, 36. CELL.l^ TEICHOE^ AND OTHER CHRISTIAN ANTIQUITIES IN THE BYZANTINE PROVINCES OF . SICILY WITH CALABRIA AND NORTH AFRICA INCLUDING SARDINIA. Vol. II. ILLUSTBATED, EDWIN HANSON JRESHFIELD, M.A., F.S^A. Printed Privately, 1918 A few Copies of this Booh may be obtained by Students from the Author through the Printers and Publishers IRijon Si arnolD, 29 poultry, XoiiDon, B.C. TO THE MEMORY OP BASIL HAMMOND, Fellow of Trinity, AND TO ARMINE F. KING, Archdeacon of Tokio, THESE notes ON HISTORY AND ECCLESIOLOGY ARE AFFECTIONATELY DEDICATED ^ Sicily* 1898, 1899, 1905, 1906, 1907, 1908, 1913, 19U. Calabria. 1899, 1905, 1914. Sardinia. 1901, 1909. Algiers and Tunis. 1910, 1911. Egypt and Nubia. 1886, 1912. Gaeta and Monte Cassino. 1912. Albenga and Capri. 1913. PEEFACE In consequence of the war the number of illustrations has been reduced, and those of well-known buildings that can be found in other publications are omitted. The Introduction is, in the main, taken from letters written to my father during our travels. The text of them is retained, and the fact recorded here by way of apology to the reader for the frequent appearance of the personal pronoun and some lack of continuity in the subjects discussed. E. H. F. Juniper Hill, Reigate, 1918. CONTENTS Preface List of Illustrations and Plans Books Consulted Introduction ... Cellae Trichorae Cellae Trichorae Sicily ... South Italy ... North Africa, Introduction Central Tunis East Algeria West Tunis East Tunis Egypt Lower Nubia Khargeh Oasis Index PAGE V vii & viii ix to xii ... 1-18 ... 19 ... 34 ... 49 ... 67 ... 85 ... 102 ... 116 ... 132 ... 142 ... 154 ... 156 ... 167 ... 173 LIST OF ILLUSTEATIONS I'I'ATE FACINa PAGE 1 Altar at S. Theresa near Syracuse ; a tomb and some masons' marks ... ... ... ... ... ... ... ... 20 2 Trinity Chapel; S. Honorat, Cannes 21 Plan 3 the same; interior ... ... ... ... ... ... 22 4 Beliquary of S. Honorat and a chapel in the Cemetery of 8. ,Callixtiis at Bomb ... 25 Plan 5 Chapel in the same cemetery and plan ... ... ... ... 27 6 Chapel of the Crocifisso at S. Germano, Monte Cassino ... 29 7 Sohag 34 (This and the following eight plates are vieivs of the White and Red Monasteries at Sohag, in the Nile Valley) 8 86 9 88 10 40 11 Plan 42 12 ... ... 43 13 : 44 14 46 15 Castiglione, Sicily, Church of S. Domenico 50 16 the same; interior ... ... ... ... ... ... 52 17 Omitted ^ 18 Ageo ; Church of S. Peter and S. Paul . ... 56 19 Omitted 20 Capitals in the crypt of S. Marcian, Syracuse 62 21 Capitals; the same ... ... ... ... ... ... ... 64 22 Gaeta and Capri 74 Plan 23 S. Qermano, Monte Cassino 82 Plan 88 90 92 viii LIST OF ILLUSTKATIONS PLATE FACING PAGE 24 Altar Screens ... ... ••• ••• ••• 86 25 Coptic screen and font ... 26 Mosaic; Tabaeca 27 Baptismal fonts ... 28 Apsidal seats 94 29 Crosses 98 30 Sbeitla; altar, mosaic, and oil presses ... ... 108 31 Haouch Khima Mta Darrouia 110 32 Henchir Gebeul and Ain Tounya Castle 112 33 Tebbssa ; basilica... ... ... ... ... ... ... ... 118 Plan 34 Tebessa; basilica... ... ... 120 35 Early Christian tombstones ... ... ... 130 36 Haidra 134 37 Dougga; basilica and plan 136 38 Dougga; details of the church ... 138 39 Mosaics at Enfida 143 39a same 145 40 Mosaics at Enfida ... 146 41 Basilica at Carthage 150 Plan 42 Basilica at Addendan, Nubia 156 43 Church at Serreh, Nubia 158 44 Church of S. Michael, Kamula; Nile Valley 160 45 Church of S. Theodore, Medinet Habu, and a roof at S. Simeon's Monastery, Assouan 162 46 Church of S. Theodore, Medinet Habu 163 47 Chapel near the Second Cataract; Wady Halfa 164 48 the same, interior; plan ... ... ... ... ... 165 49 Khargbh Oasis ; view and chapel tomb ... ... ... ... 167 Plan 50 KhXrgeh ; circular tombs ... 168 51 Khargeh; Cathedral tomb 169 Plan 52 Khargeh; various tombs 170 53 Khargeh; architectural details 171 BOOKS CONSULTED Abating, Dr.: Publication on Stilo in ' Napoli Nobilissima.' Vol. xii. page 11, February, 1903. Amari : Storia dei Musulmani di Sicilia. Florence, 1854. Antoniades : Ekphrasis tes Agias Sophias. Constantinople, 1908. \ Ballu, a. : Les Buines de Tiingad, 1911. Barreca : Catacombe di S. Giovanni in Siracusa. Syracuse, 1906. Barth : Travels and Discoveries in North and Central Africa. 1857. Battifol: L'Abbaye de Bossano. Paris, 1891. Bertaux : L'Art dans Vltalie Meridionale. Paris, 1904. Besta : La Sardegna Medioevale. 2 Vols. Beber, Palermo, 1908. Biscari: Paterno, Prince of: Viaggio per tutti le antichita della Sicilia, 1817. Brolo : Lancia di : Archbishop of Monreale. Storia della Chiesa in Sicilia. 4 Vols. Palermo, 1880. f Butcher, Mrs. E. L. : Story of the Church of Egypt. 2 Vols. 1897. Butler: Ancient Coptic Churches of Egypt. 2 Vols. 1884. Bury, Professor : Later Bovian Empire and Eastern Boman Empire. f Byzantine Kbsbarch Fund : Publications on the Churches of S. Eirene at Constantinople a7id of the Nativity at Bethlehem, 1911-12. Chalandon : Histoire de la domination Normande en Italie et en Sicile. Paris, 1906. f Chard : Voyage en Perse, Amsterdam. 1735. Clarinval Commander : Article on Tebessa in the Beceuil of the Archceo- logical Society of Constantine, Algeria, 1870. f Clarke, Somers : Christian Antiquities in the Nile Valley, 1912 f Codex Diplomaticus Caetanus, in Tabularium Cassincnse. f Cooks: Guide Book to Egypt and the Sudan. f Cordara, Vincenzo : Notizie di Francavilla. f CowPER, H. S. : The Hill of the Graces, 1897. Books added to vol. 1 list are marked f- X BOOKS CONSULTED Delattre : Bev. Pere. Un Pelerinage aux ruinea de Cartage. Lyons, 1906. f Dickie, Archibald C. : Article in the ' Quarterly Statement ' of the Palestine Exploration Fund for January, 1899. DiEHL : L'Afrique Byzantine. Paris, 1896. DiMAHZO : Dizionario Topographico delta Sicilia. Palermo, 1856. \ Duchesne : Early History of the Church. DuPRAT, M. : Article on Tebessa in the Beceuil of the Archceological Society of Constantine, Algeria, 1897. f Fabio Giordano : Belation of Capri. Edited anonymously , 1906. FiNLAY : Greece under the Bomans. f Foakes-Jackson, F. A.: History of the Christian Church from the earliest tirnes, 1905. Freshfield, Dr. E. : In Archceologia. Vol. xliv., p. 383. f Frothingham, a. L. : The Monuments of Christian Borne, 1908. f Gaetani d'Aragona : Memorie Storiche delta Citta di Gaeta, 1879. Garufi, Prof. C. A. : L'Archivio Capitolare di Girgenti. In Archivio Storico Siciliano, 1903. Also, Per la Storia dei Secoli xi. e xii, in Archivio Storico per la Sicilia Orientate, anno : ix., fasc : i. Also / Capitoli delta Confraternita di Santa Maria di Naupactos 1048-1068 nella B. Cappella Palatina de Palermo. In Bull : Inst : Stor Italiano n. 31. Gay : L'ltalie Meridionale et V empire Byzantin. Paris, 1904. Giovanni, V. di : Castello e la chiesa delta Favara di S. Filippo a mare dolce in Palermo, in Archivio, Storico Siciliano. Palermo, 1897. Grbgorovius : Borne in the Middle Ages. Trans. Hamilton. '^ GsELL, S. : Les monuments antiques de VAlgerie. Paris, 1901. . Gsell, S. : UAlgerie dans Vantiquite. Algiers, 1903. GuERiN, V. : Voyage dans la Begence de Tunis. 2 Vols. Paris, 1862. f Hanson, Captain William, 20th Light Dragoons : Short Journal of a Voyage to Sicily, 1810-1811. London, 1814. HoDGKiN : Italy and her invaders. Holtzinqer, H.: Die Altchristliche Architectur. Stuttgart, 1889. Jackson, T. G. : Dalmatia the Quarnero and Istria. 3 Vols., 1887. Kaestner : De Imperio Constantini III. Leipsic, 1907. \ Kauffman : La decouverte des Sanctuaires de Menas dans le desert de Mareotis. Published by the Societe de Publications Egyptiennes, 4, Bue du General Earle, Alexandria. 1908. Koch, P. : Byzantinische Beamtentitel. Jena, 1913. La Marmora : He de Sardaigne. Turin, 1851. BOOKS CONSULTED xi Lapotre, a. : L'Europe et le Saint Siege (Le Pape Jean VIII). Paris, 1895. f Lemaire, R. : L'Origine tie la basilique Latine. Brussels, 1911. Lenormant : La Grande Grece. Paris, 1884. f Lenthekic : La Provence Maritime. 1880. t Lombard : Constantin V. Paris, 1902. f Marra, F. della : Monte Cassino, Descrizione. 1775. Martroye, F. : L' Occident a Vepoque Byzatitine. Paris, 1904. Mas Latrie : Tresor de Chronologie. f Mason : Persecutions of Diocletian. ■\ Merimee, Prosper : Voyage dans le Midi de la France. 1835. Merlin, A. : Publications of the Department of Antiquities in the Regency of Tunis. Forum et Eglises de Sufetula (Sbeitla) 1912. f Messina e Dintorni : Guida a cura del Municipio. Messina, 1902. f Mileham and Maciver : Churches in Lower Nubia. University of Phila- delphia, 1910. MiLMAN : History of Christianity. f Moris, Henri : L'Abbaye de Lerins. Paris, 1909. Mossheim : Ecclesiastical History. Edited by Bishop Stubbs. Murray : Handbook for Sicily, 1890. Neale, J. M. : History of the Holy Eastern Church. Orsi, Professor P. : Articles on Camerina and Maccari in Byzantinische Zeitschift, 1898-1899. Scavi . . . nel Sudest della Sicilia. Estratto dalle Notizie delgi scavi; anno 1905, fasc. 11 and 12, Borne, 1906. f Pace, Biagio : I Barbari e i Bizantini in Sicilia. Palermo, 1911. Paspates: ByzantincB Meletai, Constantinople, 1877.' Patricolo : Article on Delia Chapel in Archivio Storico Siciliano, N.S. Anno V, Palermo, 1880. PiRRi : Sicilia Sacra. Riedesel, J. H. Von : Travels through Sicily, translated by J. H. Forster. London, 1772. RoDD, Sir Rennell : Princes of Achaia and the Chronicles of Morea, 1890» Rodota : Dell Origine stato e progresso del rito Greco, 1758-1763. f Roman Studies: Journal, Vol. i., 1911. Saint Nom, Abbe de : Voyage Pittoresque Naples et la Sidle, Paris, 1783. f Saladin, H. : Archives des Missions Scientifiques, 3 ime serie, Tome Xfll,, Paris, 1887. Salinas, Professor : O71 the Christian Lamp in the Palermo Museum and on. the walls of Mount Eryx, in Archivio Storico Siciliano, 1883. xii BOOKS CONSULTED f ScAOLiA, P. SisTo : Manuale di Archeologia Cristiana, Rome, 1911. ScANO, Dr. D. : Storia dell Arte in Sardegna, 1907. •f- ScHULZ : Denkmaeler der Kunst des Mittelalters in Unteritalien. Dresden, 1860. ScoBAK : Catalogus Episcoporu,m Ecclesice SyracusancB. Sebiziat, Commander : Article on Tehessa in the Beceuil of the Archceological Society of Constantine, Algeria, 1868. f Seroux d'Agincourt. Histoire de I'art jx^r les monuments, 1823. Sladen : Sicily, 1904. Theodoret : History of the Church. Bohn, 1854. f TosTi, LuiGi : Storia delta Badia di Monte Cassino, 1842. f Toulotte : Proces verbaux d'une double Mission Archeologiqtie aux ruines de la basilique d'Ujypena. Tunis, 1906. f Vento, Raimondo : Gaeta nella Storia, 1911. f Viollet lk Due : Dictionnaire de V Architecture Francaise. Wroth, W. : Imperial Byzantine Coins in the British Museum. INTEODUCTION The purpose of these notes has been explained in the Introduc- tion to the preceding volume. For the sake of continuity the original title is retained, though Sardinia now falls out of the narrative and Egypt takes its place. I may again conveniently commence with the cellee trichorge, and present an illustrated description of the chapel of the Holy Trinity belonging to the monastery of Lerins, on the island of S. Honorat, opposite Cannes, the chapels in the cemetery of S. Callixtus at Eome, and the trefoil sanctuaries of the churches in the two monasteries at Sohag, in Egypt. The last are by far the most important and interesting buildings described in these pages, not excepting the basilica at Tebessa. It is said that the monastery of Lerins was founded in the beginning of the fifth century upon the model of Egyptian religious houses in the Nile valley, and likely enough, therefore, that the trefoil chapel of the Holy Trinity was copied from these sanctuaries at Sohag. The same might be said of the chapel of Sta Theresa by Syracuse, described in the preceding volume, for it is situated close to the route taken by the early missionaries on their way from the East to Gaul. S. Theresa has recently been excavated, and a close inspection of it shows that it is almost exactly like the Trinity chapel of S. Honorat in architectural details and in plan. I conjecture that the ancient church of S. Giovanni in Sinis, on the West coast of Sardinia, represents another link in the chain of Christian stations between Syria and Gaul. The church, or rather that part of it that is covered by a dome, built like those of S. Theresa and Lerins, would then be considerably older than the date previously assigned to it. It is also note- worthy that the monastery at Lerins owned an estate about Saccargia, in the north of Sardinia. 2 INTEODUCTION From the cellae trichorse I pass on to the Byzantine and early Norman churches in Sicily and Italy. The church of S. Domenico at Castiglione in Sicily is of particular interest in connection with the cellae trichorae, for it is the first example I have come across and, so far as I know, the only one in Sicily or Italy where the trefoil plan was adopted in the construction of a sanctuary intended for the Greek liturgy. The prothesis and the diaconicon are nothing more than niches in the north and south walls. This substitution of niches for proper chapels is not unknown, and occurs in the chapels of Pavara and Ziza at Palermo and in the little late church of S. Thecla, the present Toklou Dede mesjid, in Constantinople. Something of the same kind is found in other small churches there, but the arrangement was not very common and seems to indicate a comparatively late date, about the end of the tenth century or after. Nor should I omit to mention the trefoil arrangement of the sanctuary in the eleventh-century church of S. Elias at Salonica. The provenance of the common sanctuary with three parallel apses and the date of its introduction into the church architec- ture of Sicily and South Italy are certain enough. The circum- stances in which these churches find a place in my notes has been explained in the introduction to the preceding volume. But I may repeat here that they are interesting chiefly from an historical point of view, as almost the only evidence we still possess of the Byzantine occupation, and of the presence of a large Greek community after the Norman Conquest of these provinces. In the chapter on Sicily in the preceding volume, I have noticed the fact that there are practically no remains of the Byzantine occupation in Catania or Messina. The latter, how- ever, possessed, and fortunately still possesses a very beautiful little church, built in Norman times on a Byzantine model and now known as the S. Annunziata dei Catalani, the name being derived from a guild of Spanish merchants to whom it was assigned. It is a little basilica with a square sanctuary covered by a dome supported on pendentives. The roof of the nave is supported on pillars with classical capitals of Eoman date, said to have come from a neighbouring ruin identified as INTRODUCTION 3 the Temple of Neptune. There is the usual sanctuary arrange- ment of three semicircular apses, the main or central apse for the altar and the prothesis and diaconicon on either side. The interior of the church had been carefully restored prior to the earthquake, but the exterior was surrounded by houses and there- fore invisible. Since these houses were demolished the walls have been uncovered, and they were found to be decorated with various circular and lozenge patterns in coloured marbles or stones in a style made familiar to us from Pisan examples. The exterior of the dome and the central apse are decorated with little pillars supporting a semicircular arcade. Prior to the earthquake this church was so smothered in houses that only a small piece of the central apse and the porch of the west front were visible. Though much shaken and split from top to bottom, the fabric was kept together by the dome and surrounding houses, and when the latter were removed the church was shored up and is now standing in the open waiting restoration. The central and north apses and fragments of the nave walls are practically all that remains of the cathedral. The latter are made of rubble and cement so roughly put together that the wonder is the building stood so long. As the reader is no doubt aware, the little temporary cathedral, made of wood and iron, was the thoughtful gift of H.M. the Emperor of Germany. It is perhaps worth record that after the Calabrian earthquake a large number of wood and iron churches of the same kind were ordered by the Roman ecclesiastical authorities from a London contractor to replace those that had been destroyed. When I wanted to obtain some information locally respecting two or three out-of- the-way villages in the Aspromonte, I was referred, of all places in the world, to an ofhce in Pimlico. Among other objects of minor interest saved from the wreck are the two gigantic equestrian statues representing Mata and Griffone, the local Gog and Magog. These are made of wicker- work frames, covered by pasteboard, fixed on platforms with wheels ; in company of the effigy of a camel on men's shoulders, they were dragged about and figured in the annual civic pro- cession to celebrate the Feast of the Assumption of the Virgin, and possibly also the defeat of the Arabs. The two figures appear to date from the early part of the eighteenth century. 4 INTKODUCTION One of them represents a clean-shaven individual, dressed as a Boman warrior, who is usually mistaken for a woman, and the other a swarthy individual of Berber type. I infer that they typify the native victor and the vanquished foreigner, though popular tradition makes them represent fabulous personages variously named Cham and Rea, Saturn and Cybele, or Zanclos and his wife, who came down from the Nebrodian hills to found Messina. I have often thought the history of these effigies might be worth tracing, not only for local tradition, but for the light it might throw on our Gog and Magog and other statues of the same kind that are to be found in some cities of Northern Europe. As I am writing about Messina, I may digress for a brief space and refer to the part my friend Commendatore Luigi Sofio took in the earthquake and its sequel. To him and his good parents, my old and esteemed friends the late Mr. and Mrs. John Sofio, my wife and I owe some of our pleasantest travelling recollections in Sicily. The Sofio family escaped the earthquake by only a few hours, for they left Naples on the eve, and the train they travelled in actually reached Bagnara, in Calabria, opposite Messina and in sight of the old Faro lighthouse, within an hour of the disaster.* Information of a serious earthquake had already reached the stationmaster at Bagnara, but as the light was still burning in the lighthouse the passengers did not believe the report that Messina had been destroyed. The train, how- ever, was unable to proceed towards Eeggio, and after some delay returned to Naples, where Mr. Sofio and a staff of doctors took steamer for Messina, and arrived to find the city in ruins. The rescue work had already been commenced by the sailors of a Russian squadron that happened to be cruising in the neigh- bourhood. Many interesting incidents arising out of the earthquake concern the identification of unknown children. Mr. Sofio was appointed curator, or guardian, of a host of these children, who were rescued from the ruins or found wandering about and unable to give any account of themselves or their parents. This difficult work has been so successfully accomplished that out of several hundred children under twenty remain to be identified. * Monday, December 28, 1908. INTEODUCTION 5 The following cases may be taken as illustrations. A Neapolitan doctor saw a sailor from the Eussian squadron rescue an infant, a few months old, from the top story of a house near the Via Sta Maria la Porta. He adopted the girl and took it to his home in Naples. The houses in Messina are usually let out in flats, and in this particular block lived a citizen of repute, who perished, as was believed, with all his family. Mr. Sofio knew the Neapolitan doctor, and hearing where the girl was found, had her brought this year to Messina for identification. In the meantime a claim to the estate of the family was made on behalf of two infant cousins, whose parents had also perished. The question of identity was, therefore, raised in a contested suit before the Court of Claims, and a jury of friends and relations empanelled decided in her favour. In another case two little lads were found wandering in the streets by a Russian officer and taken on board a Eussian man- of-war. Four years later a Eussian gunboat put in to Tripoli during the Italian occupation, and when the Italian Governor went on board the two boys were pointed out to him. Their pictures were taken and sent to Messina, the father was found, and he identified his sons. Then came the difficulty: the Italian Government claimed them, the Eussians did not want to give the boys up, and the father, finding his sons comfortably provided for, did not want them back. The solution of this problem is still under discussion. Some of the cases were amusing and many, as may be imagined, touching and pathetic. I hardly know which of these epithets to apply to the case of my good friend who, being a bachelor, suddenly found himself the foster parent of several hundred Messinese youngsters. Besides these duties he also has charge of the admirable model village, built in the northern suburb and maintained by H.M. the Queen of Italy. Eeggio suffered more severely than Messina, but it was, archi- tecturally speaking, a mean place, and the public buildings were of no archaeological interest. 6 INTEODUCTION By North Africa I must be understood to mean the French colony of Algeria and the protectorate of Tunis ; that is to say, the provinces known in Eoman times as Numidia, the two Mauretanias, the Byzacena, and Africa Proconsularis. Under the Eomans Tripoli and the Cyrenaica were administered as part of North Africa, but under the Byzantines they were attached to Egypt.* The extent of the Christian remains in Tripoli and Cyrene is still unknown, and owing to recent events it has been impossible for a foreigner to visit them, but it seems likely that the Church there was in rite and language the same as in Algeria and Tunis. In one respect, at any rate, the history of the Church has certainly been the same, for after the Arab conquest in the seventh century the orthodox Christians were completely wiped out, while in Egypt the Coptic Jacobite Church managed to hold its own, and survives as a flourishing institution. The student will find in the able and interesting works of Mrs. Butcher, Mr. A. J. Butler, and my friend Mr. Somers Clarke all and the latest that is known concerning the history, liturgy, ritual, and architecture of the Coptic Church. It seems plain enough that in the main the Copts followed the Greeks and not the Latins. There are no books on the Church in North Africa like the three I have just named, and so far as the architecture, at any rate, is concerned, the only sources of information are to be found in periodical publications and other works dealing with archaeology generally. The student must thank my friend M. Saladin for writing the first, and till recently the only con- secutive notice of Christian edifices in the interior of Tunis ; to his general survey must now be added the more detailed examination made by M. Merlin, the learned Director of the Department of Antiquities in Tunis, and the work of M. Diehl on Africa generally. The truth of the matter is that until recently Christian archaeology was entirely neglected and the attention of the antiquaries directed to the earlier Eoman remains and inscriptions. Concerning the constitution of the Church in North Africa, we know that after the * Peace ' of Constantino it was presided over by a Metropolitan at Carthage, that the provinces, each with its own primate, corresponded to the civil provinces, and * The exarchate of Africa also included Sardinia. INTKODUCTION 7 that there was a large number of bishops scattered all over the country, but the names of several of the sees are still unknown, and of those that are known many still remain to be identified. We also know that Latin was the language used, as appears? by the writings of the African Fathers, by dedications and inscriptions, and that after Justinian's time, at any rate, the Roman see possessed an extensive patrimony. These patri- monies gave the Pontiff much power and influence as landlord, and for a time at any rate the rectors presided over the local assemblies of bishops, so that the Church in Africa was attached to Rome by something more substantial than a common language. Moreover, it is clear that the older and larger churches in North Africa were designed and fitted upon the same model as the basilican churches in Rome. During the first period of Christianity, between the Peace of the Church when the Christians were allowed to practise their religion openly and the Vandal conquest in the fifth century, the Church was distracted by the Donatist schism. No question of faith was involved in this controversy, and as circumstances evolved themselves it developed into a dispute concerning the government of the Church and discipline. What liturgy was used or whether the Donatist service differed materially from the Orthodox we are not told. At any rate, I am not aware that the ritual required any substantial modification of the common plan and arrangement of the sanctuary and the altar, so that a Donatist church is not distinguishable from an Orthodox in the same way that the ordinary Greek church of to-day can be distinguished from the Latin. As the Donatists were accused of inviting, or at least assisting, the Vandal invasion of Africa, and as ultimately the stronghold of the sect was among the native Berbers in Numidia, it has often occurred to me to wonder whether the Donatist schism (at the outset a movement for a higher ideal or standard of conduct, like the Quakers) did not eventually develop into a kind of political home rule movement by the native clergy against their brethren of Roman origin. What I have just said regarding a structural distinction between the churches of the Donatists, and the Orthodox applies also to those of the Vandals. The Arian clergy took possession of 8 INTEODUCTION the existing Orthodox churches, and if any were built during their ascendancy the new are not distinguishable from the old. I now pass to the third or Byzintine period, that lasted from the conquest of Justinian till the closing years of the seventh century. During this time North Africa was governed more or less directly from Constantinople, at first by Justinian and his immediate successors, and later on and for the greater part of the time by Heraclius and his descendants. Nearly every church of importance that I have seen in North Africa shows that it has been extensively restored, if not practically rebuilt out of materials taken from classical buildings. There is every reason to attribute these restorations to the period when Justinian set about reconstituting the Eoman Empire in North Africa. Beside the purely military operations, this con- sisted of colonizing and cultivating territory that had been depopulated and devastated by the Berbers during the Vandal dominion, and of rebuilding the churches ; the Emperor also placed the clergy in an official position that they had not occupied before, and not a little of the subsequent trouble over religious matters may be traced to the great power and importance acquired by them at this reformation. The fact that the ritual arrangements of the sanctuaries, the altars, tribunes, baptisteries, and so on, were the same in the restored as in the ancient churches seems to point to the fact that there was no substantial alteration in the service to be performed. It seems, therefore, impossible to distinguish the date of the church from any structural peculiarity adapted to a particular and varying form of ritual as at Constantinople, and the con- clusion seems to be that the early form of service was continued after the Byzantine conquest, and that the arrangements to screen the altar with stone or wood screens across the chancel arch, in the manner adopted by the Copts and Greeks of to-day, was unknown ; and if any screening of the altar did take place it was effected by a movable curtain suspended from a baldacchino or rods and drawn to conceal the celebrant from view in the manner practised by the Armenian Church. I conclude that the African Church commenced by being Latin in language and in rite, and that during the Byzantine dominion it continued also in faith and Church government INTKODUCTION 9 substantially as it had been before the Vandal occupation, and so remained until the fall of Carthage in 699. Though this event may bo conveniently chosen to mark the termination of the Christian empire in Africa and the commence- ment of the Mahometan dominion that was destined to last for just over twelve hundred years, in Africa as in Sicily the Arab conquest proceeded slowly. We may attribute the ultimate success to the general weakness and disorganization of the Byzantine empire during the preceding fifty years, and select as the turning point of events in favour of the Arabs the revolt of the Viceroy Gregory against the Emperor Constans II. The defeat of the Greek army at Sbeitla that ensued placed the whole of Southern Tunis in the hands of the Arabs, and though the Greeks were able to defend the fortified towns for some time to come, the country districts were left to the mercy of the invaders, "We know very little about the subsequent campaigns of the Emperor Constans II in Africa. They were probably attended with some measure of success, for the Greeks were able to hold their own and some of the lost territory seems to have been regained. But these operations came to an end with the Emperor's death, and his son Constantino IV, Pogonatus, was otherwise engaged in defending Constantinople in a siege that lasted for seven years, besides carrying on a war with the Arabs in Asia Minor. Mr. Diehl, in UAfrique Byzantine, devotes many interesting pages to an account of the local history of this period, and so far as the relations between the Churches are concerned it is evident that the orthodox Latins, under the protection of the Viceroy Gregory, himself orthodox, held their own against the Greeks, who at that time, and under the auspices of the Imperial family, professed Monothelitism, During this rivalry Abbot Maximus, the champion of the orthodox party, was accused by the Greeks of having occasioned the loss of Tripoli by his evil counsels, just as later on the Pontiff Martin was blamed for the loss of all the Western provinces of the empire. x\s these charges arose out of the Latin relations with the African Church on the one hand and the Emperor Constans on the other, I may briefly allude to the circumstances here. The student is of course aware that from the Peace of the 10 INTEODUCTION Church up to the end of the seventh century, Christians, and more especially the Eastern Communities, were distracted by doctrinal disputes touching the nature, the will, or the energy of our Lord. The Emperor Heraclius attempted to reconcile the exponents and adherents of the rival doctrines on these subjects by an official pronouncement, or, as we should now call it, in diplomatic phraseology, a formula. The attempt not only failed, but raised other" issues affecting the Emperor and his family, and when he died the Church was left divided into two rival camps, the orthodox who supported the ' dual ' doctrine, and the heretics, who are known to us under the names of their founders or such cumbrous titles as Monophysites and Mono- thelites. The student may obtain a limited idea of the issues on questions of faith involved in these controversies by con- sidering the doctrine that declared our Lord to have been possessed of a single and that a Divine nature. The orthodox, among whom must be reckoned the Latin and African Churches, argued, and to us also it must apparently logically follow, that if this be the true interpretation of the mystery of our Saviour, the doctrine of the incarnation becomes a myth, the Godhead must be held to have suffered on the Cross, to have died and have been buried. I need not pursue the illustration. The dispute and this formula caused a great stir in Christen- dom, and was not finally set at rest till the close of the century in the reign of Constantine IV, Pogonatus. In the meantime the Emperor Constans II (the grandson of Heraclius and father of Constantine IV, Pogonatus) conceived the idea that the con- troversy might be stopped by forbidding any discussion about it, and this was to be done by the imperial edict or rescript issued in the Emperor's name and known to us as the ' Type.' The edict of the grandson, Constans II, produced as great a stir among the orthodox as the formula of the grandfather, Heraclius. The formula and the ' Type ' differed, however, in two im- portant respects : the former was no more than a statement of faith and expression of the Emperor's views ; the latter expressed no opinion on the points of faith, but it forbade any discussion about them and was issued in the name of Constans II as an order of State. I have already alluded to the Emperor Constans II's relation INTEODUCTION 11 with the Latin Church over the Type, to the part the Sicilian, Calabrian, and Sardinian bishops took in the Lateran Council when it was solemnly condemned, and to the prosecution of Martin by the Emperor for disobeying it. The part the African Church took in the dispute that gave rise to the Type is recorded to us in two letters addressed to the Emperor and the Patriarch of Constantinople by a synod of African bishops. The letter to the Emperor bears some analogy to that addressed some two centuries later by John VIII to the Byzantine princes of Sardinia. We see from these letters that the African bishops held the same views as Martin and the Lateran Council, and the hostile critics were not slow to take advantage of the circumstances and attribute the misfortunes of the Greeks in Africa to the intrigues of the Latins, while those in the opposite camp accused the Emperor Constans of trying to force heretical doctrines on the orthodox community. But in truth there seems no evidence that the Emperor Constans' religious views were different from those of his father or his son who were orthodox. And the Latin objection to the Type is based, not on any definite pronouncement in favour of this or that doctrine, but because by forbidding discussion on the controversial question of the day regarding the attributes of our Lord, the edict might be held to imply that one doctrine was as good as the other. After making due allowance for the part that theology played in the politics of this age, it seems likely from the course of subsequent events that the substantial ques- tions at issue underlying the dispute between the Emperor and Martin of Rome turned less on matters of faith than of language, nationality, and Church government, and that it arose out of the gradual growth of the Greek element in influence and importance at the expense of the Latin. The Church historian of Sicily * has pointed out that the change from Greek to Latin commenced when Constans II came to live at Syracuse, and that the Emperor himself probably took a part in promoting it. We are not told what happened to the Church in Africa either in the reign of Constans or of his son, Constantine Pogonatus ; the acts of the latter as affecting the • Lancia di Brolo, the present Archbishop of Monreale, Storia delta Chiesa in Sicilia. 12 INTEODUCTION Church could hardly have been objected to by the Latins on the score of faith, as he was an orthodox churchman and presided over the Council at Constantinople when the heresy that occasioned the edict of his father was condemned and finally disposed of. In Sicily and Calabria the change from Latin to Greek had proceeded so far and so thoroughly that in the Emperor Leo Ill's reign these two Churches passed from the jurisdiction of the Patriarchate of Rome to that of Constantinople. The fact that the transfer was successfully accomplished seems to indicate that by then the Greek element had jDrevailed, if not in numbers, at least in importance over the Latin. We have no means of judging by a similar test whether the Greek element also pre- vailed in North Africa, for by the time Leo III came to the throne the Arabs were in possession, the Church had come to an end, and the Christians were given the alternative of be- coming Mahometans or leaving the country. In passing to Egypt, I should not forget to mention a fourth book, upon the churches of Nubia, written by Messrs. Mileham and Maciver, and published by the Pennsylvania University for the Eckley B. Coxe, junior. Expedition of Archaeology. This and the three other books already referred to will supply the student with the best and latest information obtained on Christianity in the Nile valley. I need not do more here than refer briefly to the churches described in these pages, and point out how the Church history and architecture of Egypt differ from that of Algeria and Tunis. The first and important difference the student will notice is that whereas Christianity came, so to speak, to an abrupt end in Algeria and Tunis at the close of the seventh century, a large and flourishing Christian community still exists in Egypt under the name of the Coptic Church. He will naturally inquire how it came about that the Christians in Egypt were not only tolerated, but even, to a limited degree, protected by the Arab conquerors, while almost at the same date those in Algeria and Tunis were given the option of becoming Mahometans or leaving the country. The circum- stances are, of course, matters of well-known history, and I INTKODUCTION 13 refer to them here only because they throw some hght upon the history of the Church in North Africa, and they occurred while the Byzantine Emperors of the Heraclian family, whom I have so frequently mentioned, occupied the throne at Constantinople. I have already alluded to the questions of faith relating to our Lord's attributes and Divinity that divided the Church from the fourth to the seventh century. I say the Church, but in point of fact these questions arose chiefly among the Syrian and the Egyptian Christians, and were debated in the great centres like Antioch and Alexandria between the orthodox fathers and their opponents known to us under a variety of names, the Nestorians, Jacobites, Monophysites, and Monothelites. The founders of the first two were especially connected with Egypt, for Nestorius himself was sent to the Khargeh Oasis in exile, and the Coptic Church to-day still follows a Jacobite profession. In the seventh century the Coptic Church was, as it still is, separated in faith and Church government from the orthodox Greek Church, locally designated then as now by the name Melkite or royalist. I suspect that the distinctions between the two professions was first brought into marked prominence a century earlier in Justinian's reign, and according to one authority rival missions were actually sent to Egypt by the orthodox under the auspices of the Emperor, and by the heretics under those of his consort, the Empress Theodora. The attempt by the Emperor Heraclius to reconcile the conflicting parties under his formula or ecthesis having failed, the Arabs appeared upon the scene when the dispute was still in active progress. As circumstances eventually evolved themselves, the orthodox became identified with the royalist imperial or Greek party, while the opponents threw in their lot with the Arabs. There is no doubt that to preserve their freedom the native Copts who followed the dissentient or heretical profession appealed to the Arabs for protection against the Byzantines ; and just as under the capitulations with the Sultan Mahomet II, the Greek Church is recognized and tolerated in Turkey, so under the bargain made with the Arabs, the Coptic Church was permitted to exist and still continues to-day. On the other hand, the orthodox Melkites who identified themselves with the imperial government were rigorously proscribed until a comparatively 14 INTEODUCTION recent date. The Coptic Church has therefore had the Egyptian mission field to itself, and exceptional opportunities of preserving its identity and customs. The language, however, was lost, and Arabic took its place. The last is perhaps the most remarkable result of the separation, for we may suppose that to the rank and file of the clergy and laity the question of language played at least as important a part in the dispute with the orthodox Greeks as that of Church government, or the abstruse doctrine about our Lord that is presented to us so prominently by the clerical historians. In these notes I have endeavoured to see if the date of a church can be fixed from the style of sanctuary architecture that was adopted to suit a particular form of service, and one object of my visit to Egypt was to see how the arrangements of the primitive churches in Algeria and Tunis might compare with those in the Nile valley. I think I am justified in saying that in the main the present ceremonies of the Copts, so far as the celebration of the Lord's supper and the sanctuary arrangements are concerned, are substantially the same as those pertaining in the Greek Church of to-day. But if the Coptic Churches in general are built to suit a service that is founded in the main on the Greek as it was developed after Justinian's revival, at least two structural peculiarities that seem to be survivals of an older ceremonial have been retained in the ancient churches of Cairo. As I have pointed out in the Introduction to the chapter on Algeria and Tunis, there is no indication that the ancient churches in those provinces were ever provided with fixed parti- tions to screen the altar in the manner of the Greek iconostasis. And if, as there seems no reason to doubt, in the earliest times at any rate the celebrant was hidden from the view of the congrega- tion, movable curtains must have been used for the purpose, either hanging from a canopy or baldacchino or from rods fixed across the body of the church. I am not at all sure that the baldacchino was as common in Africa as in Kome. In some of the ruined churches there is no trace of anything of the kind, and, unfortunately, one of the only independent pieces of evidence we possess in the way of a representation of the interior of a church, the famous mosaic now in the Bardo Museum at Tunis, INTEODUCTION 15 throws no light upon the question, and represents the altar as standing exposed in the middle of the nave. But on the other hand we have the evidence from the churches in Kome and the marble carving on the west wall of S. Sofia at Constantinople to show that a baldacchino was used ; and to these examples must be added some of the old churches of Cairo. It is quite certain that in the early period no stone screens existed in Egypt, and when these were introduced at a later date, the baldacchini were no longer necessary for their original purpose. Nor are they found in the comparatively late churches in the Nile valley. In the old churches of Cairo the baldacchino or canopy over the altar figures as well as the stone screen, and I can only explain the presence of the former by supposing that the baldacchini were retained from old custom. It is also note- worthy that the Coptic altars stand close into the apses, though these are fitted with seats or tribunes for the clergy. This was certainly not the primitive arrangement, where, as at Announa, Tebessa, Carthage, and elsewhere in North Africa, as well as in Home, the same arrangement of seats is found, but the altar stood right out in the first bay of the nave ; so that it would seem that at some time the position of the altar was changed and shifted farther into the sanctuary. And as the ancient churches in Nubia are fitted in this way, the change in this part of the world, at any rate, must have been made at an early date. In another respect the Coptic practice differs from all others by having several altars in the sanctuary, and three seems to be the usual number ; this custom is certainly not derived either from the early Christians or from the Greeks. I am not prepared to draw any definite conclusion from these peculiarities, and only note them here for the student's benefit, as they may be some guide to him in fixing the date of a church. At Addendan and Debereh in Nubia, for example, there is some indication that the sanctuary of the church was altered by the erection of a stone screen that did. not form part of the original structure. The other peculiarity in a Coptic church is that there are two fonts : a large one like a tank serving for ceremonial immersion at the Epiphany, and the smaller one used for the Sacrament of Baptism. I suspect that the tank represents the 16 INTKODUCTION earliest form of font, and when the custom of baptism by aspersion superseded that of immersion (not submersion) the old fonts were retained and utilized for ceremonial purposes- only. Something of the same kind occurred at Albenga in Italy, at Frejus in France, and other instances might no doubt be found. I speak with hesitation respecting the early practice of administering the Sacrament of Baptism, for I am concerned with it only in so far as the change in structure consequent on a change in practice may be the means of fixing the date of a building. The student who is interested in this question must accept what I say with caution, especially in the use of such terms as immersion, submersion, affusion, or aspersion. In fact, I have only used the first and the last. It would be manifestly impossible for an adult to be com- pletely covered by water in any of these Tunisian or Algerian fonts that are described here. They are not large enough by a long way, for I tried the experiment in several of them. The ceremony no doubt consisted in the candidate standing knee-deep in the font and having water poured over him, and that is what I mean to convey by the word immersion. By aspersion I mean the practice of sprinkling as followed by the Church of England and the Church of Rome to-day. In the Greek Church the child is actually plunged into the water head and all. This is certainly not the general Coptic prac- tice, whatever the theory may be, for the fonts are often narrow-necked jars tucked away in niches or recesses in the wall, where nothing of the sort could practically be done. Was a distinction made between the design of a monastic and a parish church? I have not been able to detect any, but it is quite likely that the internal arrangements were not quite the same. My father suggests that the trefoil church of S. Elias at Salonica may be an example of a monastic church, as there is no provision for an iconostasis : nor would an altar screen of that kind be necessary where the congre- gation was composed of persons who were in Orders. I am not sufficiently well acquainted with the Greek monastic INTRODUCTION 17 service and practice to express an opinion respecting S. Elias, but it is certainly the fact that trefoil buildings seem as a rule to have been connected with religious communities. So the trefoils at Sohag, Bethlehem, Tebessa, S. Honorat, Montmajeur and S. Menas in Egypt were all attached to monasteries. And that brings me back to the two monastic churches at Sohag. The student may rightly consider them to be among the most remarkable monuments of early Christianity that have come down to us, and for my purpose they have of course an especial interest owing to the sanctuaries being built on trefoil plan as cellae trichorae. The other well-known, or perhaps even better known, example of this arrangement in a great Christian building of early date will be found in the church of the Nativity at Bethlehem. It seems probable that this was the oldest of the three ; that the idea was taken from an earlier building in Syria, not necessarily a church, but perhaps a baptistery or a tomb. My father draws my attention to a trefoil font, with steps leading down to the cuvette for immersion, now lying in the courtyard of the mosque, once a Byzantine church identified as the Hodja Mustafa Pasha Djarai in the Balata quarter of Constantinople; a plan will be found in the text. But I have discussed the origin of the trefoil sanctuary in a great church elsewhere, and need say no more about it here beyond expressing my agreement with M. Saladin in thinking that this peculiar termination to a church was chosen for its doctrinal signifi- cance. The reader will see for himself that my expedition to the Christian cemetery in the Khargeh Oasis to find a trefoil tomb was not successful. I have already acknowledged the sources of much of the information contained in this volume in the Introduction to the preceding one. To the list of books which is reproduced here I have added the four on Egypt already referred to, and M. Saladin's report on the antiquities of Tunis that was omitted by an oversight. These and other added books are marked with an asterisk. 18 INTKODUCTION Besides the many friends in Sicily and Italy whose kind- ness I have already acknowledged, notably Professors Orsi and A. Garufi, I must now add Professor Salinas, my friends Mr. Macbean our Consul at Palermo, Mr, Churchill our Consul at Naples, and Mr. Jerome, the United States Vice- Consul at Capri. I have again to thank M. Merlin for helping me in Tunis and M. Saladin for his interest and for much valuable information and suggestion he has given me from time to time. In Nubia I have to thank Sir K. Wingate, Mr. Drum- mond of the Education Department at Khartoum, Mr. lies the Governor at Haifa, and Mr. and Mrs. F. Llewelyn Griffith for their kindness and hospitality to me at Faras, and to Mr. Woolley. In Egypt to my friends Mr. Somers Clarke and Mr. J. W. Crowfoot. I have also to thank Mr. Strzy- gowski for allowing me to use his photograph of the dome of the Red Monastery at Sohag as it used to be. My visit to these monasteries brought me to the house of Suwaris Basta, the son of Basta Bey, whose hospitality and kindness are well known to travellers. To the authorities in the Khargeh Oasis, Mr. Hornblower, Mr. Armstrong, and to the genial Director of the Museum at Alexandria, Dr. Professor Breccia, I also owe thanks for their kind help. I venture once again to beg the indulgence of the reader for the many errors and omissions that can scarcely be avoided in a book of this kind. E. H. F. Palermo, January 12, 1914. CELL^ TRICHORD Since our visit to Sicily in January, 1908, that is, a twelve- month before the earthquake that destroyed Messina and Reggio, our friend Professor P. Orsi, of Syracuse, has discovered the debris of two more cellae trichorae. One is situated in the Val d'Ispica in Southern Sicily * in a locality he describes as follows : * Presso rincontro delle due cavuzze che unendosi formano il corso idrico del Busaidone . . . alia sinistra del grande ponte, due vaste catacombe. . . . Larderia e S. Maria. Piu sopra in una spianata rocciosa in mezzo ai ruderi di un abitato Bizantino, una fabbrica detta S. Pancrati.' The other is at a place called Coppanello on the seashore just below the hill of Staletti in Calabria. Nothing but the footings of this chapel remain, and round about are the extensive ruins of a large monastery, possibly the celebrated establishment founded by Cassiodorus.f This is the ruin that we set out to find on our expedition from Catanzaro Marina to the Eoccelletta. Unfortunately we missed our way, and instead of proceeding along the coast road we turned up a lane that eventually brought us to Staletti. STA THEBESA by Syracuse We found that the floor of this trefoil chapel had recently been excavated by the proprietor, Commendatore Vinci. Human remains were found under the floor and also a late Eoman or early Christian marble altar front in the central apse ; the latter is now deposited in the Museum at Syracuse. On this occasion we noticed that the major part of the fabric of the little nave of • Scavi . . . 7iel Sudest della Sicilia. Estratto dalle Notizie degli scavi ; Anno 1905, fasc. 11 and 12, Rome, 1906. t Vol. i. p. 85. 20 CELL^ TRICHOE^ Sta Theresa is composed of large stones taken from one of the Roman buildings that this district is full of. The style of archi- tecture, the form of altar, and the use of old materials indicate that this chapel was erected in the early Christian period, and for choice in the commencement of the fifth century before the arrival of the Goths and Vandals. The remarkable similarity in plan and architecture between Sta Theresa and the Holy Trinity on the island of S. Honorat near Cannes shows that they belong to the same early date. This, then, may be a convenient place to describe and illustrate the Proven9al chapel that we visited in 1913 on our way to Sicily. ABBEY OF LERINS In Provence The foundation of this monastery dates from the beginning of the fifth century. As a school of theology it stood in much the same relation to Gaul as Mount Athos and Monte Cassino did to the Levant and Southern Italy, and in the twelfth century it certainly rivalled them in importance and renown. Among many celebrated scholars, divines, and fathers of the Church educated within its walls, S. Patrick of Ireland and S. Vincent of Lerins will be names familiar to English students. The community, founded about 400-420 by S. Honorat on the smaller island * that now bears his name, was composed of monks and hermits who lived together under a rule probably adopted from the Egyptian monasteries of the Nile Valley, but sufficiently distinct to have been recognized at the Third Council of the Church held at Aries (454) and identified as the Rule of Lerins. To this early period belong the present cloisters of the modern abbey, the church of the Holy Trinity with a trefoil sanctuary, and possibly one or two of the little chapels or oratories dotted about in different parts of the island. Though these notes are chiefly concerned with cellae trichorae, the cloisters of the abbey and the chapel of the Saint Sauveur are described here for the sake of convenience. * In Roman times S. Honorat seems to have been called Lerino to distinguish it from the larger island Lero, the He Ste Marguerite of to-day. m PHOTOGRAPHS 1 Uf(ir, S Theresa Cuba, SyraciiS( 2. Thmh, Alexandria. ■)'. Masoii's Marks, Lerins Cloister. SKETCfntS .4_~„. n V w w n o^ n n 17 1 Z\ 3- d /. Mnsori's marks, Kirchcrian Museum, Rome. Masori's marks, Aries Museum. J. S. Satnrm'iK, Cas^liari. To face page 20 20 HORiE Sta Theresa is composed of large stones taken from one of the Roman buildings that this district is full of. The style of archi- tecture, the form of altar, and f '^f old materials indicate that this chapel was ^^^e^f^^^^f^x^cv- Christian period, and for choice in the commencement of t^i-> fifth of>ntnry before the arrival of the (ioths Aud Vandals. ^ lanty m l.lan and arc^^e^ct^e be^^en^bt^^^^^ ^ on the island of S. Honorat near Ca:— _- o^..v. ^ . . . -; to the same early data2.'a.^'[j^lhs?J, may be a convenient place to describe and illustrate the Proven9al chapel that we visited in 1913 on our way to Sicily. The found alibn of the fifth century same relation to n Gaul W his mori; a school ol[_i|ieol('Jy 1 I I', l:- :mii^' of ood in much the 'Mount Athos aj^d Monte Cassino did to the Levant and Southern Italy, and inTne twelfth century it certainly rivalled them in irapoHahce and renown. Among many o=y- ^ith a trol ^il sanctuary, and possibly one or two of the lifthpis LnpRlR or oratories dotted about in different parts of the islanu*: Though these notes are chiefly concerned wirtfc«eN»\'ft:iv"iSe*i«\ tfh*bftl^stei|g of the abbey and the chapel of the Saint .iB^fil:evMivs4itt\(Jescribed here for the sake of convenience. ?.^\^K ,ii^ftm I'wois^V. >v ^-^v^(\^ ^i'jv^\ o\ •21 .-;-, , .ii^v^ U\^''A ^^tu^ N-K^J :oes leads to th;s roinant:' across the channel witli/^n*: auctuar}^ ms dictionaj the dome /f fh.'\ A^estern tl fisherfoik a^ a kr;> lun the w and ' t' The lo\N er aall stone cingfes by la rp The larger vl ve this si in Hi' iLiiiii.' of small materials and finis/ was addr 1 during the r-^, vrl r-,( the church wa :is /re mounted on • aryof the w^g^^^y^ ) lapd in a hi w i l . cv i u - ^ d * s o .' t CL le chi' uftn \S ^'^'R^ v,i\»V ^^ Plati. To face Page 21 LEEINS 21 Chapel of the Holy Trinity. This chapel stands at the eastern extremity of the island near the ruins of a Koman building said by tradition to have been a temple. It is hidden in a grove of pine-trees and, being invisible at a distance, may easily escape the visitor's notice. From the little harbour in the straits between S. Honorat and Ste Marguerite a path winding about between the stems of the pine- trees leads to this romantic spot. On a fine sunny day the view across the channel with the green woods of Ste Marguerite as a background makes an exceedingly bright and pretty picture. The path leads up to the west front of the church, and walking round the south side to the east end we noticed that the sanctuary consisted of three apses arranged in trefoil. In his dictionary of French architecture * Viollet-le-Duc says that the dome of this church is probably the first that was built in Western Europe, but he does not describe the sanctuary, and the fact that it is arranged in trefoil came to us as a surprise, for we had not seen M. Moris's t description of it. Continuing our walk round the north side, where there is an outbuilding of comparatively recent date that seems to have been used by the fisherfolk as a kitchen, we returned for a closer examination of the west front. This is now composed of two parts ; the lower and older portion extends up to the line of the sloping roof, and the upper, made of small materials and finished off in a parapet with battlements, was added during the Spanish occupation of the island in 1635, when the church was turned into a fort or blockhouse and guns were mounted on the roof. The lower part of the west front is very roughly built with small stones laid in a hard cement and bound together at the angles by large dressed stones obtained from a Roman building. The larger blocks used to make the posts and lintel of the door give this side of the church the appearance of a prehistoric or Cyclopean building. Above the door is a lunette window, and again immediately above it a rude cross made of bricks embedded in the masonry. There is some trace of a second cross, but no other attempt at decoration on this side of the building. • Dictionnaire de V architecture frangaise, vol. vi. p. 348-350. f L^Abbaye de Levins. Henri Moris, Paris, 1900, Plon-Nourrit. ^2 CELL^ TKICHOE^ Entering by the door, placed a little north of the centre of the wall, we come into a broad nave covered by a barrel vault made of stone. This is divided into two bays by a rib following the contour of the vault and supported on each side by a detached pillar. These pillars have stone shafts and plain caps, and were no doubt obtained from a Koman building. The views of the interior of the nave taken from the points marked a and h on the plan are not much of a success, but the best that could be obtained in almost total darkness. The view of the north side shows the pillar and plain cap, and beyond it a small square window in the wall. The view of the south side shows the position of a door in the east bay. A sketch of the interior of the church taken from the door and looking eastward towards the central apse gives an idea of the general appearance, and shows an arrangement of seats round the walls similar to that in S. Giovanni at Sinis, in Sardinia. The nave and sanctuary are separated from one another by a wall pierced by a round arch springing from a plain cornice. The wall on the north side has at some time been decorated with wall paintings of figures, and a very faint trace of colour may be detected here and there. In places the black outline* of figures and designs are visible, but the damp and decay have obliterated the colours too much for the subjects portrayed to be distinguishable. The trefoil sanctuary consists of a square chamber covered by a dome on pendentives and three semicircular apses covered by semi-domes ; these semi-domes rest on a cornice at a different level and of a different pattern to that found in the nave arch. Each apse was lighted by a round-headed window ; the central apse window is still open, but the other two are blocked up. The remains of a stone altar slab and the pillar or foot that supported it were found lying on the floor of the central apse ; there were also traces of interment, showing that the chapel had been used for the same purpose and in much the same way as the trefoil chapels in the cemetery of S. Callixtus at Kome ; the term shrine might be appropriate to describe this sanctuary. The dome is slightly conical, and rests on a plain cornice that separates it from the supporting pendentives ; it is made of * Compare this method of outlining in black with that in the Abu Seir chapel in Nubia, illustrated below. TRINITY CHAPEL. S. HO NOR AT. Trinity chapel fruni the West door \(ii\ig the nave into the trefji/ sanctnarv. \1 To face page 22 . ..Ltle north of the centre of the .0 covered by a barrel vault made id mto two bays by a rib following the and supported on each side by a detached i,.:.,i. ^..- ■! .vp stone shafts and plaip '•"•^^' and were i.o doubt ru Roman building. '1 of the interior ave taken, from the points marken A on the plan art uot much of a success, but the best thai r.!.'o:.,,.,i : . -"i..ost total darkness. The view of the nu.v.. . ...v • and plain cap, and beyond it a small square he wall. The view of the south side shows the ^oor in the east bay. A sketch of the interior of the n from the door and looking eastward towards apse gives an idea of the general appearance, and • .-iiTangement of seats round the walls similar to that in at Sinis, in Sardinia. The nave and ^ v are m one another by a wall pierced by a ...... arch !u a plain cornice. The wall on the north side has • been decorated with wall paintings of figures, and •e oi'Sl9ll>M'li!X5\t¥;vVii'Si^ted here and there. In 0Utli;15%*t;s^(^V /'".'« " ' "■^•'■■i'? :n-(^ v!Kil)lr' hv,t V n^^Mfe^v,.,.-^ -->^-^ • !i sanctuary consist.-, o. .. .^^^uare . ;1 ■ vendentives and three semicircuJ^- ^^ — „„.„., i these semi-domes rest on a cornice at a different a different pattern to that found in the nave arch. lighted by a round-headed wini"" ' 1 . still open, but the other two — ['. of a stone altar slab and the pillar or foot that d it were found lying on the floor of the central apse ; ■ -^s of interment, showing that the chapel had „ .„. I..;. .,arae purpose and in much the same way as . )\l chapels in the cemetery of S. Callixtus at Borne: the t^xm shrine might be appropriate to describe th :s slightly conical, and rests on a p' !; n J .: from the supporting pendentives . 'o of * Compare this meLimd uf oiitlinink? in hlack with that in the Abu Seir chapel in Nabt», iUiutratod b< S.S. '»'^V^\ '^-JV»\ o\ LERINS 23 dressed stones and built up in concentric rings. As already mentioned, it is said by VioUet-le-Duc to be probably the earliest example of a dome in Western Europe. Eeturning now to the main door, and passing out to the south side, we come to the footings of some buildings that have now perished away. These consisted of a cloister and outbuildings, and when uncovered the steps leading into them were found to have been worn down and polished like marble by the pilgrims who came to visit the shrine. The masonry of the lower or original part of the church up to the line of the eaves and the later or Spanish addition with the parapet and battlements above can be clearly distinguished in the view of this side of the church. There seems also to be a break in the masonry where the nave and the sanctuary join, and this confirms our view that the two parts of the church w^ere not built at the same time, and that the sanctuary is older than the nave. The window in the south apse, though blocked up, can be clearly distinguished ; but the interesting detail of this part of the building is the ingenious device of the Spanish architect to support his parapet by the insertion of a kind of pendentive in the angle where the side apses join the central one. The view of the window in the central apse shows the decorative moulding that bears some resemblance to the moulding in the nave of S. Giovanni at Sinis in Sardinia.* The Cloisters The monastery was rebuilt in 1875, and on that occasion the mediaeval church and all the older buildings except the cloisters were destroyed. The accompanying illustration will give an idea of what they are like, and the plain solid masonry with the waggon vault and pilasters bear a resemblance to the fabric of the Trinity chapel. A collection of marks in the south wall deserving careful notice are illustrated on the accompanying plate. They represent two chisels, a mallet, an angle measure, and an adze. The mallet in a more primitive form, accompanied by a chisel and a Maltese cross, will be found on a stone in the walls of the early nave of S. Saturnino at Cagliari.t We took • Illustrated in vol. i. Plate 36. t Vol. i. Plate 39. The similarity in these details may not be quite accidental, for S. Honorat possessed the important abbey church of Saccargia and adjacent lands near Sassari in Sardinia. 24 CELL^ TEICHOE^ these to be emblems of our Lord's Passion, but the examples at Lerins explain what they really are, and we found some more of a similar kind exhibited in room L of the Kircherian Museum at Kome and the museum at Aries, and they are late Eoman or early Christian. Saint Sauveur. At the west end of the Island stands an octagonal chapel now identified under the dedication to the Saint Sauveur. Prosper Merimee * who visited Lerins in 1835 describes this chapel as a baptistery, and though there is no evidence of a font or water supply to serve it his conclusion is probably correct. He might have gone further and suggested that the baptistery was purposely placed here for symbolic reasons. The Christian life commenced at the west end of the island in the baptistery, and passing through the Abbey and church in the centre was completed at the east end in the place of entombment, that is the shrine in the Trinity chapel. M. Moris points out that by the decree of the Council of Aries the Abbot had the privilege of solemnizing the rite of baptism according to episcopal use.f The interior is now used as a wine-cellar by the proprietor of the neighbouring restaurant. The six sides of the interior are recessed in niches and covered with semi-domes, the seventh side is the entrance, and the remaining side projects into a little semicircular apse covered by a semi-dome. The roof of the chapel is a very flat dome. There are one or two other little rectangular chapels in the neighbourhood of the Abbey each provided with a semi- circular apse, but there is nothing else noteworthy about them except their respective dedications, t and no features in the simple construction to help determine the dates when they were built. The annals of the Abbey give no information as to the con- struction of the Trinity chapel, and the date can only be • Voyage dans le Midi di La France. t Compare this chapel with the baptisteries at Fr^jus and Albenga described and illustrated in a later chapter. I Moris, p. 380. guessed, from t ance of the : , -point in ti lu irvs had elected ■ diiced the rule of S was resented by a that ended in a moj- be plun»i . ;iie five li and re-endowed b} niagne, and f'>v From Charlem . Siif -ens became n of or but he laubt bavt^ Aries, wher-^ 1'^ i PontitT lo or reliquary reputed fifteenth, now prest ; identity of the t M.Moris con clu Eugenius III. (1161) a occun-ed long after of this opinion 1 a not later than the objects depicted ur. ano \v^\\\\^ r((\^K '^llSiiCi \ft ift'xovu>V\. .?. \o f^v>«^t\"i^ visit of a ^oden box ■ ry or early asse. The discussed. U-1124) and VI. (1522) . the last part -n that it wag and the other a the question, shape of some Latrii ^> ^■^ft<\ ^:iV'>\ oT ROME. Cemetery of S Callixtns West chapel. (For the plan see next plate. ) To face page 25 LEKINS 25 guessed, from the general history of the Society, from the appear- ance of the architecture, and by comparison with similar buildings elsewhere. After enjoying great renown and prosperity for the first two hundred and fifty j'ears of its existence, the year GOO marks a turning-point in the fortunes of Lerins. It seems that the monks had elected one Aygulph to be their Abbot, and he intro- duced the rule of S. Benedict from Monte Cassino. The change was resented by a section of the community and led to a dispute that ended in a civil war, and the plunder and destruction of the monastery. The monks had hardly settled their differences and rebuilt it when the Saracens appeared on the scene, and for a second time it was destroyed and the monks to the number of some five hundred were killed. The community was resuscitated and re-endowed by King Pepin le Bref and the Emperor Charle- magne, and for a short time enjoyed a renewed prosperity. Prom Charlemagne's death to the close of the tenth century the Saracens became masters not only of the coasts of Provence, but of all the western part of the Mediterranean, and exacted tribute even from Eome itself. At this time the Pontiff, John VIII., occupied the Koman See and made a journey to south Gaul in a Neapolitan ship. There is no evidence that he landed at Lerins, but he must have passed close by on the way from Genoa to Aries, where he landed at AVhitsuntide in 878.* The visit of a Pontiff to Lerins is, however, recorded on a painted wooden box or reliquary reputed to be of the late fourteenth century or early fifteenth, now preserved in the cathedral treasury at Grasse. The identity of the person represented has been much discussed. M. Moris concludes f that the visits of CalHxtus II. (1119-1124) and Eugenius III. (1151) are legendary, and that of Adrian VI. (1522) occurred long after the reliquary was made. With the last part of this opinion I agree, and came to the conclusion that it was not later than the twelfth centur3^ The dresses and the other objects depicted unfortunately throw no light on the question, and the suggested date rests chiefly on the shape of some escutcheons on the base of the casket that may be additions. * Itinerary of the Popes, in Trisor de Chronologie. Mas Latrie. Also Le Pape Jean VIII. Lapotre. t Moris, p. 425. 26 CELL^ TKICHOE^ The Saracens were not finally dislodged from the south of France till the time of William Count of Provence in 975. A century later the monastery had regained its former prosperity, and the fine fort that still stands on the south shore of the island was built in the abbacy of Aldebert (1075) to guard the com- munity against the Saracens. About this time, or a little later, the church and monastery, all but the ancient cloisters, were entirely rebuilt, and the works then executed lasted till the middle of the nineteenth century. A record of what they were like has fortunately come down to us from the description of a member of the Community written shortly before they were demolished. There can be no doubt that the cloisters and the Trinity chapel were built long before the time of Aldebert, and though it is of course possible that they are as late as the revival in the reigns of Pepin or Charlemagne, it seems more probable that -they existed before the community fell on the evil days that commenced with Abbot Aygulph and continued during the two centuries of Saracen oppression that ensued. There are two other trefoil chapels in Provence, one near Toulon, respecting which no information is at present to hand ; the other is the well-known building dedicated to the Holy Cross and attached to the monastery of Montmajeur near Aries. This chapel is known to have been consecrated in the first quarter of the eleventh century, and the architecture, the decorative details, and the vaulted roof instead of the dome are in the style that might be expected in a building of that date. Save in point of plan there is no similarity between these two chapels. BOME* Cemetery of S. Callixtus Nothing- is known either of the date or circumstances of the erection of these chapels, or of the persons who were buried in them. All that can be said is that they are Christian, of early * The chapels are referred to in Die AUchristUche Architektur. H. Holtzinger 1889, p. 246 ; also in Journal of Roman Stiodies, p. 107, vol. i, part i., 1911. 27 date, and chapels for th community. The first i • the western course quit [>«, .1, tnt's house • fi^llowing the 1 i round-headed i The second i i;> uh is seen i; apse the remains width metres side apses, .'' v,il fioor i- The next ,_ longer than ; carried out «< west wall, within it mination < There is >• but pr. SpOI: i .WU\«^\ that apt ws whero /, .]^ bricks . d witn'-±tfnijan ittle museum,! ; the surrounding iie total length tii apse to the rj th of the ^s deep ; wi ne tres d eep. It I'ld-fevel omMni'' ; This is a 1 re the nave VAjas ■ 4 in Ithp sQidh- Vdtn^just :al ter- door. ^e side, corre- in the astel- •arate pass Z^?. "^"^v^^ 'i::iVs\ oT X75, ROME. I. W. chapel, apsrs and South side. W. chape/. E. chapel. Plans. To face paf^e 27. KOME 27 date, and were probably intended to be memorial or funerary chapels for the interment of prominent members of the Christian community. The first picture shows the south-east front and the apses of the western chapel. The superstructure of the square nave is of course quite modern, dating from the demolition of the peasant's house that had been built up round it. This new work, following the lines of the old, takes the form of a lantern with round-headed lights in the gables. The second picture shows the interior looking into the central apse, and it is taken from the temporary steps leading on to the old floor. The remains of an altar with a tomb under- neath is seen in the foreground. At the foot of the central apse the remains of more stone coffins were found, so that appa- rently several people were interred. The picture shows where the new brickwork of the walls joins the old ; the original bricks are small and thin, of the kind usually associated with Eoman work. This chapel has now been turned into a little museum, and contains fragments of sculpture collected from the surrounding garden. The inside dimensions are as follows : The total length of the chapel from the extremity of the central apse to the door, 9 metres ; width of the nave, 5*25 metres ; width of the central apse, 4'30 metres across, and it is 2"70 metres deep ; width of the side apses, 3'50 metres, and they are 1'98 metres deep. The original floor is about 1 metre below the ground-level outside. The next pictures are of the east chapel. This is a little longer than the other, and at an early date the nave was carried out some 4 metres. The principal door is in the south- west wall, and the pictures of the interior are taken from just within it. In the foreground are the marks of the original ter- mination of the wall, and on the left of the picture, of a side door. There is no trace of a corresponding door on the opposite side, but probably one existed. It is not unusual to find small corre- sponding doors in this class of building. They occur in the Nubian church at Serreh, at Maccari, and at Delia, near Castel- vetrano in Sicily, and were perhaps intended to serve as separate entrances for the sexes, or to enable visitors to the shrine to pass across the chapel. 28 CELL^ TEICHOK^ In this chapel remains of an important tomb were found in the central apse. The altar seen in the picture is of course modern. The appearance of the brick masonry is similar to that of the west chapel. The inside dimensions are as follows : Total length from the extremity of the apse to the door, about 13 metres ; nave, 6 metres v/ide ; square intersection, 5'70 metres by 6 metres ; the central apse is 4"50 metres across, and the side apses 4*57 metres across. As the chapels in the cemetery of S. Callixtus were certainly used for burial, and it has been suggested * that the trefoil plan was copied from a pagan model, we have added to these notes a short account and illustrations of a remarkable building at the foot of Monte Cassino, near San Germano, known as the chapel of the Crocifisso. Though attributed to a pagan foundress, and said to date from the first century a.d., it is built in the form of a cross with a bee-hive shaped dome over the intersection : the plan in general resembles those of the two chapels at Camerina in Sicily and the Galla Placidia monument at Eavenna. The arrangement of bodies on three sides of a square chamber either built overground or dug out underground seems to have been common to both the Pagans and the Christians. While in Alexandria we happened to come across an example of this class in a little tomb chapel made of stone now preserved in the garden of the Museum, of this also we give a short account and a picture. MONTE CASSINO Chapel of the Crocifisso, near S. Germano In a little village about a mile out of St. Germano on the road to Kome stands the chapel of the Crocifisso. It is situ- ated at the foot of Monte Cassino, about 100 feet above the valley, and close to it on the south side are the remains of a large Roman amphitheatre. The top of this chapel, now covered by a house, and consequently invisible from outside, is on a level with the village street. The entrance is by a narrow lane with two flights of steps leading down from the main street of the village to a terrace or platform in front of the * Manuale di Arclieologia Cristiana, pp. 166-168. P. Sisto Scaglia, Rome, 1911. MONTE ( country opposite - The gru . irti. The . ^fjon vaults and th> . v.alla ri:, The intersv beehive-shapo > i I Sardinia, and •" <"'" away !'; .!ar clii'r I eat care and pi assessed of The floui „ locks of stonn The r.'. r size. lese detai •^re IS a mountain B IS arrange- once recall .fred by a Silanus it have • arance it with masons antwise ii^oiii^!^: ^^^«^^o^^-^. .-^-^^ ^^ Y"^" '^'^' -^ere clopean without 1 stones a good »'n from I'nts cut is lie over eW(^ ,^^v^.w-.o — o . .^ , _i-^ marks, .,■ decoratic^^XH- ^^•*^-' '-H— ^ ^^^ ^^ ^^""^^'^ '^^"''^^.| saints c monks in the ) Tb*»pf lave now becoini lie twelfth cent; The dimen^' iiapel from ' iietres ; the wi* 'istance acr oO that of til metres dee]> . aetres abo^• No .1. ' itiilfl; M " Qel in S. Narthex marked C on plan. lo fare paj^c 7^ CELL/E TRICHOR/E ijdt. partr of ♦ 1 . . »^.v . from the .'■:; a ]i)ile Theodora, aonuments ays apiece, tyith intro- • .rOifth, it in the liew outsi fcaids from shows* the < * I lie culti- 1, and iu the cu-Lic 1,1^ QB^^' looking with sloping walls t^ ^ \Kh the pro- lE^IPpiHBHBilllliprture. No [j^^whit^dorn e|L '^:<> fnitsidc diljflpnsion of thejtiona^ ng and about 20 lie wall that now has really become a curtain. il • . ~w f 1^ i ■ ''■ '^l ■ ■l.. .. .. •l SOHAG 35 and is so described in these notes, was originally pierced by- four doors, one on the west or desert side, one on the north side, and two on the south side ; one of the last is now used as the entrance, and the others are blocked up. In the original arrangement, the enclosure within the curtain wall was roughly divided into three unequal parts, of which the two smaller ones on the south and western sides were occupied by corridors, and the other third and larger part by a great church of basilican form provided at the east end with three apses built in trefoil shape to serve as sanctuary. All that now remains of this great church are the three apses covered by semi- domes, parts of the walls, and some fragments of decorative detail ; the nave has been completely gutted, and till recently was filled with a number of small brick houses. It will be easier for me to describe and the reader to under- stand the interesting points in this building if I refer to the letters on the plan and the corresponding letters on the pictures and proceed as if we had entered by the south door and were going round the place together. Entering by the south door we stand in the corridor at the point A. There are two pictures looking towards the door, B, that leads into a chamber said to have been a library. On the left is the wall of the south aisle of the church, and on the right the curtain wall we have just passed through. At the top of both walls there are little round-headed windows in a row, and below them cornices with corbels. The corridor we are standing in was therefore covered by a timber ceiling resting on the cornice, and the monks' living rooms above it were lighted by the windows. The original fabric of the fifth century is made of stone, while the Coptic restoration of the twelfth century is of brick, and that distinction applies to every part of the monastery. The next picture is also taken from A, but looking in the opposite direction towards C. C is the semicircular apse of a Coptic church that was built at an early date in this corridor. The apse was fitted in the usual way, with seats for the clergy and a Bishop's throne in the centre, the position of the latter being marked by a cross carved on a step-stone now lying just below the end of the fallen pillar. The apse was decorated like Announa and other North African churches, with pillars on each 36 SOHAG side of the entrance. The arched building X, seen beyond, is a much larger Coptic erection covering a cistern or well. From A in the corridor we pass through a door to D in the nave. I give a picture of this door, looking back from D to A, to show that the lintel is made of a block of stone with a triglyph carved on it obtained from a classical building, so that from this and an Egyptian fragment on the lintel of the north door and elsewhere it is clear that ancient materials were used by the architect. The next picture, taken at D, is from my sketch-book. We are looking at the wall EFG, and the dome K, just visible on the top of it. The north wall of the basilica H is on the left, and the party wall between the basilica and the south aisle marked J is on the right. This party wall was made by the Copts of brick and was built up between the great pillars of the nave, on the south side, that are still seen embedded in it. The next view is taken from the top of the wall EFG, looking at D, on the floor of the church, to show the place in front of the white door where the preceding sketch and views are taken from. This bird's-eye view shows the modern houses that still occupy the spaces Y and Z on the plan, and also the arrange- ment of the stone and marble pavement. Before the recent res- toration all this part of the church was filled with houses, and the pavement space became a lane or street between them. To return to D and the picture taken from my sketch-book : the reader will understand that we are now standing in what was the nave of the church and looking east towards the sanctuary. In the original church neither the wall EFG nor the dome K above existed, for these are Coptic additions made of brick after the original nave had been gutted. The Copts had then to satisfy themselves with the trefoil sanctuary for their reduced church, the reason for this being apparently that the original nave was covered with timber and tiles, while the roofs of the three trefoil apses were made of masonry, so that when the nave roof was burnt out (for there are traces of fire) only the roofs of the apses remained. The Copts then covered the square space N with a dome, made a narthex covered by three small domes, L, K, M, in front of it, and then built up the brick wall EFG to close their church in. If the reader will bear this in mind so HAG, White monastery. M!^ Sketch of the ancient nave looking east from D to the wall E.F.G on the plan; small domes T,.K.M above; and pulpit on the left. --; See plan un Jiy sheet J. The pictures mentioned in the text.^ taken from A to B on the plan., from D to A and from the top of the wall E.F.G, are omitted. I Tft face pa^e '^6 ..^ on the right. This naity wall was made by the t brick and was buiTt^ up be'tween the great pillars of side of the entrance. Ixie arcJied building X, seen beyond, is a ,.>w 1. hvget Coptic erection covering a cistern or well. a A in the corridor we pass through a door to D in the nave. I give a picture of this door, looking back from D to A, to show tl' ' I is made of a block of stone with a triglyph carved .. d from a classical building, so that from this and an n fragment on the lintel of the north door and elsewhi clear that ancient materials were used by the architect. Thn >u>\t picture, taken at D, is from my sketch-book. We 8n- at the wall EFG, and the dome K, just visible I it. The north wall of the basilica H is on the left, rty wall between the basilica and the south aisle This Daily it up betwt Ib.e nave, on the south sidel<."^^i^*^ ^ffrseen embedded in it. next view is taken from the top of the wall EFG, ^'t T^c^o-^'Si^^\ift^otl>^vflhi»"^wto?>4Si\ps\'^:tih^place in ■■ i'lf t,%lw<(;fe <\\«>w^\r^^-«w3V»«^^ Ww>\v^^tiJ^iMi\fe(^r& houses that ^ the spaces Y^i^d^^ o^^he^J^^^d also the arrange- ,!<• stone and marble pavement. Before the recent res- . .ill yaJi^]aft*lh\ot\:*iiie?»4iiiw«lftwwte^iWc'^'t*r^^^ l^ses, and ttko j. iVenienV«wd the ^e^ile %lfeft*?r>>?ii^y*^etch-book ; . will understand that we are now standing in what wa'- • of the church and looking east towards the sanctuary. Jn the original church neither the wall EFG nor the dome K ind, for these are Coptic additions nv ' '' ' '. after , .al nave had been gutted. The C' ^ ;ien to -V, i'^!> tiiemselves with the trefoil sanctuary for their reduced church, the reason for this being apparently that the original nave v, ' red with timber and tiles, while the roofs of the three ti . .. ..^ ies were made of masonry, so that when the nave ruof was burnt out (for there are traces of fire) only the roofs of the »p$«fi remained. The Copts then covered the square space N wit' * made a narthex covered by three small domes, L./ , orit of it, and then built up the brick wall EFG to close th^ church in. If the reader will bear this in mind fvl V 1 '1 i f if » f ?1 WHITE MONASTEEY 37 he will readily understand the illustrations and explanations that follow. It will be perhaps more convenient to leave the nave for the moment and proceed at once to the door F in the brick wall, noticing the Coptic dedication crosses over the lintel. Passing through the door F we enter the Coptic narthex and stand under the dome K, flanked by the smaller domes L and M. These are shown in the photograph taken from inside F, looking upwards. The next view is taken from the same place in the narthex looking across N into the central apse. Three structural details will be noticed here ; first that the central dome N is of Coptic brickwork and rests on four piers, next that the two piers between N and K have columns embedded in them, and thirdly that the apses and P are separated by a passage from the piers and from the central apse. These details show that the space between the apses is now only partly covered by the dome N, and must in the original church have been roofed by timbers and tiles. So that when the latter was destroyed, and when the Copts covered this sanctuary by a brick dome, they had to make a square sub- structure to support it. This was done on the side nearest K by building up pilasters or piers round the ancient pillars that supported the chancel arch of the original church. The top of the capital of the pillar on the west side will be seen on the left hand of the picture projecting a little from the face of the pilaster. No attempt appears to have bsen made to restore the basilica* as a whole ; all the Coptic architect did was to adapt the chancel of the old basilica into a small church or chapel as we now see it. Judging from what I have seen of these great Christian basilicas elsewhere in Africa, I imagine that when once a basihca had been destroyed, the same difficulty of reconstructing presented itself to the Coptic as to the Tunisian and Algerian architects, but perhaps even in a more pronounced way. As a general rule these great buildings were covered with timber and tile roofs, and when, as usually happened, these were destroyed by fire, the ways and means that enabled the original builders to obtain timber from Europe were no longer available. The restorers then adopted the expedient of dividing up the original structure into a number of square divisions or bays with transverse arches supported on 38 SOHAG piers, and then of covering the square partitions thus made with small brick domes. If in this case the same scheme was ever adopted for the whole basilica, all that now remains is in the present church consisting of the square intersection between the apses and the bay of the nave next to them. In this case the western arch of the first hsbj was bricked up, the covered space constituted the present and reduced church, and the rest of the basilica was abandoned. I now come to the ancient fabric. The semi-circular wall of the central apse is divided into a lower and an upper colonnade. In the restoration a wood architrave has been in- serted from capital to capital to replace the original one that supported the stone frieze and a little cornice above it. This arrangement is the same in both colonnades, but in the upper one the stone frieze is gone and has been replaced by bricks in three rows. There are six columns in each row, and the intercolumn space is occupied by round-headed niches : these little niches are set in kinds of frames with pilasters on each side supporting what I may describe as broken gables. This peculiar design is repeated over and over again in both the monasteries, but can be best seen in the Red Monastery where the niches have escaped destruction. The canopies over them are semi- domes and plain barrel vaults alternately, usually decorated with various pretty patterns, flowers, fruit, birds, stags and peacocks ; in one niche, on the north wall of the basilica, the peacock is represented with tail displayed facing the spectator. The barrel vault niches have usually a medallion with some device on the back wall. Owing to interposition of the wood screen and altar, it was impossible to take a picture of the whole apse ; the views taken in sections commence with the north corner of the lower cornice and continue round to the south. In the centre of the entablature above the lower colonnade is a square stone carved with a heraldic bird displayed : this I venture to say, with all deference to those who differ from me, is not a Holy dove but the Imperial eagle.* It will be noticed that all the capitals are different. Some * The Imperial eagle is frequently represented, as for instauce over the main door in the narthex of S. Sofia at Constantinople. SOHAG. White monastery. Trefoil saiicttiary of the ancient basilica now covered by a Coptic dome; a modern wood screen in front of the chancel; the north apse on the left; a Coptic wooden lectern in the centre. The pictures of the domes in the narthex are omitted. k \ To face page ^8 V.;. (•0\'-rin<^ li. ; pciLLlliuus lim^ lUHde witii sn_. If in till the same scheme was ever adopted t vhole basilica, »11 that now remains is in the present church c< 'f the square intersection between the apses and the ba\ ui ilu> nave next to them. In this case the western arch of the first bay was bricked up, the covered '••pace constitated the present and reduced church, and the rest of the; i>avHica was abandoned. ] ■ M- : the anciciii iaunt:. The seun-iiii-uiui w.m of tl ■; is divided into a lower and an up}>"r col')r. ill the restoration a wood architrave has been m- tal to capital to replace the original one that ■- frieze and a little cornice above it. This ar same in both colonnades, but in the upfK ■ Tone frieze H^^^^ft.^^^ ^^^ ^' ''^ «* cows. There are six columns in each . riii^ ■■'' IS occiipiea by rouna-neaaed niciu-^. uiesc supportiii.j^^^-|:^-5 i„j^^ (ig6^1i^'c.»v\ift:itkfe fiedk^Mofejl^fim® where the niches have unction. The canoDies over them are semi- dome^ .>^^|Vf^iS^^i?^^^^rvlfl\lft '•^!?l?natelf, usSallj decorated pretty patterns, flowers, fruit, birds, stags and - - niche, on the north wall of the basilica, the ; -ented with tail displayed facing the spectator. The barr»?l vault niches have^ usually a medallion with some the back wall. * ••-• -T^-^'Sltion 0( Liie \>ijiMi ,-i n • i: . •- >i iio iTnp. a picture of the whole ai .•■. views sections commence with the north comer of the lower .1-, and continue round to the south. In the centre of ntablature above the lower colonnade is a square stone d with a heraldic bird displayed: this I venture to say, with all deference to those who differ from > dove but the Imperial eagle.* u „;ii i>o noticed that all the capitals are different. Some I !if iriip'-rial eagle is frtvj ' i ted, as for«iiista ace over the main door in the narthex of S. Sofia a .. \ V ^ - .-^ '^y^ /^^ _j^ ■\.^-^-^ d i t'V "T' S23S^ 1 s >:.,;^.r' WHITE MONASTEEY 39 are either unfinished or what I may call shop or ready-made pattern to fill a vacancy ; others look like Eoman work, and two resemble the caps in the Eed Monastery that seem to have been made specially for the occasion. The entablature above the wood architrave is divided into two parts, the frieze, and a small projecting cornice ; the latter has a serrated edge and the face is divided up into little square panels with rosettes and leaves ; below the cornice is a running pattern of plain billets ; I shall not attempt to describe the patterns carved upon the frieze in alternate rosettes and whorls. All three friezes are different, and on the whole more elaborate in the side than in the central apse. The view of the upper colonnade shows that all the capitals are alike, and resemble closely those in the Red Monastery. On the second pillar there are two dedication crosses in high relief with the P turned the wrong way. The painting in the half dome represents our Lord seated with the right hand uplifted, blessing in the Greek way. This is Coptic work of the thirteenth century and probably coeval with the rest of the brick structure. So far as construction goes, the two side apses are repetitions of the central apse ; they are separated from the latter, as I have already pointed out, by doors leading into the rooms at the back of the sanctuary, and in that respect the arrangement differs from that of the Red Monastery where the two side apses join the central one and the doors to the back are cut in the side apse walls. I give some views of the side apses, and a general view of the south apse where a screen of wood stands in the way. In the south apse one of the columns in the upper row has a pretty dedication cross in a wreath of leaves carved in high relief. This should be compared with a similar cross on a pillar in the Red Monastery, and on the coin of the Empress Eudocia, illustrated on plate 42 in Vol. I. Some of the pillars in both the apses, and in both upper and lower rows, were obtained from an older building ; notice in the upper colonnade of the south apse a shaft in two pieces of which the upper is fluted ; and in the lower row of the north apse a hexagonal shaft. There remains the question where the altar of the basilica stood. The Copts have shut off the central apse by a screen or 40 SOHAG iconostasis and the altars within are concealed from the view of the worshippers. But it is certain that this was not the original arrangement of the sanctuary when the altar was habitually placed out in the open; and judging by the early basilicas in Tunis and Algeria the foundation of the altar here should be sought either in the intersection between the three apses, that is to say under the dome N, or if the precedent of the African churches be more strictly followed then further west in the first bay of the nave under the Coptic dome K, or even further west. The position under N is on the whole more likely, but the question could be determined by digging up the floor. What I have just said applies also to the Ked Monastery. The present Coptic arrangement is of course a comparatively late alteration. The plan will show the arrangement of the chambers behind the sanctuary. The construction of the semi-circular apse within a square outer wall and the introduction of chambers in the angles are common features in African churches. In one of the rooms on the south side is the Coptic font, and adjoining it an octagonal chamber that may once have been the baptistery. It is now used as a store room and much too dark to be photographed. This chamber has doors on the west and north sides ; the former is the main entrance, and over it is a small round-headed lunette with acanthus foliage and a Greek cross (arms of equal length) carved in stone. The six panels of the octagon are decorated with niches (half dome and barrel vault tops) similar to those in the apses ; the semi-circular half domes and the waggon vaults over them are decorated with scallop shells, and in the gable over the north-east and south- west panels are figures of a tall jar with vine and grapes issuing from it. Above the niches a cornice is decorated with rosettes and stars like that in the apses of the church. To reach the chambers on the north side of the sanctuary we have to return to the north apse by crossing the chancel. Here are two other rooms, one of which forms a corridor leading by a broad staircase to the roof. These stairs are part of the original structure, and in the walls are fragments of carving obtained from an Egyptian building. The views of the nave looking down from the roof and of the cupolas have already been referred to. to SOHAG. WHITE MONASTE Details in central apse, with dedication cross on a pillar shaft ; and the Imperial eagle in cornice : niches and cani)pies. To face page 40 40 a ic*'iiv-i...-.i^ Ai ,vi...... ,^^^. concealed froni iii.; ViCN. of the worahij ;t it is certain that this was not tt ■; original arraii}j;enjent of the sanctuary when the altar wu- habitoally in the open; and " by the early basilicas i« . Algeria the foundai.^,-. ■ . ;' * >Har here Hhoald be sou i- in the intersection betv, three fip«es, tha ^v under the dome N, or if the precedwit of tl be more strictly followed then fn -t )n ...c nave under the Coptic dome i\. fui position under N is on the whole more but the q- ould be determined by digging up the flooi What I ; t said applies also to the Red Monastery. The pr- -> ic arrangement is of course a comparatively late ai The plan sviU show the arrangement of the chambers 1 th' The construction of the semi-cir< wii — .. ,,ie outer wall and the introduction of ^.......k.c»„ in the *»»•'♦'*• are common features in African churches. In one 0' ims on the south side is the Coptic font, and adj' u octagonal chanftfe^rolft may once have been the . It is riaw^"a.^2lt£EA.Qi\liit(aaW(>Mi and much too dark i liographed,. This chamber has doors on, the west and u ■■.: Uie foriaer, is tlwi ipam entyance,.aJid over it is CTu..-, cijual length) carved in stone. The six panels of the . ; are decorated with niches (half dome and barrel vault tops) aimilar to those in the apses ; the semi-circular half dor ' ' waggon vaults over them are d' tiCii ">1 in the gable over the north-e; wt ^ures of a tall jar with vine and , from n. Above the niches a cornice is decorated with rosettes an ' bat in the apses of the church. .... ;ij!' chambers on the north side of the sanctn.'n-. w. havf to return to the north apse by crossing the cli: Here are two other rooms, one of which forms a corridor h by ' ircase to the roof. These stairs are pai"t oi ine ori^ ... „;:. .ure, and in the walls are fragments of carving obtained from an Egyptian building. The views of the nave looking dawn from the roof and of the cupolas have already been referred to. WHITE MONASTEEY 41 In the chamber marked R on the plan are some fragments and capitals that have been collected from various parts of the building. Some of them are Coptic, some Eoman, and some belong to the later period when the native artists introduced designs of their own, copied more or less from Roman models like those in Tunis and Algiers. There is nothing Byzantine about them. Duplicates of a plain and graceful capital will be found in the museum at Alexandria and in the mosque of Amr at Cairo.* To visit the west end of the church we have to return to the nave and pass through door D. The lower part of the wall H was decorated with a series of niches, above them was a row of round-headed windows, above them again square holes to take the joists of a triforium that was also lighted by a row of round-headed windows, and finally above all are the stone corbels to support the roof. The niches are constructed on much the same plan as those in the apses, and the arches over them are carved with various designs, scallop shells, peacocks, deer, flowers, vine and grapes, and interlaced foliage. The most original is a peacock with tail erect facing the spectator in a niche in the north wall. This wall was pierced by the north door that is now blocked up. The picture of it from outside is taken to show the crosses on the lintel and also a carved stone from an Egyptian frieze. Between D and W, that is to say on the north side of the nave, are the remains of a pulpit. The position should be compared with that of one of the churches at Faras in Nubia. We pass now to the west end of the nave. As the plan shows the new houses YZ abutting on the walls of the basilica are built across the middle of the nave, leaving a narrow passage or lane between them. Beyond these houses, and between them and the west wall of the basilica, there is an open space corres- ponding with the last three bays of the nave. The next view taken from this spot shows two of the pillars and the north wall Q decorated with more niches. In the north-west angle of the nave there are the remains of a Coptic building covered with a dome and of a fine squinch that supported it. The main door was in the west wall of the basilica and on * See illustrations of capitals at the end of this book. 42 SOHAG each side were two niches like those in the apse. Passing through this door we come into the narthex that separated the west wall of the basilica from the outer wall of the monastery. The view looking north shows an apse covered with a brick semi- dome supported on a cornice with the original wood architrave and pillars. The view looking in the opposite direction shows the south end T pierced by two doors, separated by a pilaster decorated with a niche. One of the doors led to a staircase and the other to a small room. On the left and right of this view are niches of the usual form and beyond them the doors U and V. The photograph of V is taken from outside. The Bed Monastery The Eed Monastery stands about a mile north of the White Monastery. The road, if such it can be called, passes along the border of the desert and the cultivated land through two small villages occupied by an agricultural population that is mainly Christian. The ground plans of the White and Eed Monasteries are substantially the same, but the latter is the smaller of the two, and though there is less of it the parts that remain are in better preservation. It has also been carefully and sufficiently restored to pres(?rve it from further damage. The main entrance was by the south door A under an elaborately-carved lintel decorated with little crosses. Opposite this door in a group of buildings outside the Monastery is a rectangular room with a domed roof containing the Epiphany tank. This room is of no particular interest, and was too dark to be photographed. Entering the south door A we come to an open space. The major part of this space, bounded by the wall opposite on the north side, was occupied by a large basilican church and a corridor on the south side of it. We are standing in the corridor C and looking across the nave to the north door D, decorated with a handsome carved lintel. The original trefoil sanctuary stood at the east end on the right of the view and has been converted into a church by the Copts in the same way as the White Monastery was treated. The work of restoration was only partly accom- II SOHAG. WHITE MONASTERY. Apse til West iKirtlicx at S on plan (see fly sheet J ). RED MONASTERY. Smith Door. To face page 42 iyjicii hide M-ere two i. ^ • nose in the apse. Pas-i this door we come le narthex that separate^ , wall of the basiliisa frtmi the outer wall of the monaster) view lo^ ^vs an apse covered with a brick semi- d(>me su; ,: '-ornice with the ; ■ ; ' '■ and pill;i 'W looking in the o} ; -i:wv,_ the sour' reed by two doors, separat oilaster deer* niche. One of the doors led to r; 1 ^' -l room. On the left and rigni ml form and beyond them the dooi of V is taken from outside. The Red Monastery i Monastery stands about a mile The road, if such it can be called. .'.le desert and the cultivated land thruugu i .ipierl hv an agricuK^^^^^pulation that i- V the same, but the latter is the smaller of the two, there is less.'^5^iS^TBeK]ft©ft*itlaRSfemain are in better i. It has also bee%)gQj't^iii^ and sufficiently restored it from further damage. n entrance was by the soum uoor A under an u ved lintel decorated with little crosses. Opposite group of buildings outside the Monastery is a loom with a domed roof containing the F. 111 is of no particular interest, and was tc the south door A we come to an open space. The ills space, bounded by the wall opposite - occupied by a large basil- "" "^ uthsideofit. We are stan and loo.i he nave to the north do me c&rv«;o liatel. The original ' stood at ' '/^ of the view and ni^ ia< m (...averted into :• . '• tibe same way as the White Monaster} was he work of restoration was only partly accom- ViH\\Z >^^ \^\vs. .31 V -^ ^ JD n CD -^ V "O^ ^-o < rr o ^^ 91 J— o. .uo\UA?,u^^a\ 'n^VN-vK Viu\\ \\oi\iJ^x-v?)?,\\\ 'ji^'i^o'^ ?iV ''>'^^^i '^''^^^ ^^"^ 2nd fly Sheet Coptic Inscription and Arabic translation. To face page 46 T'r £1 f roi : arc ! and N H' they ^^ original sim church in ti I will (X ' the wall caj arch V, suppoi ; pilasters nf • nearest tlu ported an ar arcn across t'l' the corners ' " - Thenicl: It is flanked cap) anl much i arch cr^' .%m&i ^M ^\\»'A«ft^^&^^^ ^'5^'5^ tv •^■^J^^ "^'^^X ^"^ 2nd Hv Sheet 12 ^ ~^ R £tT-M ONA STER Y. CV. riUTln chancel apse. o "fnscrifituman Coptic characters. -> ~--fwtth atranslation in a(h^ie^) ^]0 The illusitration and plan^-^ferrf^to (VT-' rec on ^i page ypp^site 43" are om^idd,-^ It w^Ube tmder- ij are omutm^~^Jt w^uoe tmder- j . gifoimd plan q/ the tvm-basilica» -^^ \ sUDstantiattyl^e same/ I C^"***^ * '>> V- o o To f^^yTTgT'JJ' J^ EED MONASTEKY 43 plished at the time of my visit, and a small piece of brick wall that formed the west end of the Coptic church was still standing between the pillars of the ancient church. A modern wood roof was erected to protect the old stone masonry behind. It will be seen that the adjustment of the three apses R, S, T forming the trefoil sanctuary is different to that of the three apses in the White Monastery, and as the intersection Q is square they fit close up to one another. The nave was separated from the sanctuary by a great wall from L to P. This was pierced in the centre by the chancel arch V, at the two extremities by two large doors at L and P, and on either side of the chancel arch by two smaller doors at N and N. If the reader will refer to the pictures and the plan they will explain how this wall was divided up, and how the original sanctuary had been closed in and converted into a church in the same way as the "White Monastery. I will commence with the view X to Y because the details of the wall can be clearly distinguished. This shows the chancel arch V, supported by two detached pillars K, K. Y and Z are pilasters of the pier adjoining K, K. The cap on the pilaster nearest the spectator Z is a restoration uncarved, and suj^- ported an arch across to the pillar H. Y on the opposite side of the chancel arch has the original carved cap and supported an arch across to the pillar F. H and F are the two pillars seen at the corners of the Coptic brick wall G in the general view. The niche over N is seen in the same view taken from X to Y. It is flanked on one side by the pilaster Z (with the uncarved cap) and on the other by the smaller pilaster c with cap at a much lower level than that of the capital on Z, From c an arch crossed to the pillar d, and from a on the opposite side an arch crossed to b. I must confess that this arrangement of piers and pillars in reference to the colonnade of the nave is to me very puzzling and difficult to account for. Passing now into the sanctuary Q, we see that the plan of decorating the walls of the apses with a double colonnade and niches is again repeated here. In this case the capitals are all alike and obviously made for this church. The interesting part of the sanctuary is the roof over Q, and 44 SOHAG I give a view of it as it now is, and, by M. Strzygowski's permis- sion, as it was before restoration when the Coptic brick dome still covered the space Q. The intersection was covered by a square lantern lighted by three windows in each side; each window was flanked by two little pillars standing on a cornice. The superstructure is modern, and the present dome is, of course, merely a makeshift resting on the walls of the lantern that have been carried up above the pillars. Strzygowski's view shows how the Copts had replaced the original roof by a dome lighted with small windows ; notice the way the Coptic architect made use of the little pillars to support his squinches in the angles. The photograph is interesting also because it shows how the inter- column spaces were filled up with bricks to support the entabla- ture where the wood architrave was gone. I do not suppose there is any doubt that in the original building the little pillars in the lantern flanked windows, and supported little gables or canopies, and that the whole square was covered by a timber and tile roof. The student who wishes to know what an inter- section of this kind looked like when covered by a timber roof will get a good idea from the Norman church of S. Spirito at Palermo. The view of the dome Q is taken looking up into the lantern ; the arch of the chancel apse is in the bottom left-hand corner^ the arch of the north apse in the top left-hand corner, and the arch of the south apse on the bottom right-hand corner. The apex of each semi-dome is ornamented with a flat scallop like that in the octagonal room in the White Monastery. Coming now to the apses themselves, it will be seen that there are four pillars in each colonnade of the chancel apse and five in the others. As all the pillars and their caps are of the same pattern it seems probable that they were made for the occasion, and not taken from an older building. The intervening spaces are again filled up with little gabled niches in which domed and vaulted roofs are placed alternately. The wood architrave has been restored, and above it is a plain frieze with a carved cornice on the top. As in the Y/hite Monastery so here the roofs of the apses- are covered with paintings of saints and Coptic legends. The ■3 SOHAG. Red moHastery. Trefoil sanctuary of the ancient basilica ; Coptic Tvood screen in front of the chancel ; north apse on the left. To face page 44 SOi: ^i I give a view of it as it now is, and, by M. Strzygowski's permis- sion, as it was before restoration when the Coptic brick dome still covered the space Q. The intersection was covered by a square lantern Hghted by three windows in each side; each window was flanked by two little pillars .standing on a cornice. The superstructure is modern, and the present dome is, of course, merely a makeshift resting on the walls of the lantern that have been carried up above the pillars. Strzygowski's view shows how the Copt« had replaced the original roof by a dome lighted with small wii' ' notice the way the Coptic architect made use of the litt;^ , ..as to support his squinches in the angles. The phoi 'i:raph is interesting also because it shows how the inter- coiu.'iM eipaces were filled up with bricks to support the entabla- "ere the wood architrave was gone. I do not suppose ^ any doubt that in the original building the little pillars lantern flanked windows, and supported little gables or >, and that the whole sQ^*if^Q*^s covered by a timber ■ roof. The student w^j^^^^^lg^ ^^J5^now what an inter- of this kind looked like when covered by a timber roof r a g^'»ia^^o'^r6fl!?^^^%^]^»«^fmi^^'^it11^^\^15«^'S^T^ at ■'"^ <• view of the dome Q is taken looking up into the lantern ; ■ !i of the chancel apse is in the bottom left-hand corner jT M of the north apse in the top left-hand corner, and the the" south apse gn the bottom right-hand corner. The each semi-dome is ornamented with a flat scallop like the octagonal room in the White Monastery. (iicg now to the apses themselves, it will be seen that iht:,i are four pillars in each colonnade of the chancel apse B,ni\ Fivt- in the others. As all the pillars and their caps are of * "'Mie pattern it seems probable that they were made for .ision, and not taken from an older building. The interver* 'i spaces are again filled up with little gabled niches in \ " 1 and vaulted roofs are placed alternately. The woo, ; has been restored, and above it is a plain frieze with a cai'Vod lornice on the top. As in t J ■' Monastery so here the roofs of the apses are covered w.tii paintings of saints and Coptic legends-. The -0.'.- -Ijr~^^\ ^OHAC- T?..^Mona.sr • ...y of these early monasteries. and in th< church of the Nativity at minent place of the sanctuary, or in ^V ^>,vs\ ^:ivi\^ o\. SOHAG 47 subsidiary places as at the monastery of S. Menas, at Tebessa, in the Damns el Karita at Carthage, at Monmajeur and S. Honorat de Lerins in Provence, in Armenia, and elsewhere. In these two monastic churches, as in the church of the Nativity at Bethlehem, we have examples of the cella trichora using the words in their real meaning as the trefoil choir or sanctuary of a great church and not merely as detached cells or chapels. In considering the decorative detail we stand on surer ground. There is evidence that in the White Monastery some of the pillars and capitals were obtained from an older building ; but in the Ked Monastery they are certainly contemporary with the church itself. There is nothing at all here like the true Byzantine capital as we find it in Europe, in Tunis and Algiers, or in the neighbourhood of Cairo or Alexandria. The forms are essentially Eoman and they only resemble Byzantine work in so far as the sculptor departed from the stiff conventional repre- sentation of foliage in Roman patterns and attempted to copy nature more closely. Viewed in that light, these transition styles which are no doubt to be attributed to the native provincial workers and usually dubbed debased, are a great improvement. It needs but a glance to see that the decorative details at Sohag are based on Roman and not on Byzantine models, and as the sculptor was certainly not influenced by the Byzantine designs that did find their way into lower Egypt, the inference seems to be that they were produced before the middle of the sixth century. The presence of materials taken from Roman buildings raises other considerations. In Algiers and Tunis we may certainly take Justinian's conquest in 535 as a convenient terminus a quo for fixing the date of a building where Roman materials have been used, for the Roman buildings in those two provinces were in the main destroyed by Genseric and his successors for military or poHtical reasons. Owing to incessant war between the Vandals and the Berbers, no attempt was made to rebuild the Roman cities that were destroyed until Justinian's generals had broken up the Vandal kingdom and resettled the provinces. By that time paganism was so far discredited that the temples became useless, and were either converted into churches or broken up for the materials. 48 SOHAG The truth is that the history of Algiers and Tunis in the fifth and sixth centuries is so well marked off into distinct periods by the invasion and ultimate expulsion of the Vandals and the final destruction of the Eoman Empire by the Arabs in the seventh, that it is comparatively easy to assign, if not a date, at least a period to any building of importance either before or after the Vandal occupation or before the Mahometan conquest. In Egypt nothing exactly corresponds to the Vandal conquest that in Algiers and Tunis so profoundly affected the older Eoman architecture and caused the destruction of fortifications and public buildings and the use of the materials for building churches. There is little doubt that for an appreciable period after the proclamation of Christianity the pagan temples in Egypt were not disturbed ; but from the reign of Theodosius onward many were destroyed either by the zeal of monks and converts or by the State as part of an organized endeavour to suppress paganisrn. I offer these considerations for what they may be worth and the student must judge for himself whether the slender evidence obtilinable justifies the traditional foundation of these monas- teries by Schenouda. It seems certain that these monasteries were founded long before the Arab conquest. SICILY CASTIGLIONE S. Domenico Since we visited Sicily in 1908 Professor Orsi has discovered a Byzantine church near CastigHone. This building is of con- siderable interest, as vi^ell for the architecture as for the ritual arrangement of the triple apse in trefoil, showing that this plan of providing for the Greek rite was not unknown in Sicily. The term Byzantine is strictly applicable to this church because there are no Arab or Norman features about it. The situation is remote and difficult to find, but the following directions will enable the traveller to locate it. He will probably make the excursion, as we did, from Taormina, and by starting early and using a motor-car the chapel at Malvagna can be visited on the same day. After leaving Giardini the high road to Catania skirts the sea- shore and crosses the Alcantara by an iron bridge. Just beyond the bridge a branch road runs in a westerly direction, following the course of the river towards Randazzo. The first town of importance reached is Francavilla, and here the road bifurcates, leading on the right to Eandazzo by Mojo and Malvagna, and on the left to CastigHone, a picturesque mediaeval town perched on the top of a high rock commanding the valley. The road follows the river Alcantara, and passing by a little Norman chapel and over a bridge bifurcates again at the foot of the Castiglione hill. There the road on the left leads up into the town and on the right goes to Randazzo. At this point it will be necessary to enquire for the property of a local squire, Signor 60 SICILY Cav. Giuseppe Sardo Euggieri, situated about half a mile down the Randazzo road and close to the right bank of the Alcantara. The church of S. Domenico stands in a field about a quarter of a mile from this gentleman's house. It is unfortunately now used as a cattle-shed, and on the occasion of our visit was in a ruined and filthy condition. Enough however is left to show that when perfect it was a graceful little building neatly planned and skilfully built of stone rubble set in a hard cement, with some bricks and blocks of lava used for decorative effect. But as so frequently happens in these parts, no account seems to have been taken of the proximity of Etna,* and the whole fabric has been riven from top to bottom by earthquake. Indeed the church would now be a shapeless mass of rubble but for the cupola over the nave ; and S. Domenico and the church of the Catalani at Messina illustrate what I have noticed elsewhere, that in practice the dome proved to be the most enduring and binding form of roof. Reference should here be made to the ground-plan and the illustrations. The former shows that the church is correctly orientated and consists of two parts, a choir with a semicircular apse for the sanctuary at the east end and narrow aisles on the north and south sides. The prothesis and diaconicon, necessary adjuncts to a Byzantine church, required for the due performance of the Greek liturgy, and usually placed in the east wall on either side of the central apse, are represented here by two little niches placed in the thickness of the north and south aisle walls facing one another. The nave, separated from the choir by a lofty round arch, is a rectangular room divided into two bays and flanked by narrow aisles. Three different methods of roof are represented in this church. The choir is covered by a voute d'aretes and the adjacent aisles by barrel vaults. The weakness of the voute d'aretes as com- pared with a dome is again illustrated in this building, where the crown of the vault has fallen in. The nave is covered by a dome, supported on skilful brickwork pendentives that may be des- cribed as a series of fans fitting into one another. The illustration will explain better than a verbal description this * In the eruption of 1912 the lava from Etna flowed down to within three- quarters of a mile of the church. is S. DOMENTCO. Castiglione. 'ielow N.W. and W. sides; ahotr S. side and prothesis niche. 1 Plan. To face page 50 r>o R« C»v. Giuseppe Sardo R I ted about half a mile down lo the right bank of the Alcantara. t-Tids in a Held about a quarter of I il^e. It is unfortunately now on the occasion of our visit was in a on. Enough however is left to show 'I eraceful^Uttk building n* "- - '- /] a m used as a (*at ruined and that - ' -: - and - lie rubble 'st^t in a hard cemen bricks and blocks d Java used^?8r^ecorative effect. But a.s JO ..V'i.n MtftMm<^**be!ife^|)aftsr«6^6dd\iSt •IJete^d'l-^^ ■ • -ty of Etna,* and the whole fabric has i oottom by earthquake. Indeed the cbi. : peless rQ«f9^..crfTTrtTbi^^>H^t for the cupola over Domeirtco^''a:Sd tTJ&sdicvrch of the Catalan! notice most ap6t / narr^v \iMted with L uld -heije be" ♦made to tbe groiindl-plan an I he farmer shows ^tbai the ch\ircli is correi Nd consists ojt two pp,!^, "Jfc- dioir with ^ semici -J? winctuai^ all the/east end and narirow aisles o i t le ' :L •i4.QS< — iPh e • protbesis- arPd .-d-i •jvaftn^fei lie church, requi red f or Jii^dueteeB'fnjy! jraKa^^, a*e represented here by twq; |(i)^e iii ch( s l/okdt^ls of: the north and south iisl»-|vaik^*cirg ;saTf" , — i 4r-^"n!|,Yi, separated from the choirjby |L;]x)lt>;'f(|ii'i|d rn divided iijito two bafs ^s ani; ti^jkj ! b-'I^^l^o^? 0^ ^^^^ ^^^ representeji iq^-ihi'S-e-^fcl j'V.rediby alvoute d'arStes and the ac^4cent a sles .f fans fittin^'^^^to one anoi I' ho ratiop will better than a verbal description this iho eraption of 1912 tiu -' - rrr' .f the church. l»va from Etna flowed down to within three- 07, "^^.a*^ 'itiviX o\ CASTIGLIONE 51 ingenious device for building up a dome that also occurs in one of the churches at Salonica. The nave aisles also show that considerable care was bestowed on the edifice, for instead of the plain barrel vault that might have been expected, the roofs are made of little voutes d'aretes supported on corbels. These deserve careful notice for the sake of the comparison they may afford with similar work elsewhere, and so be the means of fixing the date of this church. The church was well lighted by a three-light window in the western fa9ade and smaller windows in the aisles. The window in the central apse has two lights. There is a large door in the centre of the west front and a smaller one by it to correspond with the north aisle. Beside the corbels and the fan pattern in the dome already referred to, we noticed the following decorative details. Near the niche for the diaconicon are the fragments of a wall painting. On either side of the main apse there were ornamental pillars made by leaving square blocks of lava projecting from the wall and then trimming them up to resemble pillars. The angles of the pilasters at the west side of the choir are recessed to take two other ornamental pillars to correspond with them. Leaving the interior by the main door we have a view of the western fa9ade. The only attempt at decorating seems to have been the undulation of the top instead of making a straight cornice. Buttresses are introduced to mark the nave, and aisles and similar buttresses occur on the north and south sides, and there are two to support the main apse. The chief points of interest in this church are the absence of Norman or Arab features in the architecture and the arrangement of the sanctuary with the central apse and the ritual niches in trefoil. In both respects this church is unique in Sicily, and it is the only building to which the term Byzantine in the strict sense of the word can be properly applied. At the conclusion of these notes on S. Domenico we have described and illustrated the church at' Agro for the sake of comparison, and the difference between the two styles, the pure Byzantine and the hybrid Norman- Arab imitation, is at once apparent. In trying to assign a date to S. Domenico we are confronted with the difficulty that there are no inscriptions, ornaments, 62 SICILY or crosses to assist us. Up to the time of the Norman con- quest there was little variety in Byzantine methods when applied to so small a building. So far as the architecture goes it might be as early as the church at Salonica* or as late as the churches at Eossano or Stilo. But immediately after the Norman conquest the new composite style, familiar to us in the Palermitan churches and seen at Agro, became so popular in Sicily that it seems to have been adopted to the exclusion of every other style, even where as in the Trinita di Delia and at Agro the churches were built for the benefit of Greek monks. S. Domenico would therefore be a notable exception to a general rule. It is likely that documentary evidence relating to this little ■church exists in th^ early Norman archives. Dimarzo's edition of Amico's topographical dictionary in the articles on Castiglione and Prancavilla refers to a Basilian monk named Chremes. It appears that this Chremes petitioned Roger, the Great Count, for leave to found the Monastery afterwards known as S. Salvatore della Placa.t The petition was granted, and the newly established house was endowed under the Norman law with the fief of the locality whence it took its name. The establishment to which Chremes belonged was then described as ancient, and there was a tradition that a * protopapas ' had ruled the neighbourhood so far back as the time of the Emperor Leo III the Isaurian in the beginning of the eighth century. The local historian quoted by Dimarzo dismisses this tradition as a fable and also the legend that Francavilla was founded by Franks in the time of Charle- magne. He might have added that Francavilla was a Eoman and probably a prehistoric settlement, and that Christianity had existed there from the earliest times of our era. The important fact noted by Dimarzo, however, is that under the Norman charter the newly founded monastery received much territory in the neighbourhood of Castiglione, and if S. Domenico existed at that date it is almost sure to be mentioned in the parcels of the charter or grant of land. The inquiry into the Norman archives and the grant to Chremes seems to be well worth following up. I * S. Elias. t See Appendix, p. 171. I The authorities quoted by Dimarzo in his editorial note on Francavilla are : Pirri, Sicilia Sacra, Not. 20, lib. 4. SS. Salvatoris di Placa, Villabianca, Sic. Nob. i6 INTERIOR looking N. W. from chancel to nave : showing vaulting of dome over nave. To face page 52 62 81C1LY or crosses to assist us. Up to the time of the Norman con- ^1 quest there wan little variety in Byzantine methods when applied to so small a building. So far as the architecture goes it • -'''^* !)e as early as the church at Salonica* or as late fts the ^ ' at Kossano or Stilo. But immediately after the Nonaan conquest the new coniposite style, familiar to us in ^ : ;m churches and seen at Agro, became so popular in 1" it seems to have been adopted to the exclusion of • . style, even where as in the Trinita di Delia and o the churches were built for the benefit of Greek mr)!iiv> b. Domenico would therefore be a notable exception to — ai rule. likely that documentary evidence relating to this little exists in the early Norman archives. Dimarzo's edition of topographical dictionary in the articles on (' <\ .villa refers to a Basilian monk named Chremc. . a^„ .. ^i^.v c^rs 1 IS Chremes petitioned Eoger, the Great Count, for leave to i the Monastery afterwards known as S. Salvatore della ' The petition was granted, and the newly established was endowed und^j^^^^^-^glggi^anglaw with the fief of the y whence it took.u^^Lm^-._JI^ establishment to which < iremes belonged was then described as ancient, and there was a ■ a that a ' protopa^ft^^^'^^^Ied the neighbourhood so k as thet!ft5^o^ fe^^^ft^^^ -Sfc^IS^'te^Isaurian in the :;ung of tbg^^mtP a^*i?^.^^^\m^1oWB*orian quoted by ) 'nudrzo dismisses this tradition as a fable and also the legend rancavilla was founded by Franks in the time of Charle- He might have added that Francavilla was a Roman '!)ably a prehistoric settlement, and that Christianity had there from the earliest times of our era. The important : i)y Dimarzo, however, is that under the Norman charter founded monastery received much territory in the • iod of Castighone, and if S. Domenico existed at that dai Umost sure to be mentioned in the parcels of the ch 1 ! of land. The inquiry into the Norman archives "• ' f'hremes seems to be well worth following up.t * S. Elms. t See Appendix, p. 171. J The authoriiiefl q«otfid by Dimarzo in his editorial note on Francavilla are : Pirri, Sioilia S<*cra, Not. 20, Hb. 4. SS. Salvatoris di Placa, Villabianca, Sic. Nob. ^t ^"^"sA ^"-^A. '^^ CASTIGLIONE 58 On the other hand, there are practical difficulties in the way of assigning a pre-Norman date to this church. The Norman conquest, for instance, succeeded an Arab occupation of Sicily that had lasted for over two centuries, and though the Christians were allowed during that time to retain and repair the existing churches, they were not allowed to build new ones. In Palermo, at any rate, the Arabs, like their co-rehgionists the Turks, seized the churches and converted them into mosques. It is, however, a remarkable fact, that with the exception of S. Giovanni degli Eremiti at Palermo, there is not a single church now extant in Sicily that has been changed, first from church to mosque and back again from mosque to church. It is, of course, possible that this fate befell S. Domenico, but I could find no trace of it in the building itself. And as a general rule the Normans made a clean sweep of the old buildings, as they did in England, and started afresh with new ones. Judging by the fabric alone it would, however, not be impos- sible for S. Domenico to have been built either before the Arabs came to Sicily at all or at some time in the early part of their occupation, and before they conquered all the island. From their first landing at Mazzara, in 827, to the fall of Eometta, it took them 150 years to conquer Sicily, and even then the Christians remained in a majority in the Eastern provinces. The only other detail to guide us in assigning a date is the substitution of niches for apses. The arrangement in trefoil is of no help, for so far as we know it is not found in a Byzantine building anywhere else. The only Sicilian examples of the use of niches placed on each side of the main apse are at Palermo in the chapels of the Mare Dolce or Favara Castle, built during the Norman siege, and the Zisa. The little chapel at Constantinople dedicated to S. Thecla, dating from the reign of the Emperor Isaac Komnenos (1057-59), also has niches. On the whole I am inclined to think that this church is not earlier than the Norman conquest ; that it was built by immigrants from the Levant during the revival that ensued,* torn. 2, p. 267 et seq. MS. che conservasi nella Bibl. Com. di Palermo, Q q. R. 64 ; and on the grant of the fee, De Ciocchis Sacro Regice Visitationis decreta torn. 2, p. 454 et seq. Also Notizie di Francavilla by Cav. Vincenzo Cordara. • Vol. i. p. 83. The new Norman law regulating the tenure of land gave the monastic houses a new lease of life and much prosperity. 54 SICILY and that it was probably the church of a BasiHan monastery. But this opinion cannot be considered as final till the site has been carefully explored. The arrangement of the apse and niches in trefoil is inter- esting for another reason. It has been suggested that the neighbouring cell a trichora at Malvagna was built with three apses arranged in this way to suit the requirements of the Greek liturgy. We have never come across such an arrangement in a Byzantine church built in Sicily or Calabria for the elaborated Greek liturgy. But it does exist in the much earlier churches at Sohag and Bethlehem. The arrangement at S. Domenico shows that this plan was actually used in Sicily, and the suggestion with regard to the chapel at Malvagna is quite likely to be well founded. But the cella trichora was not an invention to suit the require- ments of the elaborated Byzantine liturgy, as the Koman baths at Thelepta and Lambessa the churches at Sohag and the very early Christian chapels in Tunis clearly prove. AGRO This favourite carriage excursion from Taormina involves driving two miles up the bed of a mountain torrent, and is only feasible in dry weather. Before starting the traveller should ascertain that the river-bed is dry and practicable for a carriage or he may make the long drive to Sta Theresa la Eiva in vain. And even if the traveller is unfortunate enough to find the river in full spate he will witness a sight well worth seeing. The river-bed or moraine, for that is really what it looks like, is in places nearly a mile wide, and after the torrential rains that come in autumn the water shoots down from the hills above with incredible violence, carrying all before it, undermining or sweeping away the roads and railway bridges, flooding the orange and lemon groves with large deposits of shingle, and doing an immense amount of damage in a very few minutes. Apart from the archaeological interest in the church and the beauty of the surrounding scenery, the excursion is well worth making for the sake of seeing a large Sicilian mountain torrent or fiumara at close quarters. AGKO 55 S. Peter and S. Paul Dimarzo says that the church of the Crocifisso in Forzia d' Agro was turned into a barrack and occupied by British troops during the campaign against Napoleon when Murat tried to invade Sicily in September, 1810. The French then held Scylla and Keggio in Calabria, and on that occasion a detachment of Neapolitan Infantry and some Corsicans, in all about 3,500 men under the command of General Cavaignac, came across the Straits and landed at Mill, near Giampilieri. The English regi- ments engaged were the 2nd Light Infantry, the 21st Eegiment, the 3rd and 4th King's German Legion, and the 20th Light Dragoons. The upshot of this raid, described in the language of the day as a ' singular incursion,' resulted in the surrender of 800 Frenchmen with 50 officers and a stand of colours as against three wounded on our side. An account of this is given in the Appendix to the diary * of my great uncle Captain Wilham Hanson, of the 20th Light Dragoons. Captain Hanson was then on the way from England to join his regiment at Terranuova, near Messina, and he has left an interesting account of his journey from Messina to Syracuse along the coast road. His experiences in Sicily seem to have been much the same as my wife's and mine were exactly a century later. The Sicilian country inns were terribly dirty and bad, the fowls for dinner were half starved, just killed, badly boiled, and very tough, and he and his brother officer ' went to bed with not a few mosquitoes and bugs for companions.' Of Nicolosi at the foot of Etna he says, ' Here there is no inn, and we therefore went to the house of a man thought in this country a gentleman, who, provided they pay him exorbitantly, is ready to allow Englishmen shade and house room, for little else was to be had : his name is Giamellaro ; and this same man contracted for a certain sum (of part of which I am a subscriber) to build a house this last summer for the accommo- dation of travellers on Etna.' t On the way from Augusta to Syracuse he passed by the Torre di Marcello near Priolo, describ- * Short Journal of a Voyage to Sicily, 1810-1811, by Captain William Hanson, of the 20th Light Dragoons. London ; J. Darling, 1814. t This, I think, must be the Casa Inglese, and he refers to it again when he ascended Etna. 66 SICILY ing it as a large pyramid of stone marking the place where the Syracusans defeated the Athenians. The little church dedicated to S. Peter and S. Paul stands on a bluff or spur of the mountain-side jutting out into the bed of the fiumara, and seen at a distance it looks like a small fort with battlements round the roof. The plan shows that it is a rectangular structure very narrow in proportion to the length and height. It is divided into two parts, a nave with lateral aisles and a chancel with three apses. Of these apses the main or central apse is square on the plan and the prothesis and diaconi- con are semicircular. In front of these apses there are three rectangular compartments representing the chancel in the centre and the lateral aisles on each side of it. The nave is divided into three bays of equal size and separated from the lateral aisles by four pillars, two on each side. At the west end there is an open narthex outside the main door, and this is flanked by staircases built in the prolongation of the aisles. The arrangement for roofing the different parts of the church is very peculiar. Commencing at the east end we found that though the central apse is square on the ground-plan it takes a semicircular shape about two metres from the ground on the inside, the break from square to half-circle being negotiated by little squinches in the angles. The semicircular shape is then carried up to the top of the apse and is covered by a semi-dome built up on little squinches or semi-domes arranged clusterwise after the Arab manner. The apses of the prothesis and diaconi- con are semicircular within and without and covered by plain semi-dones. The rectangular spaces in front of the prothesis and diaconicon are covered by cross vaults, and the chancel between them is carried up square to the roof and then covered by a little oval dome supported on the same cluster squinches of Arab origin. The central bay of the nave is also covered by a rather larger dome supported on squinches of the usual and simple shape, while the two other bays east and west of it are covered by timber and tile roofs. The aisles of the nave have cross vaults. All the interior of the church is smeared over with white- wash excepting the four monolith pillars of grey granite and their stone caps. The latter are carved to represent bunches of acan- i8 AGRO. Abbey Churchy S. Peter and S. Paul. INTERIOR sJvowinf^ honey comb vaulting in the apse: and squinches supporting the central dome. The porch xvith Greek inscription above. To face page _^6 SICILY ing it as a large pyramid of stone marking the place where the Syracusans defeated the Athenians. The little church dedicated to S. Peter and S. Paul stands on a blufif or spur of the mountain-side jutting out into the bed of the fiuuiara, and seen at a distance it looks like a small fort with l)attlement8 round the roof. The plan shows that it is a rectangular structure very narrow in proporti(^ i^th and height. It is divided into two parts, a nave ,*.i^i- . ''-iles and a chancel with three apses. Of these apses th or central apse is square on the plan and the prothesis and diaconi- r ;n are semicircular. In front of these apses there are three '• '^o'lar compartments representing the chancel in the centre .ateral aisles on each side of it. The nave is divided into bays of equal size and separated from the lateral Aiftles by 'iUars, two on each side. At the west end tl • X outside the main door, and this is fla>'^ m the prolongation of the aisl^^^i^tv The arrangement fo^^^^e^figgvite 4JgQf e^t ^^^^to^iiS^church IS very peculiar. Commencing at the east end we found that ' ■ -"^'h the central apse is square ^ifii\1fVS^'^"V^nd-plan it takes a ircular shap^i^d<»W^^r»>f%^^l¥W?ii^«6^ ^i^n^a^WP^e iii'^ide, the break from-^?k^^'^>:;3^^cm»).?^^?i^*^^^??^«^ ■ squinches in tH^^gF^^^TH^ h^^^r^m- ^^H^ then ^>d up to the top of the apse and is covered by a semi-dome . up on little squinches or semi-domes arranged clusterwise %ftw the Arab manner. The apses of the prothesis and diaconi- are semicircular within and without and covered by plain dones. The rectangular spaces in front of the prothesis Haconicon are covered by cross vaults, and the chancel them is carried up square to the roof and then covered " oval dome supported on the same cluster squin ' m. The central bay of the nave is also covered by a tie supported on squinches of the usual and simple shap» the two other bays east and west of it wre covered by tiuiijci and tile roofs. The aisles of the nave hftve cross 'aults. AH th* «?Tt<>rior of the church is smeared over with white- ■i four monoHth pillars of grey granite and their btiiuc cap.-., i h*j Utter are carved to represent bunches of acan- AGRO 57 thus leaves of the usual and conventional Norman pattern. The arches they support are all pointed, and so are the windov^^s in the clerestory and the aisles. Passing now into the porch and looking back into the church through the main door v^^e get a distant view of the timber roof of the third bay and the squinches supporting the dome over the middle bay of the nave. The door itself has a pointed arch, and in the tympanum is a cross in a circle with circular ornaments in the quarters made of coloured stones. The Greek inscription on the lintel is hopelessly abbreviated, but sufficiently clear to be legible in the larger photograph. The text seems to be as follows : * {First line) avtK(ai)vi(T{dri) 6 va(oc) ovt(o(t) tu){v) ayi{wv) aTr{oaT6X(i)v) T[t{Tpov) K{ai) 11 (auXou) Tr{apa) 9e{o)(TTri (Second line) piKT{ov) K{a)dayovfi{evov) tov ravpo/xevirov inrb oIkh {Third line) mv avaX(t)paT{u)v) fivngOnt} avT{ov) K{vpio)g erei g\ir {Soffit) o Trph)Topai(TTop Tipapcog o (ppajKog Rendered into English this inscription is as follows : ' Was renewed this temple of the holy Apostles Peter and Paul by Theoste- riktes Abbot the Tauromenian at his own expense may remember him the Lord in the 6680 AM. (1172, a.d). The master builder Girard the Frank ' The combination of brick and stone and the general appear- ance and colour of the outside of the church are very pleasing. The view of the east end shows that the central apse is carried square up to the top of the church while the prothesis and diaconicon are semicircular. Notice the bands of bricks arranged in the herring-bone pattern and the method of finishing off the flat buttresses by interlaced arching. This is the style adopted in decorating the apses of the cathedrals at Cefalu, Monreale, and Palermo, but here this decoration is carried right round the church. The view of the west end shows more pieces of herring- bone pattern, and the peculiar porch or narthex in the fa9ade. * I am indebted to confrere Mr. Cyril Davenport, the Assistant Librarian of the Order of St. John of Jerusalem in England, for obtaining the transcript of this inscription. Salinas, Notizie degli Scavi, January, 1885, pp. 87, 88. 58 SICILY The view of the south side shows a door, now blocked up, that corresponds to the central bay of the nave. It needs but a glance at this building to see that there is nothing Byzantine about it except the arrangement of three apses, the Greek inscription over the west door, and the orna- mental device known as the herring-bone pattern. The squinch and pendentive arrangements are copied from Arab buildings, and for the rest the church is purely Norman in plan and archi- tectural details. Another building of the same date and much the same style but rather larger, known as the Badiazza, at Messina is now partly embedded in a fiumara in the mountains behind the city ; it was severely shaken and considerably injured in the earthquake of 1908, when a part of the roof fell in. The capitals of the pillars deserve notice, and like the rest of the building they are purely Norman. The only Byzantine feature about this edifice is the triple apse, and like the church at Agro it was built for Basilian monks. MESSINA In the church of the Catalani at Messina we have another example of a building riven and split from top to bottom in an earthquake but kept together by the dome. The accompanying illustration is from a photograph taken after the debris of the surrounding houses had been cleared away. The church is of late date, and was built after the Norman Conquest. I have already referred to it in the Introduction. CASTELVETBANO Trinita di Delia In the preceding volume I have given two illustrations of the outside of this little chapel showing the west front and the triple apse at the east end. On the wall of the central apse outside, about 2J metres from the ground, I noticed a few Greek characters cut in the stones, but can make no sense of them. The interior of the church and also of the low crypt below have been thoroughly renovated ; the crypt is a little cruciform chamber SICILY 59 with a cross vault for the intersection and barrel vaults for the transepts. In the church itself two of the large pillars are made of red granite, and the little recessed pillars are made of white marble, and are quite modern. Only one of the capitals of the large pillars seems to be old and it is certainly Roman ; the others are probably Norman imitations. ALTAVILLA The ruins of the Norman church said to have been founded by Guiscard and Roger for Basilian monks, stand on a bluff overlooking the Palermo-Messina railway about two kilometres east of Altavilla station towards Termini. The church was made of blocks of stone, and presents even in its ruined state a very substantial appearance. The ground plan shows that it was a cruciform building with one large semicircular apse at the east end of the chancel and semicircular apses on each side of it built in the east walls of the transepts.* There was a crypt underneath the chancel, and numerous fragments of carved stone that have been stowed away there show that the church was Norman. A peculiar square recess in the wall between the chancel and the apse on the south side deserves notice. SYBACUSE The Cathedral There is a tradition that the Doric temple at Syracuse now used as the cathedral was first converted into a Christian church by Bishop Zosimus, who occupied the See shortly before the Emperor Constans II. came to take up his abode in Sicily, in the middle of the seventh century. This tradition rests on no very sure foundation, and judging by the dedication to the Blessed Virgin under the style of Maria Theotokos, the conver- sion might well have been two centuries earlier. In 1911 permission was obtained to remove the plaster and to make a thorough examination of the cathedral from floor to roof, and in January, 1913 and 1914, the work of cleaning the interior was in active progress. * See the plans of the churches at Poictiers and Fontevrault. 60 SICILY It will be seen from the sketch-plan that the nave and choir occupy the room of the cella of the temple, and that the aisles were made by filling in the spaces between the great pillars with masonry, and so making an outside wall. The walls of the cella were then pierced with eight arched openings, and carried up to provide a clerestory room to light the nave. The roofing was effected by barrel vaults in the aisles, and by timber and tiles for the body of the church. The west end was walled up in the same way as the north and south sides, and probably had a narthex, but the exact arrangement is uncertain^ because an elaborate west fagade with a porch was built up in front of it during the baroque restoration. The major part of the east end also suffered the same fate at that time, when a large square chapel was fitted up to the end of the choir, and a smaller chapel was built out at the south-east corner ; but as a semicircular apse covered by a semi- dome was discovered at the east end of the north aisle in the place of the chapel of the prothesis, it seems probable that the church had three apses, that the south aisle had a similar apse to correspond, and that the present choir terminated in a larger apse of the same shape. It will be noticed that this semicircular apse in the north aisle is built up between the pillar that stood at the north-east corner and the second pillar of the east front of the temple, and if the contour line be prolonged, it will be seen that the temple ended about midway down the present choir. The footings of the eastern extremity of the cella were found under the steps of the bishop's throne, so that the present nave corresponds roughly with the cella. We noticed the following peculiarities : there are five clerestory windows on each side of the nave ; four are grouped together in the western half of the nave, and the fifth, which is rather larger than the others, occupies a place by itself over the second pilaster from the choir. The photograph of the window on the south side shows a large arch above it spanning two bays. As this arch has no connexion with the present roof, it seems probable that it supported an earlier and lower one either of the kind found in S. Spirito at Palermo, or a cross vault or a dome. The same picture shows a portion of the timber roof supported on wooden corbel figures. The latter are unmistakably Norman. SYKACUSE Gl This roof runs uniformly the whole length of the church from the west end to the eastern extremity of the choir ; it has been repaired and restored at various times, and the portion over the choir is now concealed by a ceiling of plaster. The small picture of the window in the masonry between the Doric pillars is taken in the north aisle. These windows are made of finely dressed stone, have rounded heads and, as the section shows, are made on the bevel in the Norman way. The rim or flange on the outer side was no doubt intended to take an open-work stone lattice or screen. There seems no reason to doubt that prior to baroque restor- ation the cathedral had the three apses incidental to a church intended for the Greek liturgy. As the Normans usually adopted this form of building the sanctuaries for the churches they restored or built in Sicily after the conquest, this triple-apse arrangement might certainly be as late as the eleventh century, when, as the roof and other parts of the fabric now show, the cathedral must have been extensively restored. One would like to know whether this restoration was the result of the great earthquakes that occurred shortly after the conquest in 1140 and 1169, and whether the displacement of the Greek pillars in the north aisle was also the result of them or of an earlier earthquake that took place in the sixth century. The difiicult question, however, in regard to these apses is rather not how new but how old they are likely to be. Were the building in Constantinople or the Levant instead of in Sicily, we should expect to find the triple apse as a matter of course. But the Greek element in the Sicilian Church did not begin even to pre- dominate till the Emperor Constans came to Syracuse, two years after the bishop's death,* and fifty years more were to elapse before it passed from the jurisdiction of the Bishop of Kome to that of the Patriarch of Constantinople and the Latin liturgy gradually gave place to the Greek. The evidence from the works now in progress seems to show that the fabric inserted between the pillars of the temple is in the present form Norman, and that the early Christian work on the cathedral fared at the hands of the conquerors as that of every other ancient church in Sicily did. * Zosimus in 661. 62 SICILY S. Marcia7i's Chapel Two very different opinions are given as to the age of the marble capitals in S. Marcian's crypc. According to Barreca and many archaeologists they are early Christian, while other and more recent authorities attribute them to the Normans. In point of fact the capitals consist really of two parts, a little Ionic capital below and a rectangular abacus above ; this combination was much favoured by the early Christians and Byzantines, and con- tinued in use for a long period. We are inclined to think that the truth lies somewhere between the two opinions ; that the capitals are ancient, that is, either Koman or early Christian, and that the abaci are very much later. We found four out of five of these little capitals used with the abaci above to decorate the angles of the square hall of the crypt ; the fifth has been turned upside down and placed on a square block of stone near the altar, and is now described as S. Marcian's throne. We think there is little doubt that these capitals are early Christian, that they originally stood on granite shafts and belonged to an older church that was destroyed either by the Sara- cens or by earthquakes in the twelfth century. In the restoration made in Norman times, four out of the five were appropriated to serve their present purpose, and the granite pillars were used as old materials to make up the walls. The abaci deserve more particular notice, and it will be con- venient to take them in the following order, premising that they stand about nine feet from the ground, and that as they are embedded in the angles of the crypt nave only two sides of them are visible. The Lion is at the angle marked (a) on the plan ; quite close to it are pieces of a granite pillar that originally supported this capital. We found that the diameter of this pillar and of the capital now called S. Marcian's throne corresponded exactly. The inscription on the sinister side is from S. Mark's Gospel, c. i. V. 5, as follows : ^ SECVNDVM MAKCVM. ET EGEEDIEBATVE AD EVM OMNIS IVDEAE EEGIO ET lEKOSOLVMITE (sic) VNIVEESI ET BAPTIZABANTVE AB EO IN lOEDANIS FLVMINE CONFITENTES PECCATA SVA, 30 CAPITALS In the Crypt of S. Marctan^s Chapel at Syracuse. EMBLEMS of S. Luke and S. John. To face page 62 oc 62 SICILY S. Marcian's Chapel Two very different opinions are given as to the age of the marble capitals in 8. Marcian's crypc. According to Barreca and manA '^ts they are early Christian, while other and more recei ..jus attribute them to the Nc-'^-^"^ In point of f^rf -iitals consist really of two parts, . l^mic capital and a rectangular abacus above ; this corabin.ttion was favoured by the early Christians and Byzantines, and con- 1 in use for a long period. We are inclined to think that truth lies somewhere between the two opinions; that the '.-•ituls are ancient, that is, either Roman or early Christian, and ' the abaci are very much later. We found four out of f'^'f' '^f these little capitals ur^^'' ^r'^^h the i above to decorate t of the square hall o ot ; th» fifth has been tnmed upside down and , are 'k of stone 9ear the altar,, and is now described as . .... ae. We think there is little doubt that these l..j,...».., ,..^ early Christian, that they oii^lHi,lf^^lbod on granite shafts and belonged .4@«!isiderable and continuous intercourse between Syracuse and the I .atin church of Africa, and more especially after the Christian communities in Africa were finally broken up by the Arab con- quest of Carthage ; it was mainly in South-eastern Sicily that many then sought refuge. We find some indirect confirmation of our view from the fact that things Latin were favoured in Sicily for longer than might have been expected ; so for instance at an early date the Sicilian bishops protested against Greek innovations introduced by the Pontifif Gregory the Gr^k> again at a later date the i^nw>f tOtJU^o HI- thought it necessary to forbid all communication between the bishops in Sicily and the KomanCiina?''^^^ ^"^ ^^^"^^^ i'«v.n^^U .^ ^e \<^h^0 ^^ «\ But we find two difficul^M>?^^M%ting Barreca's view as t<^ these ^^^^^.'^^^T^if^fil^l^lt^^ tee^^lS^^Wefe ikf^n'^T&Msaiit^. a*^ two Evangelist emblems, the Eagle and the Lion, preserved over the main door in the church of S. Maria in Cosmedin at Kome. These are generally supposed to belong to the twelfth-century restoration of that church. The positioh they now occupy, however, may not be the original one, and it is not one we should expect to find them in ; we are therefore prepared to hear from further research that in company with many other frag- ments discarded in the twelfth century, they belong to the restoi ation undertaken in the Pontificate of Hadrian I. (772-795). At any rate they are facsimiles of the examples in S. Marcian's chapel \nd both belong to the same date. '^' (^ second and more serious difficulty arises in the style of . voters or lettering of the inscriptions. It is quite true that the little leaf-shaped rubric seen in the eagle inscription is frequently found in early Christian inscriptions in Africa and in ' ? that apart, we know of no characters in the Christian :-....,.. ! Xorth Africa, ranging from the beginning of th( ^^^ SYEACUSE 65 fourth century to the end of the seventh, or indeed elsewhere of that date, that bear any resemblance whatever to these in S. Marcian's chapel. The latter seem to us to be too well executed for the sixth century or the seventh, and they appear to resemble more closely the characters used from the tenth century on till the Norman conquest. But we think that the abaci are anterior to the Norman conquest, probably by at least half a century or perhaps even more. From the texts themselves no evidence can be derived. The mistakes in spelling count of course for nothing, and the sequence of words follows the Vulgate. The Basilica hij S. Marciaii's Chapel The capitals in the crypt are not the only archaeological puzzles in the precincts of S. Marcian's chapel. It has been usually supposed that two Doric pillars in the nave of the mediaeval church of S. Lorenzo belonged to a pagan temple. The church of S. Lorenzo, like the modern French chapel at El Kef, occupies the narthex of a much older church that for convenience I will call the basilica. As the plan shows, the axes of the basilica and the church were consequently at right angles to one another, and the west front of the basilica decorated with very rough carvings now forms the west side of the church. It was eventually determined to excavate the site of the basilica, and the discoveries then made leave no room for doubt that these so-called temple pillars are in reality Norman imitations. This is clearly shown from the heads that decorated the corners of some of the ' Doric ' capitals found in the debris, and from a» tongue pattern on the base of the pillars. It was found, moreover, that the capitals were made not in round drums, but in segments of circles. The site has not been fully explored, but in the meantime the apse has been uncovered and the arrangement of the dais or floor show that there was an earlier church. There may be some difficulty in fixing the date of this earlier church as well as the Norman restoration on account of two great earth- quakes in 1140 and 1169 ; but it seems likely that the Norman church was destroyed in one of them, and that the site was then abandoned for some years till the later church of S. Lorenzo was 66 SICILY built up in the narthex. The south front and the graceful cloister porch of S. Lorenzo are familiar objects to every traveller who visits the catacombs, and too well-known to need description here. The archaic figures over the old west door of the Norman basihca deserve notice, and should be compared with similar figures on the early church at Aregno in Corsica. SOUTH ITALY S. Giuseppe at Gaeta S. Costanzo at Capri S. Maria delle cinque torri at Monte Cassino 1 HAVE mentioned these three churches in the notes on the Cattolica at Stilo and S. Mark at Kossano in Calabria, because in a sense they are all built on the same plan, and the architects of the first two obviously intended to secure a cruciform effect inside with a dome for the centre of the cross. But the absence of the triple apse * distinguishes them from the Sicilian churches of the Norman period like the Cappella Palatina and the Martorana, and shows that they were not built to suit the requirements of the Greek liturgy. The term Byzantine is commonly applied to all of them, but with the exception of the two Calabrian churches and Castiglione it is altogether misleading, as well from the architectural as from the liturgical point of view. GAETA S. Giuseppe or S. Giovanni al Mare The history of mediaeval Gaeta is well known from the early charters and records published at Monte Cassino. f Constantino * Schulz shows three apses at S. Maria, but I think he is wrong. t Codex Diplomaticus Caetanus in Tabularium Cassinense. Monte Cassino, 2 vols., 1889-91. For the general history, Memorie Storiche delta Citta di Gaeta, by Gaetani d'Aragona, 1879 ; Chalandon, Histoire des Normands, vol. i. p. 4. 68 SOUTH ITALY Porphyrogennetos reckons it among the cities dependent on the theme of Sicily,* and hke Naples, Amalfi, and the seaport towns in Sardinia, it belonged in theory to the Emperor at Constanti- nople ; but in practice the city was left in a great measure to manage its own commercial affairs and develop local self-govern- ment under a magistrate called hypatos, or consul, who occa- sionally bore an imperial title, and whose office eventually became hereditary.! "When the relations with the Empire were more than usually close the magistrate adopted the years from the Emperor's accession for dating his official acts. Duke John, the supposed founder of this church, was the successor of a certain Count Anatolius, sent to Gaeta as * consul ' by the Pontiff Gregory II. at a moment when the city had rebelled against an unusually heavy burden of Imperial taxes. The precise character of Anatolius' mission seems to be uncertain ; possibly Gregory II. did not intend to do more than protect the large estates of the Church in the neighbourhood, but this mission was the first step towards the Patrimony near by passing under the control of the Dukes of Gaeta, and so the origin and ground for the claim made by John XII. to temporal authority over the city in place of the Byzantine Emperor. Count Anatolius came upon the scene in the reign of the Emperor Leo the Isaurian, about ten years after a Council of the Church held at Eome had condemned Iconoclasm, and con- sequently at the beginning of the great change in the political and ecclesiastical relations between Eome and Constantinople that ensued. He was still ' consul ' during the creation of the Pontifical dominion, when Gaeta though nominally dependent upon but virtually abandoned by the Byzantines, became hemmed in between the Lombard States and the newly created Roman territories. From these powerful neighbours the little commercial community maintained its freedom, sometimes with the help of Naples, at other times by an alliance with the Saracens, and occasionally by resuming the position of subject to the Emperor at Constantinople. There is evidence that Naples and Gaeta had already joined together while Anatolius held office. In 757, the Pontiff Paul * Gaeta nella Storia, Raimondo Vento, 1911. t The same thing happened in the provinces of Sardinia. See vol i. p. 52. GAETA 69 wrote a letter to King Pepin, calling upon him to constrain the men of Naples and Gaeta to restore the Patrimonies of S. Peter situated at Naples, and to allow their bishops to come to Eome for consecration.* The two complaints made in this letter respecting the temporalities and the spiritual prerogative are the same that were made in regard to Sicily and Calabria when the Emperor Leo III. confiscated the Patrimony. At that time the revenues there were diverted into the Imperial exchequer, and the bishops were put under the Patriarch of Constantinople and forbidden to communicate with Rome. The Patrimony in the vicinity of Gaeta remained in fact under the control of the Dukes of Gaeta for more than a century later and the retention was, as already said, one of the grounds for the claim made by the Pontiff John XII. to temporal authority over the city.f In practice these Patrimonies scattered about Italy, Sicily, Africa, and elsewhere, gave the Pontiff extensive power and influence as landlord, and the rectors who managed them took a leading part in the ecclesiastical administration of their respective districts. We are not told what side the Gaetans took in the iconoclastic controversy. The authorities consider that in Italy the clergy and laity were as a rule opposed to iconoclasm, and that for political reasons the Emperor Leo and his successors did not press their religious views upon their Sicilian and Calabrian subjects. The local historians, writing from the Italian point of view, claim that the original secession of the citizens from the Empire arose from their opposition to the Iconoclasts, t Their allies, the Neapolitans, on the other hand, seem to have taken a different side in the controversy, for they refused to receive a bishop named Paul, a well-known enemy of the Iconoclasts, sent to them in 761 by the Pontiff Paul. § It seems to me likely that the refusal of the Neapolitans and Gaetans to recognise the spiritual prerogative and allow their bishops to go to Rome for consecration arose more out of questions of language, nationality, and finance than of * Hodgkin, Italy and Her Invaders, vol. ii. p. 263. t Gay, L'ltalie Meridionale et VEmjyire Byzantin, vol i. p. 292. X Gaeta nella Storia, p. 13. § Chalandon, vol. i. p. 10. 70 SOUTH ITALY faith or image- worship. In importance, at any rate, the Greek element certainly prevailed over the Latin in Naples and Gaeta at this time. For example, we know from a letter written by the Pontiff Hadrian to the Emperor Charlemagne about the Patrimonies in these parts, that the Viceroy of Sicily had come to reside at Gaeta in 778.* Later, in 812, the Gaetan ships were placed under the order of another Sicilian Viceroy, and helped the admiral of a Byzantine fleet.! Again, later, in 830, the acts of the city are dated by the years of the reign of the Emperor Theophilus. I During the ensuing century the principal event of importance to Gaeta as a dependency of the Byzantine Empire is the con- quest of Sicily by the Saracens. By the capture of Palermo the Arabs obtained the command of the Tyrrhenian sea, and after the fall of Syracuse the Byzantine seat of government was moved first to Calabria and afterwards to Apulia. The Gaetans and their neighbours at Naples and Amalfi were then left to shift for themselves and hold their own as best they could. At this time Southern Italy was divided up into the following territories. Calabria and the province of Otranto belonged in part to the Saracens but mainly to the Byzantines. Between these districts and the Eoman State the Lombard principality of Beneventum occupied the eastern side of the Appennines ; on the western side were the duchies of Salerno and Capua, and on the seashore the three Greek republics. After the fall of Palermo and the withdrawal of the Byzantine fleet, the coast of Italy from Civitavecchia to Messina was left exposed to the mercy of the Saracens. It is chiefly in connection with their exploits that we hear of Gaeta, and they became in turn besiegers, alHes, and finally tributaries of the city. In 846 they were defeated in a battle near Ostia, and an unsuccessful siege of Gaeta, that lasted for two years, was raised by the assistance of a NeapoHtan fleet. But it is beyond my purpose to trace the varying fortunes of these raids or of the feudal wars between the rival Lombard princes that made combined action against the Mahometans for a time impossible. • Hodgkin, vol. viii. p. 63. f Gmj, vol. i. p. 22. I Chalandon, vol. i. p. 8. GAETA 71 The next events of importance in the general history of the period that affected Gaeta were the reconquest of Bari by the Greeks and the accession of John VIII. to the Pontificate. Upon this great prince-bishop devolved the duty of championing the cause of Latin Christendom in South Italy, and during the ten years of his reign he laboured strenuously and unceasingly to reconcile the conflicting interests of his neighbours and bring about an alliance against the Mahometans. His efforts were only partly successful at the time, because of the hostility of the Lombards, and the three republics, who had recently come to terms with the Saracens, refused to join it. The upshot of these negotiations, so far as the Gaetans are concerned, seems to have been the cession of the territories of Fondi and Trajecta to the hypatos of Gaeta and the removal of the Saracens to the River Garigliano, where we find them first as vassals and later on as allies of the Gaetans in a battle against the Neapolitans and the Lombards. But by skilful diplomacy John VIII. induced the Byzantine Emperor to send the Imperial fleet back to the Tyrrhenian sea, and this important event marks a turning- point in local history.* It fell to the lot of a successor and namesake, John X., to accomplish the task, and the joint action of all the Christian States at the battle near the Garigliano drove the Saracens out of Western Italy altogether. The Gaetans were rewarded for their prominent share in this campaign by a con- firmation of the grant of Fondi and Trajecta and the hypatos, Duke John III. received the title of Imperial Patrician from the Emperor. The Christian aUiance was, however, soon dissolved, and we next hear of Gaeta in connection with the doings of three leading characters in the history of the tenth century, the Pontiff John XII., and the German Emperors, Otto II., and his son. Otto III. John XII. claimed the suzerainty over Gaeta in virtue of the cession of the territories of the Patrimony made to the city by his predecessors John VIII. and John X., in * Gay, vol. i. pp. 123 and 124. The Byzantines had captured Bari in 876. This was a most important event in the history of Southern Italy, and more than com- pensated the Byzantines for the loss of Syracuse two years later. The Byzantines sent two naval expeditions to the Gaeta Naples coast in 879 and 880, and won a victory over the Arabs near Messina (id. p. 112). 72 SOUTH ITALY the circumstances already narrated. As he was on friendly terms with the Byzantine Court we may, I suppose, infer that the Emperor had by then ceased to claim any authority over the city, and in point of fact the successor of the Duke John III. had dropped the title of Imperial Patrician, and the acts or charters were dated either according to the Pontifical reign or that of the Duke. The claim did not, however, avail to protect Gaeta from falling into the hands of the Lombards a few years later. In the meantime Gaeta was visited first by the German Emperor Otto II., and shortly afterwards by his son. Otto III. The former came * with his wife, the Byzantine Princess Theophano, and was received by the Duke Marino. The latter f was received by the Duke John and his wife, the Duchess Emilia, and on that occasion the principality of Pontecorvo was added to the Duke's dominions. And that brings me to S. Giuseppe, or more correctly S. Giovanni al Mare. The local historians attribute the foundation of the church to this Duke and his consort at the close of the tenth or beginning of the eleventh century, and I must be content to take their word for it, for I can find no mention of a church of S. Giuseppe in the Codex Dlplomaticus Caetanus. The earliest reference to the parish of S. John * al Mare ' occurs in a deed of conveyance of some private land dated in 1277. , At least three other churches dedicated to S. John are mentioned in the charters of the ninth and tenth centuries. A note by the editor of the Codex connects one of them with the Duchess Emilia, but it was apparently situated outside the town. Another was situated in the village of Formia, about five miles out on the road to Naples. And lastly, the recitals in a document relating to the third church of S. John, which was in the city, are insufficient to identify it with our church. The * Probably 981, when the Emperor went to Naples (Gay, vol. il. p. 332.) This Emperor was married in Rome on April 14, 972 (p. 319), and died in December, 983. He was then twenty-eight years old (p. 341). His son, Otto III., was not crowned till 996 at the age of thirteen (p. 371). t In 999 Otto III. had been on pilgrimage to Mte. Gargano. It was on this occasion that he was received by the famous monk, S. Nil (Gay, vol. ii. pp. 372, 381). GAETA 73 plan, elevation, and general style are not inconsistent with the date assigned by the local historians, and • that is all that can be safely said. The church lies SE. and NW. between the sea and the foot of the rock crowned by the modern citadel : the sanctuary is consequently at the north-west end, and the main door faces the seashore. The ground-plan is that of a little basilica with a single apse, and as the ground slopes slightly from the door the floor is raised by steps placed at intervals up to the sanctuary. The accompanying sketch of the outside gives an idea of all that is visible from the street, the chancel and transepts being shut in by the adjoin- ing houses. The view is taken from the south side, and shows the modern west front, the wall of the south aisle, and the wall of the nave above, with two clerestory windows now blocked up. The little circular tower with a round top contains the cupola or dome of the usual high pitch found in Byzantine and Norman buildings of late date. The bell-cot and the west front appear to be modern additions. Passing now through the west door we get the general view of the interior shown in the next photograph. The church is divided into nave and aisles, separated by slender columns made of granite. To secure a cruciform effect the architect adopted the usual plan of varying the roof levels. Thus the nave and the central bays of both aisles are raised to the full height of the church and covered with a barrel- vault, while the other bays of the aisles are covered with cross vaults at a lower level. This break in the aisle vault produces the effect of transepts, and the intersection of the latter with the nave is covered with the little dome. The third picture is taken looking upwards, to show that the dome is supported on squinches in the angles of the square. This important architectural detail of Arab origin identifies S. Giuseppe with the churches in Palermo, and made me think that it was not built, at any rate in its present form, till after the Norman conquest, and much later than the date assigned to it. Upon this question the plan and architecture of the church and the decorative details now visible throw no light. The single apse points to the archi- 74 SOUTH ITALY tect being a Lombard or an Italian, for the Greek architect would probably have built three apses, as his compatriots did in Sicily. But the two methods of building a sanctuary according to the service to be performed were used con- currently over a long period and are no help to fixing a date. The decorative details are now covered up by plaster and barroque work of the eighteenth century, but where a small piece has been removed two layers of wall paintings one above the other have been found underneath it : the amount at present uncovered is, however, too small to be of any practical assistance. Neither do the granite pillars and their capitals help, for they were obviously taken from a classical building. The capitals are made of stone, and much damaged, and appear to be late Roman work. One of them is used as a base for the last pillar on the north side. The only other ancient object in the church is a carved stone slab used for an altar front in a chapel on the north side of the church. Schulz gives a drawing of it in his book but he does not mention it in connection with the church, and the numbering of the plates in his letterpress has been confused. It is a remarkable composition, and a comparison of the work on the cross and the two supporting griffins suggests that it is not all of the same date. The cross and the little ornaments are certainly designed by a Byzantine artist, and as they are similar to the work on carved screens in the churches of Sta. Sabina and Sta. Maria in Cosmedin at Eome, probably belong to the eighth century.* An artist of the tenth century would hardly have produced two beasts so well executed. I believe they are Roman, and that the cross is a later substitution for some pagan device escutcheon or emblem to which the griffins acted as supporters. Nothing is known locally about this slab. Schulz, to whose picture we have already referred, visited Gaeta, and gives a ground-plan and elevation but no historical information about this church. He points out, as the fact is, that the old cathedral at Capri and the churches of S. Andrew at Trani, and S. Nicholas and Cataldo at Lecce are like it. The * S. Maria in Cosmedin was restored and redecorated in the Pontificate of Hadrian I., 772-795. See similar beasts as supporters on a sarcophagus in the Aries Museum. Gaeta. " S. Giuseppe. Capri. S. Costanzo. To face page y^ 71 m ace cnn datt teci . iOmbarJ Italian, for the Greek arci wouid probably have built three apses, as his compatriots did t the two methods of building a sanctuary the service to be performed were used con- ver a long period and are no help to fixing a tiecorative details are now covered up by plaster .0 work of .tiwk^Bighteenth centiiiy, " - -pre a ' i^! bee%«^^§ia*Oe^ two layers of v. •--- other bavei-vbeen found undern* is, however, too sma ither do the granite piiiais i!.%ufefiitlO^QK ^^\*^ere obviously taken from i ^ \icliolas and Cataldo at Idoes [ling lis a the il..f unly the and hth [or some pagan u IS acted as supporters. It; ot h. Aaidrew at are like it. ^^ - Mu»cvuii. in Cosm 1-7913. Si- ll and redecorated in the Fontifi«ate of as supporters on a sarcophagus in the Ailes V\ ^"^«i\ ^^ft\ oT CAPKI 75 date of the last, built by Tancred Count of Lecce and afterwards King of Sicily, is known to be 1180. A conjecture that the church at Trani and this church at Gaeta were also built by Tancred, or under his auspices, presents itself temptingly, for these three cities, Lecce, Trani, and Gaeta were all in turn honoured by Taucred's special favour. So Lecce apparently remained part of Tancred's royal domain, the citizens of Trani received com- pensation for damage sustained during the war, and an important charter was granted to Gaeta, conferring privileges and remitting certain obligations to furnish men for the galleys.* But there is no evidence to support it, and though, prima facie, S. Giovanni appears to be a Norman building of the eleventh or twelfth century, there is nothing in the style of architecture inconsistent with the traditional foundation, or more likely restoration, of the church by the Duke John about the year 1000. I may add in conclusion that the Duchess Emilia survived her husband and became regent to her grandson during his minority. An alliance between Gaeta and Naples made at this time against the Lombards brought about the downfall of the Ducal family, and the city passed first under the dominion of Pandolf, Prince of Capua, and shortly afterwards under that of the Normans. The Ducal title then passed to the Hauteville family, and my wife's kins- man, Geoffrey de Riddell, was appointed by Count Eoger as the first Norman Governor, with the title of Consul, and ruled over Gaeta from September, 1072, to April, 1077. CAPBI S. Costanzo This little church is situated on high ground at the north end of the Marina Grande in a district that is full of Roman remains. In an adjoining garden we were shown a beautiful shaft of Egyptian onyx about nine feet long, and two large blocks of uncut green serpentine that had just been unearthed. These and a number of other pieces of marble found in the same spot were perhaps the stock of a Roman marble mason. The present edifice consists of two parts, the original church • Chalandon, vol. ii. pp. 62, i48, and 449. 76 SOUTH ITALY and the annexed buildings that were added in the twelfth century or later. In the accompanying plan the former is indicated by black and the latter by red lines. It will be seen from this plan that the axis of the original church is almost due north to south. It was a rectangular building consisting of a nave divided into nine compartments, a single apse for the sanctuary and altar at the north end, and a double narthex at the south end. In the reconstruction of the twelfth century the axis of the church was, so to speak, turned about, a large square hall to contain the altar was built out on the west side, a door was cut in the east wall to make the main entrance, and the original apse and the doors lead- ing into the old narthex on the south side were walled up. The alterations then made subsist to the present time and the church is arranged, so far as the altar and sanctuary are concerned, as it was in the twelfth century. The reader who has studied the plans of the Cattolica at Stilo, of S. Mark's church at Eossano, and S. Giovanni al Mare, other- wise S. Giuseppe at Gaeta, will be familiar with what may be called the nine-compartment division, and notice that the ground- plan and roof arrangement of S. Costanzo and of the last-named church are almost exactly the same. The obvious intention of the architects was to secure a cruciform effect within, and for this purpose the central compartment was covered by a cupola, the four-angle compartments were covered with voutes d'aretes at a low level, and the remaining compartments with barrel vaults at a much higher one. As at Gaeta, so in this church, the cupola is a high pepper-pot dome supported on pendentives. All the roofing is supported on two square pilasters of masonry* and fourteen pillars of various coloured marbles, grey granite, green cippollino, and jallo antico from Numidia. Presumably no ancient capitals were available, and plain stone slabs were used instead. From this short survey the student will obtain a general idea of the church. The original narthex at the south end is now used as a wine-cellar, and consists of two parallel chambers covered by * Replacing two shafts of Numidian marble that were carried of! by the Bourbons to decorate the Palace chapel at Caserta. All these pillars were obtained from a Koman building. The domes are of the same shape, but this one is suspended on pendentives while that at Gaeta is on squinches. CAPKI 77 barrel vaults supported on two square pilasters. The south end seems to have been walled up to form a cistern, and three holes were cut in the vault to drain the water from the terrace above, and from the roof of the church. The entrance to the church was by three round-headed doors, corresponding to the nave and aisles and the space between the pilasters of the narthex. Light was obtained by three vents on the south side. Some additions have been made to this narthex, and the doors into the church have been blocked up, but otherwise it is substantially in its original condition. Passing out into the little fore-court or piazza we enter the church by the door cut through the south wall in the twelfth century. This leads into a narthex of three bays, covered by cross vaults, that support the priests' apartments above. The henitier near the door stands on the fragment of a cippoUino pillar shaft, originally cut in baluster form. The companion of it will be seen in the nave. The division of the nave into nine compart- ments has already been explained, and with the aid of the plan the accompanying photographs will give an idea of the original church taken from two different points. The first of these is taken from the modern choir looking east. The second is taken from the door looking in the opposite direction, and the choir is seen in the distance, through the pillars of the original church. The choir beyond is a plain rectangular room covered by a ribbed cross vault, resting on four corbels of Norman character, two grotesques, a well-carved head, and an ornamental design that could not be distinguished. The original apse of the earlier church has been cut in two to make a passage. We could find no indications of a screen or other altar furniture, but in the garden my wife found a post with a cross carved on it that belonged to a chancel screen. In the modern room adjoining the apse is the fragment of a marble console decorated with the acanthus leaf, and beneath it the remains of a mediseval tombstone of doubtful date. A good view of the exterior of the church could not be obtained owing to the surrounding garden, but the rough sketch will suffice to show that the dome is almost exactly the same as that of S. Giovanni at Gaeta. The outline of the original church is difficult to distinguish owing to the additions of priests' rooms, 78 SOUTH ITALY over the modern narthex, and the two angle compartments on the east side. The approach to these rooms is by a staircase at the end of the passage cut across the original apse. In considering the date of the original part of the church we are again confronted with the usual difficulty that there is no contemporary documentary evidence, and the general history of Capri is so involved and confused that little help can be derived from it. The foundation of S. Giovanni at Gaeta is locally attributed to the Duke John in 999 ; but we found no reference in the Codex Cassinensis respecting it till the thirteenth century, and this important source of information is silent as to S. Costanzo at Capri. The two churches are, however, so much alike, save in one detail,* that it seems reasonable to suppose that in the main they were built about the same time. We hazard a conjecture that the present edifices are not the original ones ; that the date of the foundation of the preceding churches, though not necessarily the first to occupy the site, is to be sought between the middle of the eighth century and the middle of the ninth ; and we do so because of the altar front at Gaeta and the little screen post we found in the garden near S. Costanzo. We consider that these are of the eighth century, and that if it be the fact that the Duke John had any hand in building S. Govanni it was as a restorer of an earlier church to which this altar belonged, and not as the original founder. We would venture to go farther and say that it seems likely that the originals of the three churches f were built under the auspices of the monastery of Monte Cassino. There is a tradition that Capri was given by the Byzantine Emperor to the monks of Monte Cassino. There is a sufficient resemblance in plan in these churches to justify the assumption that they were copied from one another upon a common plan that happened to be fashionable at the moment. The practical difficulty of assigning so early a date to the churches at Gaeta and Capri is that all these coasts were infested by the Saracens in the ninth and tenth centuries, but the complaint of the Latins against Naples, Gaeta, and the other maritime republics shows that the * The dome at Gaeta is supported on squinches, and the dome at S. Costanzo on pendentives ; but the two systems were used concurrently, and too much importance should not be attributed to the difference. t S. Maria delle cinque torri is the third. CAPKI 79 latter were able to hold their own, and either by treaty or by force to keep their territory free from the Mahometans. The local historians of Capri supply us with four pieces of information respecting the early Church history of the island. As the first two relate to events in the sixth century that hap- pened long before S. Costanzo was built they need no more than a passing reference. The third event arose out of the political history of Southern Italy. The fourth is in a measure connected with the third and relates to the creation of the bishopric of Capri of which S. Costanzo became the cathedral church. The third and fourth may therefore conveniently be taken together, though they were separated by nearly a century. There is a tradition, and I will call it no more for the present though it is quite likely to be founded on fact, even if the persons said to have been principally concerned had no hand in the trans- action, that the Emperor Louis II. took Capri away from the Duchy of Naples and gave it to the Amalfitans in consideration for services rendered by the latter in fighting the Saracens.* * I quote the following from pp. 93-4 of Fabio Giordano's Relation of Capri, edited anonymously and printed at Naples in July, 1906 : ' It is said that the Emperor Ludwig took Capri from the Neapolitans and gave it to Amalfi as a reward for services rendered in the Athanasius incident. I cannot trace the original authority for this donation. Modern writers copy the usual account of it from Mangoni ; who quotes Pansa to this effect ; who quotes, or says he quotes, from Marino Freccia. Even Freccia, with all his learning, would be no authority ; besides, I have been able to discover no passage in his imposing De Subfeudis which speaks of such a donation. Both Camera (Istoria di Amalfi, p. 100) and Anastasio (Lucuhrationes in Surrentiiwrum, etc., vol. i. p. 110) name Freccia in this connection, although he merely states that after the Roman period Capri became subject to Amalfi. Anastasio's learning on this subject is mostly derived from Chioccarelli ; he goes so far as to copy that writer's wrong reference to the tenth volume of Baronius. Among other older writers, Chioccarelli {Antisitum, etc., p. 88) and Baronius {Ann. EccL, vol. 14 p. 264) both give accounts of the incident which is supposed to have led to this donation, but without mentioning Capri at all. The chroniclers I have consulted are equally silent. Leo Ostiensis (Peter Diaconus), Paul Diaconus, Giovanni Diaconus, Erchempertus, the Chronicum Amalfitanum, the Chronicum Anonymi Salernitani, the Chronicum archiepiscoporum Amalfitanorum, most of whom touch upon this episode, do not speak of Capri as the supposed reward of the Amalfitans. Nor is there any mention of it in the documents concerning St. Athanasius published by Muratori, nor in Paolo Regio's life of that saint. At the same time, there is no doubt that, whatever its ecclesiastical status may have been, Capri was colonized during this long period from Amalfi, and not from Sorrento. Thus Edrisi, in an important passage, writes: " L'isola di Caprie abitata da uomini d' Amalfi che vi tengono loro greggi . . ." (L' Italia descritta nel libro delReRuggiero, M. Amari e E. Schiaparelli, Roma, 1883, p. 18). The existence of a large number of Amalfitan names of families and places on Capri testifies to this long domination.' 80 SOUTH ITALY It is also said that in 987, that is a little more than a century- later, the Archbishop of Amalfi was commissioned by the Pontiff John XV. to consecrate a certain John as the first Bishop of Capri. The date 987 corresponds closely with that usually given for the construction of S. Giovanni at Gaeta, and as S. Costanzo is almost exactly like the Gaeta church there seems to be good ground for thinking that they were erected as we now see them at the same time, S. Costanzo being built as the cathedral church for the newly created see. The tradition respecting the grant of Capri to the Amalfitans mast be relatively late, for the early authorities do not mention it. The student who is acquainted with the history of Southern Italy in the middle and close of the ninth century, and the relations between the Greek maritime republics or communities and the Saracens on the one hand, and the Franks, the Lombards, and the newly created Pontifical States on the other, will probably see no reason to doubt that such a grant was made for much the same reasons that Fondi and Trajecta were given to Gaeta by the Pontiff John VIII. And in that connection an inquiry into the financial relations between these Republics and John VIII. would be interesting. As the country round Eome was either completely wrecked or in the hands of the Saracens, one would like to know where John VIII. obtained the money to carry on the almost incessant campaigns with the Saracens he was engaged in during the eight years of his reign after the Emperor Louis II. was dead. Did these Eepublics find the funds and were these grants of territory the price paid for them? The creation of a bishopric at Capri is usually attributed to the revival that took place in the Latin Church in Southern Italy towards the end of the tenth century. This revival was the answer of the Latins to propaganda of the Emperor Nicephoros Phocas for Hellenizing Calabria and Apulia through the medium of the Greek Church and the Patriarchate of Constantinople.* To follow the sequence of events that led up to this movement in the Latin Church it will be necessary to refer back to the confiscation of the Patrimony of the Eoman Church and the establishment of the Patriarchal jurisdiction in Italy and Sicily by the Emperor Leo III. in the beginning of the eighth century. Against the confiscation of the Patrimony protests were made from time to • Gay, vol. ii. pp. 358, 359. OAPEI 81 time, for instance by Hadrian I. in the eighth century and by Nicholas I. in the ninth. Their successor, John VIII., main- tained friendly relations with the Byzantine Court and the Greek Church under the Patriarch Photius, a policy that stood the Christians in South Italy in good stead, and was in contrast to that of his predecessors ; but his letters show that he too did not forget to assert the claims of the Roman Church to spiritual allegiance or reprove the Princes of Sardinia, then nominally at any rate dependents of the Emperor at Constan- tinople, for failing in obedience. We may account for these claims to the patrimonies and the patriarchal jurisdiction not being pressed in the hundred years that elapsed between the murder of John VIII. and the Pontifi- cate of John XV. on at least two grounds. First, that no less than thirty-two bishops occupied the Roman see during that time, and, secondly, that as in Rome so in the other Latin cities of Southern Italy, the patronage of the greater bishoprics passed into the hands of the local nobility. John XV. was a protege of the Crescentius family, and he appears on the scene at the time when the prestige of the Roman see was at the lowest ebb. We know little about him beyond the fact that after a short interval he succeeded his predecessor John XIV. (sometime Bishop of Pavia), who had been murdered shortly after the death of his friend and patron, the Emperor Otto II. At about this time certain of the Southern Latin sees, including Salerno and Amalfi, were raised to the rank of archbishoprics. This had already been done with Capua at the instance of the reigning family, and the twofold object of these steps seems to have been to gratify the local magnates, and to act as a set-off to the Byzantine Metropolitans in Calabria and Apulia, then recently created by Nicephoros Phocas.* The promotion of the see of Amalfi and the creation of one at Capri are usually attributed to this revival. • The student who is interested in this period will find a review of it in the second volume of Gay's LHtalie Meridionale et I'Evipire Byzantin. €2 SOUTH ITALY MONTE CASSINO Sta Maria clelle cinque torri at S. Germano The historians of Monte Cassino attribute the erection of this church to Abbot Teodemaro* in the last quarter of the eighth century. The monastery recently destroyed by Goths and abandoned for just over a century, had then been rebuilt about seventy years. The plan of a square church divided into nine compartments, with small domes to cover the angle compart- ments and a larger dome to cover the central one, is common enough in Byzantine churches of the tw^elfth and thirteenth centuries. The south of Italy possesses two examples of it in the churches known as the Cattolica at Stilo and in S. Mark at Kossano, both of them in Calabria, f But in those two churches and in the many other examples of the use of this plan in the Levant, the domes are cylindrical of the high pitch and pepper-pot shape afterwards adopted by the Normans in building their Sicilian churches. I am not aware of any other church in Southern Europe, of this date at all events, where square lanterns take the place of domes. There is a church | built on this plan in Lower Nubia, at a village called Addendan, twenty miles north of Wady Haifa, and other examples have been found in Upper Nubia and Abyssinia, where the Ethiopian ritual requires the altar to be placed in the centre of the church with an ambulatory round it. With the exception of the pillars and their caps and portions of the walls, it is impossible to tell how much of Teodemaro's fabric now remains. We have not been able to find either an accurate description or any satisfactory account of the founda- tion of this church ; and if any records on the subject exist in the annals of the monastery they have not, so far as I know, been published. Schulz describes the church and gives a ground-plan * Storia della Badia di Monte Cassino Luigi Tosti, 1842. Also Monte Cassino, Descrizione. By F. della Marra, 1775. t Vol. i. p. 95. I It was excavated by Messrs. Mileham and Maciver in the Eckley B. Coxe, jun., expedition to Nubia. See their account in the publication of the University of Pennsylvania, Churches in Loiuer Nubia, published at Philadelphia in 1910. I am indebted to Mr. J. Crowfoot for the reference to Abyssinia. ^3 S. Maria. Cinque torn'. vf Looking diagonally across the nave Pi ait. 1 1 ----• ■ 1 (|) J^vt^J-Cil -Q Q 0: Ji^/. 7 ' (:>:::: nine compartments, with siuail domes to cover the angle couipart- iDcrtta and a larger dome ^ttJ^^ver the central one, is common I aottgh in Byzantine , church e s of ithe twelfth and thirteenth • t>nturie8. 'I?iT5~gmithf''o|''rfiiI^p<%3;^88es two examples of it in the. churches kiukiySyps [i ai Rossano, | both of i\) (■liurcTi cs and in ^he maj uy ^Trtmer ipvunpi,.^ . plan n the LevanJ;, Jthe domes are cyli^dirical of tiiti high pitch and p<:pper-p^t s^{|,pe afterwards ado^t^d vfej^Jthe Normans in buildup their Si dilian churches. I am ijiot aware Df any other I'hurc h in-Sout5ei-6r..KHH"eB»yf>f-tl>i»^l^ vt ere square lante :n 5" fake th'e place "ofaomei. "Th"el'ells a church 1 built on this )h,u m Lower 'Nubia, at a village {ciUed Addei id m, twenty miles north of W|«k^ Haifa, and other ^f|,mples havypeeu found m Up])er Nubia' anad Aiwl&iitt^ whef*e.' the EthiDpian ritual requires the altar ijto be placed in the centre of the [cljurch with 1— i-H' ,t Stilo and in S. Mark l>iit in those two »n aipqulatory round it V ''it h the exception of the pillars anjd Itheir caps >f th J ?*^ fabrid Jiow remai4s' We have not be^nJ able to )ii< Tits d6s«j^ti6a or any satisfactory ^ if ;his church ; [ and if any records (Ji ; s l >f the muilAiJLeiy ihay ItiivT' ^s*^ [ir d portions l^iodemaro's It her an i lie I'ounda- e xist in the now, been lied. Schulz describes the church and give;^ a ground-plan * tiiaria della Badia di Monte Cassino Luigi Tosti, 1^42.. A'-^o* Monte Cassino, Desci ..'■"«/• By F. della Marra, 1775. "^^^^ I N I fv 95. ; ' 'vated by Messrs. Mileham and Maciver in the Eckley B. Coxe, jun., cxpci]' :!iia. See. their account in the publication of the University of I .s in Loioer Nubia, published at Philadelphia in 1910. I am .vvfoot for the reference to Abyssinia. S.*?) ^■%»«\ s;>vi\^ o\ MONTE CASSINO 83 and section, but the latter appears to have been copied from an older work * and we suspect that Schulz was never inside the church but took the information he gives from Leo d'Ostia and Tosti. In the first place the plan is wrong, for there is only one apse and no indication that the side apses shown in it ever existed. Schulz observes that Leo d'Ostia does not mention three apses, and though not impossible it is extremely unlikely that a Latin church of this date and in this locality would have more than one apse. In the next place, the section represents all the pillars as fluted. Schulz correctly states that two of the twelve pillars are made of cippollino and the others of granite, but only one, the second on the north side, is fluted, and the others are quite plain. The accompanying plan does not profess to be more than a rough sketch intended to show the shape of the church and the place where the picture was taken from. The photograph is from A looking towards B. The view of the exterior, taken from the roof of a house, is not very satisfactory, but it shows the odd arrangement of the lanterns and roof that dis- tinguish this church. This roof is made of tiles supported on joists and rafters exposed to view inside. The intervening spaces have ceilings of plaster in the form of elongated cross vaults. Beside the round-headed windows in the side walls the church is lighted by square windows in each face of the central lantern and by round-headed windows in the outer sides of each of the side lanterns. The tower on the south-east side is a modern addition. There are three entrances to the church, on the north, south, and west sides, and four steps lead from the ground outside down to the floor. The general appearance inside is very pleasing, owing to the handsome pillars and the effective play of light on them from the windows above. The shafts are monoliths, and as Tosti observes, the caps were obviously taken from a Koman building. They are so well preserved that at first I took them to be mediaeval copies, but a closer inspection shows that they are * Seroux d'Agincourt. Histoire de I'art par les monuments, 1823. Plate XXV. 43 and 44. And for text, vol. iii. p. 13. See also the illustration of an Armenian church, Plate XXVII. 28. Also three churches two miles from Erivan. Voyage en Perse, Chard ; pub. at Amsterdam in 1735, vol. i. p. 214. 84 SOUTH ITALY Eoman and that they were all made at the same time. There are no Christian emblems upon them, but a small device in a rosette, probably intended to represent a flower.* The present apse is a square room covered by a dome, both apparently later additions, that have been smeared, like the rest of the church, with plaster and whitewash. Nothing is now visible of the wall paintings or texts mentioned by Tosti, but no doubt they still exist under the plaster. One may, perhaps, be permitted to conjecture that the plan of this church is merely a variation of the common Levantine plan of dividing a nave into nine squares and covering the central and angle squares with domes. It must be remembered that Monte Cassino was the resort of pilgrims and travellers from all over the Christian world, and consequently just the kind of place where one might expect to find a peculiar design imported from abroad. * Cf. some of the capitals in the Christian basilicas at Rome. The Christian copies of classical originals are sometimes good enough to deceive the very elect. NORTH AFRICA The history of the Church in North Africa after the proclama- tion of Christianity may be divided into three periods. The first, or Roman period, commencing in 314 and ending with the con- quest of Carthage by Genseric in 437 ; the second, or Vandal period, ending with the capture of GeHmer by Behsarius in 534 ; and the third, or Byzantine period, ending after Carthage was taken by the Arabs, when the Christians were given the option of becoming Mahometans or leaving the country. From that time Christianity disappeared from North Africa, and by degrees the country, like Morocco of to-day, lapsed into a condition, of anarchy and desolation. It is probable that the Christians did build some churches before the ' Peace ' of the Church in 314, but archaeological research has so far failed to identify any particular church with this early period in North Africa. From the writings of the African Fathers it is certain that soon after 314 many churches were built, and as the ruins show, usually in the suburbs of the cities. This was done, it is said, not to offend the pagans, but the more probable object was to be near the Christian cemeteries. It is also certain that directly after the Arab conquests in the end of the seventh century when the Christian clergy were dis- persed and exiled, the native Berber population in a body became Mahometans, and church building came to an abrupt end. If the proclamation of Christianity affords a convenient ter- minus a quo for the study and dating of Christian architecture in North Africa, the Arab conquest with the proscription of Christianity and the dispersal of the Christians that ensued, supplies a correspondingly convenient terminus ad quem. 86 NOETH AFEICA To fix the date of any particular church within this period is a very difficult matter, as the architecture, the basilican plan, the timber and tile roofing, the position of the altar, the choir or tribunes, the semi-circular apse with clergy seats in tiers, and the bishop's throne in the centre, were the same throughout the whole period. And in this part of North Africa nothing, as a rule, is left of the old churches above the footings, so that the archaeologist must be content with the ground plan, the mosaic pavement, or such stray fragments of the fabric, capitals, carved stones and other pieces of decorative detail, as may have had the good fortune to escape the destructive hands of the Arabs. It is hardly less difficult to determine whether a church was built in the Eoman, the Vandal, or the Byzantine period and by the orthodox, the Donatist or the Arian communities. But it may be assumed that the great majority in Tunis were built by the orthodox, either in the Eoman or the Byzantine period, that some were built by the Donatists, and few or none by the Arians, The latter at any rate had no need to build, for the churches, the sees and the revenues of the orthodox were con- fiscated by the Vandals and handed over to the Arian clergy. And in that connection the names Vandal and Arian may be used synonymously and applied to what in fact became a garrison Church, that lasted so long as the successors of Genseric were masters of this part of Africa and came to an end when they were conquered. In Tunisia I have not come across any indication, such for example as the symbol 'Deo Laudes,' that would identify a church with the Donatists * or other schismatics ; some no doubt existed, and S. Optat complains of the number of churches built by them ; but these were chiefly in Numidia where the sect prevailed. We may account for the absence of evidence of the Arian or Donatist occupation of the Latin churches by supposing, and I think with reason, that when the Orthodox came into their own again in Justinian's reign all traces of it were obliterated. * On the Donatists : Milman's, History of Christianity, vol. ii. p. 229 ; Mason's Persecutions of Diocletian, p. 154 ; Broglie, vol. i. p. 254 ; Foakes Jackson, p. 286 ; Duchesne, Early History of the Church ; Diehl, L'Afrique Byzantine ; Gsell, L'Algirie dans VAntiquiti. 3»4 :airo. Anba Schenouda. Screen and altar rail in Alexandria mnseiuu. To face page 86 86 NORTH ^t no date f^articular church within this period is a very difficult m- the architecture, the basihcan plan, the :, the position of the altar, the choir n lilar apse with clergy seats in tiers, iQ the centre, were the same throughout the wh \nd in this part of North Africa nothmg, as a he old churches above the footinjjg, so that ih ■ tuust be content with the ground plan, the ] , or such stray fragments of the fabric, capitals, car hi other pieces of decorative detail, as may have J fortune to escape the destructive hands of the A.. ly less difficult to determine whether a church was ' Uoman, the Vandal, or the Byzantine period and ;. the Donatist or the ^V ; od that the great majoiiiv ju j-.u;. , 1 ., either in the Roman or the Byzantine periou, 1 ':re built by the Donatists, and few or none by the rhe latter at any rate had no need to build, for the 'he sees and the rg^n^i^gj of the orthodox were con- ',he Vandals .^^JoM^^fi^j^^^ ^^ *^e ^"^^ clergy. connection the namas Vandal, and Arian may be laously and applied to what m fact became a • iUrch, that lasted so long as the successors of were masters of this part of Africa and came to an •t they were conquered. ia I have nolJ come across any indication, such for ^ho symbol 'Deo Laudes,' that wi^-' ' -'^.' * - •- Donatists * or other schismatics^ Optat complains of the n but these were ch: the We may account foi i iJence Donatist occupation ol iches by ' think with reason, that when the Orthodox >vv'n again in Justinian's reign all traces of it '■■^. History of Christianity, vol. ii. p. 229; Mason's ., vol. i. p. 254 ; Foakes Jackson, p. 280 ; :; Diehl, L'Afrique Byzantine; Gscll, . .tnttqwttf And ^?) ^"^ft^ "i^viX 0\ INTRODUCTION 87 The student of mediaeval ecclesiology who is famiHar with the fundamental difference between the structure of a twelfth century Latin and Greek church in the position of the altar, the concealment of the sanctuary by a screen, and the presence of the chapel of the prothesis and the diaconicon, might perhaps suppose that there would be no difficulty in determining whether a Tunisian church belonged to the Eoman or the Byzantine period, whether it had been built by the early missionaries from Rome or by the Greeks who followed the conquest of Belisarius. Unfortunately it is not so, and the obvious explanation seems to be first, that minor differences of use apart, the holy mysteries were celebrated in early times substantially in the same simple way throughout Christendom, and secondly that Christianity in this part of North Africa did not last sufficiently long after the Byzantine conquest for the great change introduced in the Byzantine liturg}?- by the Greek ecclesiastics after Justinian's time to have, affected church building in these distant provinces. The most evident architectural changes necessitated by the altered ritual was the addition of the prothesis and the dia- conicon on either side of the central apse containing the altar, and the shifting of the altar itself 'from the nave close up to the central apse. At S. Eirene, for example, the altar must have stood well out in the nave of the church and not in the semi-circular apse, for the latter is provided with seats for the clergy arranged in tiers round the wall as we find them in the Tunisian and Nubian churches. The altar, therefore, probably stood in a place where it could be seen by all the congregation, and there was certainly no high screen in front of it as there is in the Greek and Coptic churches of to-day.* It does not, however, follow that the congregations of that early date witnessed the celebration of the holy mysteries at the altar as in the English and Latin Churches of to-day any more than they do in the Armenian Church, where the altar, habitually exposed to view, is concealed by a movable curtain drawn in front of it at the celebration. * Cf. The screen in the church of Anba Shanudah in Dair Abu-s-Sifain in Old Cairo, and the tribunes with the altar in front in the church at Faras, Lower Nubia ; both illustrated here. 88 NORTH AFRICA This curtain arrangement must be of very early date, for there is a marble in 8. Sofia representing the sanctuary of a church provided with curtains partly drawn aside to show the Holy Table with a cross on it. The curtains appear to have been attached to a canopy or baldacchino of the kind com- monly found in the Coptic churches of Old Cairo, and many pavements of the larger Tunisian churches have traces of the foundation of some such erection over the altar. There is in the Bardo Museum at Tunis a mosaic representing the interior of a church.* The drawing is conventional and out of perspective, but the artist obviously intended to represent a basilican church and the altar stands exposed to view in the middle of the nave towards the sanctuary. What object three upright lines are intended to represent, whether the pillars of a canopy over the altar, curtains, or even altar lights is not clear, but there is cer- tainly no indication of a closed screen here or in any of the ruined churches in Tunisia. It is likely, however, that these African churches may have sometimes possessed open screens like the *pergula' in S. Maria in Cosmedm at Rome, and the upright lines may be intended to indicate an open screen of that kind. Again the triapsidal arrangement occurs in the earliest as well as the latest buildings. In the basilica at Tebessa, for example, there are the prothesis and diaconicon, though the church was built long before the reign of Justinian, when the Byzantine liturgy was elaborated. In Constantinople this triapsidal ar- rangement affords an unfailing test of the age of a church, because the three apses were provided to suit an elaborate ritual, introduced after Justinian's period and still practised in the Greek Church of to-day. A conjecture that this later and more elaborate Byzantine ritual I have alluded to was copied from a ' use ' of the African Church presents itself temptingly, but there is no evidence to justify it, or indeed an assumption that these side chapels in the early African churches, whether basilican or trefoil, had any ritual significance. There is a strong probability that they had, and that is all that can safely be said. In one case, for example, where the church has been burned, broken fragments of the * From Tabarka, illustrated here. See also the reference below under Sidi Abich and Bir bou Rekba. 35 f Coptic screens and font. S Michael^ Kamula. Nakadeh^ Egypt. To face page 88 NOETIT ATPT^icA ihis curtain arrangement must be of very early date, for S^ there is a marble in 6. Sofia representing the sanctuary of a church provided with curtains partly drawn aside to show the Holy Table with a cross on it. The curtains appear to have I. iched to a canopy or baldacchino of the kind com- monly ..... m the Coptic churches of Old Cairo, and many pavements of the larger Tunisian churches have traces of the foundation of some such erection over the altar. There is in the Bardo Must-um at Tunis a mosaic representing the interior of a church." The drawing is conventional and out of perspective, but the artist obviously intended to represent a basilican church and the altar stands exposed to view in the middle of the nave tov> ' 'le sanctuary. What object three upright lines are ini — .') represent, whether the pillars of a canopy over the altar, curtains, or even altar lights is not clear, but there ia cer- tainly no indication of a closed screen here or in any of the ruin<^ churches in Tunisia. It is likely, however, that these African churches may have sometimes possessed open screens like the *pergula' in S. Maria iiiv(^steMdf»^t^iK(5!i*^4nd the upright lines may be intended to indicate an oDen screen of that kind. the tnapsidal arrangement^ccurs m the earliest as well as . _:st buildings. In th'e basihca at Tebessa, for example, there are the prothesis and ^gtedluysi^u, though the church was built long before the reign of\^ii|^nian, when the Byzantine liturgy was elaboi'ated. In Constantinople this triapsidal ar- rangement affords an unfailing test of the age of a church, because the three apses were provided to suit an elaborate ritual, introduced after Justinian's period and still practised in the Greek Church of to-day. A conjecture that this later and more . , ritual I hiive alluded to was copied from a * i an Church presents itself temptingly, but thert ^ idcnce to jufetify it, or indeed an assumption t1 ' , chapels in the early African churches, whether b;: .. .rcfoil, had any ritual significance. There is a strong probability that they had, and that is all that can safely be said. In one case, for example, where the church has been burned, broken fragments of the • From Tabarka, illustrated here. See also the reference below under Sidi Aoich and Bir bou Rekba. ?>?> ^-^ft^ ^:)a\ oT INTEODUCTION 89 glass Communion vessels were found on the floor of the side apse. From this it has been surmised that these side chapels were used as vestries and for the offerings of the faithful ; and if that be so, then they may be regarded in a way as the pre- decessors of the Byzantine prothesis and diaconicon. Their presence in an early church like the basilica at Tebessa came to me as a surprise, and shows that this architectural peculiarity cannot be used, as in Constantinople, to fix the date of a church, for it occurs also in a later church like El Kef, which was certainly built long after the Byzantine conquest. The difference between the Byzantine triple apse chancel and the chancel arrangement found in the earlier African churches, like Announa or Tebessa, must be distinguished. In the latter the altar stood, as it does in the Western Church, out in the open, and the apse, copied from a secular Boman basilica, was fitted as a kind of theatre with a bishop's throne in the centre and seats for the clergy on either side of it. This plan is made familiar to us by the basilican churches in Eome, like S. Clemente, or the church at Torcello, Parenzo Cathedral in Istria, or S. Eirene at Constantinople. But in the Byzantine arrangement the small central apse was put to enable the celebrant to walk round the altar at the celebration, and the altar itself stood concealed from the view of the congregation by the screen or iconostasis. The use of Boman materials taken by the Christians from temples and other public edifices for church building affords a surer indication of date, but the test is not always reliable. The edict of Constantine did no more than recognize the right of the Christians to practice their religion and worship openly. They certainly availed themselves of the new law to push forward an active propaganda and the conversions proceeded rapidly. No organized steps were taken by the Government to suppress paganism till the reign of Theodosius, and some temples continued, up to that time, to be used for the heathen ceremonies. It is true that in Egypt, for example, the temples were destroyed as part of the Christian propaganda, and some- thing of the sort may well have occurred in this part of North Africa. But as several temples were appropriated by the Christians and used as churches, it would seem that the sup- 90 NORTH AFRICA pression of paganism did not of necessity involve the destruc- tion of the temples, and in some provinces the law actually forbade it.* The Roman theatres, forts, courts, market places, triumphal arches and the like stand, as secular buildings, on a different footing to the temples. In Constantine's reign these, too, were comparatively new, for they had been built during the period of great prosperity in the reigns of the African Emperors Septimius Severus and Caracalla. It is difficult to believe that the solid and substantial masonry used by the Romans for these buildings had in little more than a century and a half so far perished and fallen into decay as to be fit qnlj for breaking up and providing material for newer edifices. It is known that Genseric destroyed the forts for strategic reasons and the Vandals have consequently been credited with the wanton destruction of all Roman monuments. This assump- tion is altogether unwarranted, for the Vandals, like their neighbours the Goths in Italy, aspired to replace the Romans and carry on their method of government and civilization. It is more probable that the native Berbers are to blame for the wholesale destruction of these Roman monuments. However that ma,y be the use of Roman materials in quantities in either forts or churches may generally be taken to indicate that those buildings were put up after the Byzantine conquest, for in the military treatise of Justinian's period, called the Nea Taktika, the engineers were expressly advised to use materials from ancient buildings in constructing their forts, and to select some site near an ancient town where such materials could be readily obtained. There is ample evidence that this advice was almost invariably adopted. The task that Justinian set himself in reclaiming North Africa for the later Roman Empire involved much more than the overthrow of the Vandal Kingdom. The country devastated during the century of Vandal dominion had to be re-settled and defended against the Berber tribes and clans, * At Dougga the temples of Jvipiter and Minerva, andCcelestis (No. 6 in Gu^rin's list, vol. ii. p. 131) were appropriated by the Christians. The foundations of the semicircular apse added to the cella of the latter, when it was converted in a church, still remain. See also Gsell, vol. ii. p. 122. a6 i^'IU 'f^^l osaicjr'om Tabarca representing an ancient hasthca. To face page go ds: i>ression of paganism did not of necessity involve t tion of the temples* and in some provinces the law forbade it.* ' The Roman theatres, forts, courts, market places, triumphal arches and the like stand, as secular buildings, on a different footing to the temples. In Constantine's reign these, too, were comparatively new, for they had been built durmg the period of great prosperity in the reigns of the African Emperors Septimius Severus and Caracalla. It is difficult to hat the solid and substantial masonry used by the Komans for these buildings had in little more than a century and a half so far perished and fallen into decay as to be fit qnly for breaking up and providing material for newer edifices. It is known that Genserir d< and the Vandals have cl j... .... ., wanton destruction of all Ptornrtn inonnments. tion is altogether unwarra; 'andals, lik neighbours the Goths in Italy, aspired Lo replace the Komans and carry on their method of government and civilization. It is more probable that the native Berbers are to blame for the wholesale destruction of theso I\oman monuments. quantities in either forts or churches may generally be taken to indicate that those -buildings were pat up after the Byzantine conquest, for in the raihtary treatise of Justinian's period, called the Nea Taktika, the engineers were expressly advised to use materials from ancient buildings in constructing their forts, and to select some site near an ancient town wht re such materials could be readily obtained. There is ample evidence that this advice w adopted. The task that Justinian set 'ti ^ North Africa for the later Roman Empi !i more than the overthrow of the Vandal K > ountry (I 1 during the century of Vandal ■ nad to be r jiiid f1(>l"(MML-fl :i,f/airist, fho. P,or\i*^v ',ind clans. * At Dougga the temples of Jupiler and Minerva, and Coelestis (No. 6 in Gu^rin's li?;t ■■]. H. 1. iSl) were appropriated by the Christians. The foundations of tin added to the cella of the latter, when it was ponvei .;ri Kr.P nl,r. Prvipll. vr>l ii ii 122. OQ ^"^s^^ "^'^tiX o'X INTBODUCTION 91 whose subjection neither the Eomans nor the Vandals had succeeded in accomplishing. Unfortunately for Justinian, one of Genseric's first acts when he conquered Africa was to destroy the Koman fortifications. So soon, therefore, as GeHmer had been taken prisoner and the Vandal forces were broken up, the Byzantines set to work to build forts and block houses all over the country to keep order and protect the colonists. A great number of these forts still exist in a more or less complete con- dition, some few, indeed, are almost as perfect as the day they were put up. They are all built with stones of large size taken from Roman buildings, obviously roughly and hurriedly put together, and they show that the advice in the Nea Taktika was almost invariably adopted. In the ruined towns of Central Tunis it will be noticed that as a rule the churches were built, like the forts, with Eoman materials. At Sbeitla and at Uppena, churches had been rebuilt in part at least, with Roman materials on the sites and over the debris of earlier churches. The Byzantine basilicas at El Kef, at Announa in Algiers, and the recently discovered church in the outskirts of Dougga, all contain Roman materials. So also do the garrison chapels built in the Byzantine forts at Haidra and Sbeitla. And to this short list many more examples might be added. The choice of dates then will rest between the Roman and the Byzantine periods. Of these ruined Christian churches by far the most interesting and important example for the study of Roman and Byzantine archaeology is at Uppena, where two large basilicas were built on the same site. The earlier basilica built before the Vandal occupa- tion, probably in the end of the fourth or beginning of the fifth century, contained a mosaic tomb with an inscription recording the names of a number of Christian martyrs of the early persecu- tions, and the later basilica, built after the Byzantine conquest, contained a mosaic inscription in which the earlier inscription was repeated and decorated with an ornamental border and cross in a common Byzantine pattern. In this church, and at Sidi Abich near by, the tombs of several bishops were found, and it might be supposed that there would be no difficulty in identifying them by their names with the 92 NOKTH AFRICA bishops who are commemorated by the early fathers as having attending Councils. But unfortunately this is not so, and the number of cases where an identification of this kind can be made may be reckoned on the fingers of one hand. It is very odd that the year is never given on these tombs though the age of the deceased, the month, day, and even hour of death are often recorded. The ornamental border round the inscription of the martyrs at Uppena, illustrates what has been said respecting the use of decorative detail as a guide to date. The pattern on this particular border is frequently met with * and from the example at Uppena, it may be safely taken to indicate that the work belongs to the Byzantine and not to the Roman period. A few churches in seaport towns on the East coast of Tunis were apparently decorated with capitals, consoles, and other carvings designed from Constantinople patterns of the Justinian period. But in the Interior the churches were decorated either with poor copies of Roman models or native designs derived from them, and they afford no evidence of date. I speak with less certainty concerning the fonts and the baptisteries containing them ; probably they will afford some evidence of date when more is known of the ceremony they were built to suit, and whether they were intended for the Sacrament of baptism or, like the Coptic Epiphany tanks, for some periodical and commemorative service of ablution. The distinction will at once be appreciated by those who are familiar with the arrange- ments of the Copts of Egypt, where every church is provided with a ceremonial tank as well as a sacramental font. These tanks, usually 2 to 3 metres square, and about three-quarters of a metre deep, are sunk in the floor of the church near the main entrance. They are generally covered up with boards and only used at certain times of the year, such as Maundy Thursday, for a service that is certainly nothing but a commemorative act of ablution. The font proper for baptisms is usually a large earthen jar sunk in a niche or hole in the wall about 1 metre from the ground. This niche, converted into a sort of cupboard with a door to it, is generally tucked away in some obscure corner of the church. No branch of the Christian Church has been more thoroughly * Sousse Museum ; small basilica on North-west side of the city of Sbeitla. ^ S. Michael Ka inula, ^g^'pl^ ^^^^^^ ^^^^^ ahich Enfida, Tunis. ■•^V" Sta. To face page Q2 92 NOKTH AFRICA iij list this y bishops who are commeraoraitjd by the early fathers as having attending Councils. liut unfortimately this is not so, and the number of ' ' ation of this kind can be made may be rec „ . _i one hand. It is very odd that the year is never given on these tombs though the age of the ~'3d, the month, day, and even hour of death are often 'al border round the inscription of the rates what has been said respecting the detail as a guide to date. The pattern on r is frequently met with * and from the ' may be safely taken to indicate that the ■ Byzantine and not to the Koman period. ^ in seaport towns on the East coast of Tunis lecorated with capitals, consoles, and other from Constantinople patterns of tl; ' ian iie Interior the churches were decora aer eg of Roman models or native designs derived from itford no evidence of date. less certainty concerning the fonts and the lining them ; probably they will afford some f 3 metres square, and about three-quarters of are sunk in the floor of the church near the main ire generally covered up with boards and only . ..lues of the year, such as Mair- ' - '""' ^ for a service that js certainly nothing but a couj of ablution. The foi^ proper for baptisms is usually a large earthen jar sunk in a niche or hole in the wall about 1 metre from the ground. This niche, converted into a sort of cupboard with a door to it, is generally tucked away in some obscure corner of the chinrcli. No branch ol the Christian Church has been more thoroughly • Seuaie MuHsum ; small basilica on North-west side of the city of Sbeitla. pc with } theiu of ail OQoe } n.' tanks, US I. a metre dcM\) entrance, used at cv..\. , > ,* but because it seems to belong to the period of J ^ conquest or of the Church revival that ensued, iH'culittrity that occurs, not in every lobed tank, but h to make it common, indicates that this style . :a some earlier model. The sequence may be coiutaieutly illustrated from the fonts in the Bardo Museum and in the ancient basilica at Bir bou Kekba, that seem to belong f ' id and are of different shape but have the same -..ey are small circular tanks s\v^'^■ i'^ t)-,.. fl..nr vvith two flights of steps facing one ano; iji the floor level into the cuvette. At Bir bou Kekba of full size and obviously intended for the descent .- into the water on one side and the ascent on the ' r the administr^ittac \^-t^il^ Sacrament. :i^'^^1 '.iria in Cosmedin. The names and purpose of t! ions are too well .teBBO^^Q the examples just given and ! ancient and modern writers on ecclesiastical archasology Meed repetition herjH^^^fife' choir enclosure was called the >ia c,^^i^^and M. Gsell found an ins^^i^on in an Algerian church showing that a part screened off for women was called the ■■ ;incellus.■^*g?nium.*»■»f»^l*M»^^reek and-^O^ptt^hurches of to- , ^ the women's quartSS^^^ also screened off, but they are . > 11 Y placed in the kind of gtJlery or triforium tliat is only occa- ' alK iiicl.with m some of tlie African chuivlu^ as tor mstance m the basilica at Tebessa. Whether these ancient churches in North Africa ha !>ne altar or not is an open question, but the ruins rtainly testify that the multiplicity of altars cor i wi I' Coptic and Latin churches of to-day I 'eady mentioned certain chuiT^ « with q>sed chapels in the form of a d these, . had a single altar apiece for minor or fvccai^ional services on to the high altar in the nave. But, an a general rule, it ' the churches had but a single altar, and this was invari- M the second or third bay of the nave, counting from the ciiief apse.' It was not raised up on steps according to iiitt luudem Western fashion but stood on the floor level and was • Lfjt Anliquiiis de lAlgMe, vol. ii. p. 148. %Q ^"%»\ %^\\ oT INTRODUCTION 99 apparently either like a Coptic altar, a mass of solid masonry, or, as with ns, a Communion table of wood or stone with legs and top. It was usual for the table slab to contain relics of saints, and the top of a stone table, showing the kind of receptacle where these were placed will be found among the illustrations of the Sbeitla churches. These early altars were I expect furnished in much the same simple way as those of the Copts are at the present time. The mosaic representation of an altar in the west wall of S. Sofia at Constantinople has already been referred to. In the larger churches the clergy were accommodated in the central apse and sat facing the congregation. The seats were either arranged in tiers, as at Announa, S. Eirene at Constanti- nople, and Torcello, or more commonly in a single row with the bishop's seat in the middle, as in the basilican churches of Rome and the Coptic churches of old Cairo, and the Nubian churches at Faras and Debereh. This space allotted to the clergy was invariably raised as a dais above the floor of the nave and approached either by a central flight of steps or a flight on each side of it. There remains the space between the apse and the altar to be accounted for, and this was probably occupied by the pergula or open screen used for hanging lamps. This kind of screen will be found at Leprignano in Italy or S. Maria in Cosmedin at Rome. In some cases this seems to have been a place of interment for the saint or martyr by whom, or in whose honour, the church had been founded. At Dougga a sanctuary of this kind was placed in the crypt under the chancel. At Bir bou Rekba it seems that a memoria was actually under the altar, and that was, I believe, the more common practice. There were two systems of providing shrines for the remains of saints or distinguished persons venerated in connection with a church. First, what I may call the crypt system, as at Dougga, and familiar to us in the arrangement at S. Miniato and other European examples, where the shrine is placed under a raised chancel ; and secondly, the overground system where the interred were placed in a chapel. The most interesting examples of the latter were built apsewise at the extremity of the church opposite the high altar. This arrangement is not uncommon in 100 NOETH AFEICA Germany but unusual elsewhere.* The examples here, at Sbeitla, and at Uppena, must be among the earliest in existence. The destruction of these churches has been so complete that practically nothing remains in the way of architecture. We may, however, get a very good idea of what the churches looked like from the basilicas of Eome, and from other well known buildings such as S. Apollinare at Kavenna, or S. Gavino at Portotorres. The church at El Kef, of which more remains than of others in Algiers and Tunis, shows that the aisles were covered with cross vaults, the voute d'aretes, and the nave with timbers and tile roofs. The most interesting feature in roof and arch building was the use of small tubes of pottery f made to fit into one another in the manner described in the preceding volume in the notes on El Gebioui. This ingenious device, of Eoman origin, is still resorted to by the natives of Southern Tunis, in the districts of Gafsa and Tozeur. The basilica at Tebessa had a triforium or gallery, and must have resembled S. Agnese at Eome. These galleries are occasionally met with in Africa, but they appear to have been the exception rather than the rule, and occur only in the larger churches. They are, however, common to the Coptic churches in Cairo and are reserved for the women. The churches in general are built of lime or other stone quarried locally ; some of it in the district of Feriana is poor stuff and weathers badly. Marble was only used sparingly for decorative details such as capitals, pillars, fonts, or for veneering. Most of it seems to have been imported from abroad. The pink marble seen in the font at Sbeitla and elsewhere was no doubt obtained from near Ghardimaou, in Northern Tunis, where other varieties and shades of this marble are also found. But the cippolino, which frequently occurs in large blocks, certainly came from Europe. I have found here and there, a few pieces of red and green porphyry from Egypt notably at Bir bou Eekba. Glass mosaic seems to have been fairly common for covering the roofs of domes, or semi-domes, and for decorative panels, as for example the Tebessa ' jewel.' * Mainz Cathedral for example. Also S. Gavino at Portotorres in Sardinia ; illustrated in vol. i. t Illustrated in vol. i. Plate 70. INTEODUCTION 101 As the earliest Christian settlements beyond the sea-coast were in the plains of central Tunis, behind Sousse, I may begin with those churches that are situated on the Sbeitla railway, taking the latter city as a centre and starting point. Owing to the season of the year I was unable to proceed from Feriana to Tebessa across country, but the distance is not great, and as all the towns described in the Sbeitla group were on the Koman roads leading up to Tebessa, where the legion was quartered, Tebessa with Announa and Timgad, in Algeria, form the second group. The towns on the road from Tebessa towards Carthage, Haidra, El Kef, Dougga, and Tibar form a third group, and the sea coast towns, Uppena, Sidi Abich, and Carthage, a fourth group. The two churches at Uppena, on the whole, are the most important for the study of early Christian archae- ology, because of the wonderful collection of mosaic tombs they contained. CENTRAL TUNIS SBEITLA The ruins of Roman Sufetula, now known as Sbeitla, have been frequently described and illustrated by travellers, and it is beyond my purpose to do so here.* It seems strange, but apparently the fact that nothing is known about the city at the time of its greatest prosperity in the reigns of the African Emperors, Septimius Severus and Caracalla ; but the triumphal arches, the triple temple, and numerous ruins in the neighbour- hood show that under the Roman rule it must have been a flourishing provincial town in the centre of a rich agricultural district. During the later Empire we hear of Sbeitla as the capital of the Patrician Gregory who had revolted against the grandson of the Emperor Heraclius, and the defeat he suffered at the hands of the Saracens was the first of a series of disasters that led to the ultimate loss of the North African provinces to Christianity and civilization for nearly twelve hundred years. The principal churches in Sbeitla are a large basilica with a cloister and baptistery rebuilt during the Byzantine period ; another and larger basilica with a baptistery and a chapel attached to it, probably dating from the early Christian period ; a small basilica on the north-west corner of the city, and the chapel in the court of the triple temple that was converted into a fort by Justinian's generals after the conquest of the Vandals. Church of Bellator The Byzantine basilica, of which much more is left above ground than of the other churches, lies to the north-east of the * Since these notes were made the Sbeitla churches and their dependent buildings have been described by M. Merlin in the publication of the Department of Antiquities at Tunis. Notes et Documents publies par la Direction des Antiquites et Arts 1912. Forum et Eglises de Sufetula, par A. Merlin, pub. Leroux, Paris. SBEITLA 103 triple temple. M. Merlin identifies this building as the church of the bishop Bellator, from a mutilated inscription found in the western aspe, and probably belonging to one of the sarcophagi. The name Bellator has not hitherto figured in the list of bishops of Sufetula, and his date is not known. This church can be recognized by the south wall that still remains intact from plinth to eaves. This wall is made, like the fort, of large stones taken from Koman buildings. The ruins show that there were two churches, and the greater part that remains, the walls, the sanctuary, and the choir, belonged to the second or Byzantine edifice. In these notes it is the second church that I am more particularly referring to. In fact the two churches occupied the same site and were of the same width but the later church was, as the inner apse shows, rather shorter than the earlier church ; also the floor of the earlier church seems to have been at a uniform level throughout, while the sanctuary of the later church was raised on a platform or dais, and the place of the altar shifted further east. The body of the church was divided into a nave 24| metres long and 16^ metres wide, single aisles on each side of it 4| metres wide, one large semicircular apse for the sanctuary at the east end, and a similar but smaller apse used as a funerary chapel at the west end. The latter was enclosed in a square outer wall and had a chamber on each side of it. The choir, occupying the last two bays of the nave and raised 60 centimetres above the nave floor, is a later addition. The principal entrance was by a door in the west end of the south wall ; it had a porch in front of it, and the bases of columns and a piece of a pavement along the outside of the wall seem to indicate that there was a kind of cloister court on the south side of the church. Two narrow doors on the north side give access to a more important cloister containing a baptistery. The view of the inside is taken from the south-west corner of the church showing that the nave was separated from the aisles by pillars of marble in pairs ; these supported a roof made of timber and tiles. The floor was covered with mosaic pavement and here and there stone footings mark the gangways that divided the tribunes, choir, and the part allotted to women. This division is found with little variety in all the earlier churches I have seen in North Africa. In the middle of the 104 CENTKAL TUNIS nave, in front of the two steps leading to the raised choir, there is a square enclosure to mark the place where the altar of the original church stood, for, as the floor shows, the raised choir or chancel beyond is a later addition. Proceeding up the centre of the nave we come to two steps leading to the raised floor of the chancel. About a metrebeyond them a line of stones laid across the chancel marks the screen that surrounded the later altar; the latter stood in the middle of the chancel. In the chapel of the prothesis I noticed a piece of a fluted pillar made of white marble that may have been one of the legs belonging to an altar slab lying close by. The places for the four legs to fit in are seen in the four corners, and the larger horseshoe-shaped recess in the centre was made to hold relics. Beyond the altar of the second church there are the founda- tions of an apse made of rough rubble. This was the end of the second or later church which was therefore a good deal shorter than the earlier one. The original apse made of stones lies a little beyond it to the eastward, and it is made of large dressed stones and contains the foundations of the seats of the clergy arranged in a semicircle within it according to the usual plan. The original aisles terminated in square walls, and there is no sign that they were pierced with doors to lead to the two chambers usually found in North African churches. M. Merlin's plan shows the central apse projecting from the rest of the building and not contained in an outer square wall. This arrangement is so very unusual that there may be some doubt as to the correctness of his plan. It would seem that the raised floor in the north aisle of the chancel that belonged to the second church indicates the chapel of the prothesis. The floor on the south aisle is not raised to correspond with it, and there is there- fore no indication of the diaconicon. It was, however, not unusual to have one chapel only. Eeturning now to the west end, the apse which is rather smaller than that in the chancel had been excavated and con- tained two or three coffins of stone, one of them nearly 2^ metres long. These probably belonged to the Saints or Bishops for whom the apse had been built. M. Merlin records the dis- covery of as tone with the following inscription engraved on it : BELLATOR EPCS vixi(t)inp(ace). . . . This probably be- SBEITLA 105 longed to one of the sarcophagi. Above the inscription was a cross in a circle. The approach to this apse from the nave was by a flight of three broad steps flanked by two pieces of cornices, from a Roman building. A small door on the north side of the church leads out into the cloister of the baptistery. In a corner of the cloister on the floor is a large sarcophagus with the cover lying by it. The cover has a Latin cross in an oval medallion, and an inscription in well cut capital letters giving the name of a bishop in the beginning of the fifth century. The only letters I could read were HIC, and below them EPSC, but M. Merhn was able to decipher the legend and renders it thus : HIC INVENTA EST DP SCI JVCVNDI Erse PER INQVISITIONE AMACI EPSCPI Near by is a console carved with a crude representation of peacocks drinking from a chalice,* and another stone carved with a vine pattern. The ground plan will give some idea of the general appear- ance of this baptistery and cloister. The font enclosed in a balustrade, and apparently covered by a canopy on pillars, was of the usual hexagonal variety with two sets of small steps lead- ing from the floor level down to the basin. At the south side the balustrade projected into a small apse. The limestone columns and their capitals are of the late Eoman type and apparently were newly made for the building.! Church of Servus About a hundred and fifty yards south of this basilica there is another group of Christian buildings consisting of a still larger basilica lying north and south, and a baptistery and chapel adjoining it all built at different dates. The basilica is identified by M. Merlin as the church of the priest Servus. Of the former * Illustrated as No. 35 on Plate 71, to vol. i. f Merlin : Notes et Documents, Forum et Eglises de Sufetula, p. 42. 106 CENTKAL TUNIS nothing but the pavement and the foundations of an apse containing empty stone sarcophagi remain. Two or three inscriptions were found here, and among them the following dedication on the lintel of a door in the words, hic domus ORATiONis. The epitaphs to Servus, who was a priest and one Natalika, are in the usual phraseology. The complete destruc- tion of this church makes it impossible to form any idea of the general plan or of the internal arrangements, but it must have been a very large building and probably had a nave and double aisles on each side. The Baptistery The adjoining baptistery was built in a large square Roman edifice of which only the angles remain, and as the photograph shows it looks like the remains of the cella of a Boman temple and similar to one at Ain Tounga. This was closed on three sides and open on the west side facing the basilica. In the centre stood the font enclosed in a little area with a low wall screen round it. The font was divided into three parts ; the first or upper part is hexagonal, and on two sides are little flights of steps leading to the second or middle part made in the form of a cross with four lobes each containing three steps leading down into the cuvette. The whole was covered by a canopy supported on pillars that have now disappeared, though their bases can be seen in the angles of the screen. This floor as well as the screen was lined with a veneer of pink marble, while the cuvette was lined with white marble. Lying about among the various fragments of decorative carving is a block of stone with a dedication cross carved on it.* Near the baptistery were several tombs and one with an inscription to Viola with a peculiar form of G made like a J. Adjoining the baptistery on the south side is a small church divided into nave and raised chancel terminating in a single * Examples of this kind of cross will be found all over the Christian world. There are three in the chapel of Guildford Castle, Surrey. See No. 49 on Plate 73 of vol. i. SBEITLA 107 semicircular apse enclosed in a square wall. In the floor of the latter I noticed the stumps of the feet of the altar ; these were fluted pillars of white marble set in the mosaic floor. The nave and aisles were separated by a marble screen supported by grooved posts with pear-shaped tops, and some of these are still lying about on the ground. There is some doubt whether the building with three apses on the north-west side of the city was a church. M. Saladin so describes it and judging by the ground plan, the general appearance of the ruins that clearly belong to a building of relatively late date, and some ornamental carving that is cer- tainly Christian he would appear to be right. The central building consists of a nave terminating at the west end in one large and two small semicircular apses made of brick. The nave is about eight metres square and as the footings show there was a narthex or porch at the east end with a square forecourt in front of it containing a colonnade or cloister. This central building was closed in on the north, east and south sides by rooms and halls of different sizes. On the south side there are two long rectangular halls with semi- circular ends like the so-called bath at Ain Tounga. Among the out chambers on the north side I noticed several mangers or troughs of the kind found in the so called stables of the monastery at Tebessa and at Haidra. The only parts of this building that remain more than four feet above ground are the brick pilasters of the chancel arch and portions of the main apse. It would seem from what is still left of the walls that the latter was decorated with niches and windows placed alternately, the niches being like those in the chancel of the church at Tibar. On the wall of the little north apse are some well-preserved mosaics with a flowing vine pattern ; the two little shafts to hold a screen in front of the apse are still in situ. The floors of the nave and apses are covered with mosaics Only fragments are distinguishable, but I noted several figures, a boy riding on a dolphin, a child walking with an adult, and 108 CENTRAL TUNIS the head of a woman, perhaps the Blessed Virgin, and the angel of the Annunciation. There is also a border pattern like that in the apse in the tomb of the martyrs in the basilica at Uppena now in the modern church at Enfidaville and on a piece of pavement on the walls of the museum at Sousse. I could not find the signs of the zodiac that are said to be there and have been taken to indicate that this building was not a church. In the church a number of Corinthian capitals are lying about and in one of the chambers on the south side the carved stones figured in the preceding volume on plate 71. Among the objects represented are a cross in foliage, a chalice and fish and, a pretty design of scrolls with whorls and rosettes. The Fort After the defeat of the Vandals, Justinian's generals con- verted the forum, containing the triumphal arch and the triple teinple into a fort. This arch and temple are too well known to need description here. The north-west corner was made into a chapel and a little apse built out in the curtain wall. The general arrangement of the chapel can still be distin- guished from the masonry and the footings, and among the fragments lying about is a stone capital of the Byzantine period with eagles at the four corners, illustrated as No. 25 on plate 69 in vol. i. BEYOND SBEITLA Since the opening of the railway from Sousse to Sfax there is a continuous line round the great plain of Central Tunis by Metlaui and Gafsa. Across the circle so formed a high road is being constructed from Sfax to Sbeitla in a north-westerly direction along the line of the modern aqueduct ; this road will eventually open up a district full of Roman remains that has hardly been explored by archaeologists. After leaving the little rest house at Sbeitla the traveller who ventures further into the Interior must be prepared to rough it in the full sense of the word for there are no civilized Second 0y sktti. 30 f'^T^BMKr',- iil^i^. -X:.- S*^ SBEITLA. A mosaic pavement^ from the West church ; p. 108. Altar slab showing the reliqp.ary in the centre and holes for supports^ p. 104 ; and oil presses at Hcnchir Chond el Rattal betivcen Kassciiuc and P^riana. u "^i- J^ To face page 108 108 CENTEAL TUNIS **^ the head of a woman, perhaps the Blessed Virgin, and ^^^ ai.j^^i of the Annunciation. There is also a border pattern like that I'' ''" he tomb of the martyrs in the basilica at Uppena ri<»w in me j ' ' .h at Enfidaville and on a piece of piivemtMit oi , . : the musemn at Sousse. I could not find ti tbe zodiac that are said to be there and have I . -.ndicate that this building was not a church. In a number of Corinthian capitals are lying about )f the chambers on the south side the carved stonch fie preceding volume on plate 71. Among the ^^*: ' are a cross in foliage, a chalice and fish and, '!> i'"^ -crolls with whorls and rosettes. The Fort lefeat of the Vandals, Justinian's gen im, containii^:itt\'afiamphal arch and the triple ."661 .\'{^l5^xw^Tte^^<^k\M¥\\t^WJRl§v^«4^e^-^lj«<^s#'%IJ known \iWVN S^\itv^WfJWWi44\% >fl^'«l'ilWJgesi6»feicift. ti^vjsdiaiidk^wan \«tk.i\o^ distiu- ' the masonry and the fo(jtings, and among the iig about is a stone capital of the Byzantine period ^ at the four comers, illustrated as No. 25 on plate 69 BEYOND SBEITLA opening of the railway from Sousgu^ to Sfax there s line round the great plain of ^»j '"a. Across the circle so fo ' i^ 'd from Sfax to Sbeitla iii i ly : line of the modern aqueduct ul will n a district full of Komiin remains that has by archaeologists. little rest house at Sbeitla the traveller who venttires farther into the Interior must be prepared to rough it in the full sense of the word for there are no civilized ?)0\ ^"^a^ ^-)v\ oT Second fly sheet. 30 "^-S. To face page to8 O} 0£ Aa^i "^ViiWi^?, %oi ^S^^ -itiiX oi KASSERINE 109 resources, and food and drinking water have to be brought up from Sousse by the daily train for the railway staff and the few European colonists. The country, however, is inhabited by nomad Arabs who live in tents and wander about with their flocks from one pasture ground to another, and there is a large and prosperous native population all along the line. By day the monotonous landscape seems to be quite lifeless and deserted, and the passing traveller would hardly credit that there was a living soul about, but at sundown the whole country-side is illuminated with countless camp fires dotted about as far as the eye can reach. From Sbeitla the line continues to rise very gradually for about twenty miles as far as Kasserine, another important Roman town commanding the Southern approach to Thala, Tebessa, and Northern Tunis. The choice of position for Kasserine and many other Roman cities, like for example Thelepta, Henchir Gebeul, Haidra and indeed Sbeitla itself, is to be accounted for by the configuration of this country. The land lies in a series of plateaux or plains varying of course in size, but as a general rule from twenty to thirty miles long and ten to twelve broad. These plateaux commencing at the high level of Tebessa descend gradually like a series of steps to the plains of Kairouan at sea-level. In the upper plateaux there is a considerable rainfall and the rivers formed by it pass from one plateau to the other, from the higher level to the lower through small cafions or gorges, eventually losing themselves in the chotts or marshes about Kairouan and central Tunis ; few if any of them reach the sea. As a rule the roads followed the river beds, and near the cafions, where the water could be easily dammed into reservoirs, the cities and larger towns were built. These cities formed a series of fortified out-posts linked up with the head-quarters of the Roman legion stationed at Tebessa, Timgad and Lambessa. Kasserine Nothing is known of the history of Kasserine as a Roman colony, but the name and rank are preserved to us from an inscription on the front of the principal gateway or triumphal 110 CENTEAL TUNIS arch on the east side of the city. The other important remains are a barrage, a water tower, an early church and some tombs. Among the latter is the well-known monument to members of the family of Flavins Secundus. On this tomb there is the sententious and portentously long inscription in Latin hexameter verse reproduced by M. Guerin in his book. The church was built of Eoman materials and nothing remains of it but the west fagade with two doors and the footings of the nave, chancel, and single apse. In front of the doors at the west end was a porch, and close to it are a monolith tympanum very roughly carved with birds and a vase or chalice. The interior has not been excavated. There were no side chapels, and the sanctuary is at the w^est end. Kasserine station is about two miles east of the town and near it are the remains of several oil presses showing that in Roman times olive-trees were grown in this district. After leaving Kasserine station the line ascends on to the escarpment of the upper plateau and then continues for twenty miles on fairly level ground to Thelepta. The north side is bounded by the mountain range called the Chambi, and on the south by downs that separate the plateau from a large valley at a much lower level. From the maison cantonniere, about the eighth kilom. on the line, a path leads across the plateau in a southerly direction to the Dernaia Pass through these downs and so into the valley I have just mentioned. The path ends at a place called Haouch Khima Mta Darrouia, where there is an interesting early church discovered by M. Saladin. HAOUCH KHIMA MTA DABBOUIA This is the name given to the site of a buried Eoman city about 20 kilometres due south of Kasserine. It is situated in longitude 7' 5" E. and latitude 38' 85" N. The best way to visit it seems to be to sleep at a maison cantonniere on the railway half-way between the stations of Kasserine and Thelepta. It should be about four hours' ride from there across the moun- tains bounding the Kasserine-Thelepta plain on the south side. It is a small but substantial building of large dressed stones. p ■* FfDM and apse. '"' '* Photnirraplis taken by the department of .Intiqiiities ^ Moiis: Merlin. Moils: Saladeiis sketch of reconstructed chapel. To face page I jo no CENTRAL TUNIS arch on the east side of the city. The other important remains !' ' barrage, a water tower, an early church and some tombs. A he latter is the well-known monument to members "' illy of Flavius Secundus. On this tomb there is the is» and portentously long inscription in Latin hexameter verse reproduced by M. Guerin in his book. The church was built of Romati materials and nothing remains of it but the west facade with two doors and the footinp:^ of the nave, chancel, and single apse. In front of tho di the west end was a porch, and close to it are a ipanum very roughly carved with birds and a ...t.e. The.tilsKie^lCit fel^\5c©^tteen excavated. There -ide chapels, al^<3^il^^i^'i^(t||^ry is at the west end. i!it= station is about .two. imlesjieasttaf the W)wn and . tlfe rem^ms Q|^^s^erajl^^^^ presses si iti Mies olive-trees were grown in this 11,0^^^,^. ,k;it;r ' .^'^Wii^itsliR^fixcfe^ ^^:,W^H'^mk.to ,l^\fescarpment upper plateau and then continues for twenty miles on ! ground to Tlieleptjjt . . -utain ran^Je that sep level. F on the lili •n to the illey I h.iy> Haouch/ in Ka The north side is bounded by inbi, and on the south by -a, Mrge ' valley at a much ,. abotf tr the eighth i-^au in a southerly j^^ns and so into ends at a place is an ing early'''6}jafc he^^^there adin. MTA DAIiliOUU the name given to the site of a buried 1 , ^_, kilometres due south* of Kasserine. If iied in 5" E. and latitude 38' 85' " best way u iO be to sleep at a ii niuere on the rai Dotween the stations l ^^ ^luj and Thelepta. It should be ab'.iit four hours' ride from there across the moun- tams bounding the Kaeserine-Thelepta plain on the south side. ^ small but substantial building of large dressed stones. 0\\ ^"^R^ ^:iv\ o\ ^^^^^^^^^H^ ' " ««*^*Jta ^ta ni ^^^^KL Hi -'■* tw. ■\ K. r^gMI \-S^^^. ^^-^ A 4 .1 ^^: ^ ^HlS^ HAOUCH KHIMA 111 The construction is altogether superior to that of any of the buildings I have been describing, except Tebessa. It consisted of a square nave covered by a cross vault, and a single semi- circular apse covered by a semi-dome, and a shallow narthex or porch with columns in front of it. The general appearance of the building can be best understood from M. Saladin's essai de reconstruction. It bears a strong resemblance on a larger scale to the upper part of the mausoleum at Haouch Taacha I have already referred to. It is obviously copied from a Roman building and, from the style of masonry and the resemblance to some of the chapels at Bin Bir Kilessi in Asia Minor, is probably one of the earliest Christian churches built in the North of Africa. About three miles from the maison cantonniere along this path, and about half-way to the Dernaia Pass are the remains of a large agricultural establishment.* These remains consist of nine sets of upright stones, 2| metres high, arranged in pairs with a stone laid across the top of each pair. They originally occupied the centre of a large stone building of which the debris are lying about on the ground. In front of each pair on the south side is a little circular platform with a groove or gutter leading to a tank or reservoir. The purpose these stones were intended to serve can be best explained by reference to the accompanying rough sketch suggested by M. Saladin's drawing from which I have purposely omitted the minor details. There seems to be some doubt whether these presses were used for oil or wine, but the natives of to-day adopt much the same kind of plant for pressing oil, substituting a screw-jack worked by a long beam for the vertical pressure operated by means of ropes and pulleys. The strength of the plant suggest that oil, and not wine, was the produce treated. The purpose of these stones is, of course, obvious enough when the method of using them is explained, but it must be said, in defence of an eminent antiquary, that when seen at a distance an isolated pair does look remarkably like a prehistoric dolmen. Students of North African archasology may recollect that in 1897, a traveller set out to prove that some upright stones of this * This seems to be the place visited by M. Saladin and described in his report under the name Henchir Clioud el Battal. 112 CENTKAL TUNIS kind were the horns of altars to which the ' sacrifice ' was bound, the circular basin in front was the altar itself, and the gutter or groove was to carry away the sacrificial blood into the tank below.* At present the neighbouring plateau is used for pasturage, and a little wheat is grown in places, but there would seem to be no great difficulty in replanting and maintaining olive- trees, for there is sufficient moisture to make a marsh and justify, as my guide explained, the name Garaat Krechem el Kelb f being applied to a small lake which is always wet. Tlielepta The station at Thelepta is the only habitation on the site of one of the largest Koman cities in this part of Tunisia now represented by the modern village of Feriana about a mile further west on the southern escarpment of the plateau. The principal ruins of Thelepta are the Eoman baths containing two trefoil chambers made of brick, a group of four pillars with their consoles and the fragments of a large church in the citadel. There has been little or no excavation here, and the site offers one of the most promising fields for research. The most important Koman remains here are those of the thermae, a large brick edifice situated on the bank of the river and just above a small canon where it has forced its way through the escarpment and so past Feriana into the valley below. These baths, of which I give a plan taken from M. Saladin's notes, and a photograph, are now in ruins. There are two details of interest in them. First, the trefoil apartments, and secondly the vaults which have been built on little tubes of pottery X fixed in the cement. In one corner there is the spring of a cross vault built in this way and almost exactly like that of the chapel at El Gebioui. The size of these thermae show that the city must have been large and important. At the * The Hill of the Graces ; a record of investigation among the trilithons and Megalithic sites of Tripoli. Published by Methuen, 1897. t By interpretation ' the lake of the dog's nose.' I A photograph of one of these tubes will be found oh Plate 64 in vol. i. 3a AIN TO UN G A. Castle and view of inside of a bastion showing method of consti'nctinn. HEN CHI R GEBEUL. Details of a Roman building- probably a hath^ with canopied niches. 112 CENTRAL TUNIS kind were the horns of altars to which the * sacrifice ., ^^ the circular basin in front was the altar itself, and ti or groovo was to carry away the sacrificial blood into the tank below.* At present the neighbouring plateau is used for pasturage, and a little wheat is grown in places, but there would seem to be no great difficulty in replanting and maintaining olive- trees, for there is sufficient moisture to make ' and jiwrifv as. my guide explained, the name Garaai .. " '^I l>eing applied to a small lake which is always . Thele/pta The station at Thelepta is the only habit of one of the largest Roman cities in this now represented by the modern village of Feriana ab< mile further w^est on the southern escarpment of the plateau. The principal ruins or j.iiejepTa are the Roman baths pillars with their consoles and the fragments of a large church in the ciiadel. There has been little or no excavation here, and the site offers one of the most promisir" >"■• i'^- <'■^'• research. The»*te§sf^ i%6?l^to^^!^Mii»f^ft'!^ifts^giS^^ll% those of the thermae, a large bric^^^^'^^mM^'^^ the bank of the river and just above a small canon where it has forced its way through the escarpment and so past Feriana into the valley below. These baths', of which I give a plan taken from M. Saladin's notes, and a photograph, are now in ruins. There are t if interest in them. First, the trefoil ap'^^" ■•"•*- 'y tlie vaults which have been built or- fixed in the cement. In one con spring cross vault built in this way and aii v like that ol the chapel at El Gebioui. The size v thermae show that the city must have been large and important. At the the Graces ; a record of investigation among the trilithons and Mef.' s of Tripoli. Published by Methuen, 1897. t By II ^1 lion ' the lake of the dog's nose.' J A photograph of one of these tubes will be found oh Plate 64 in vol. i. S\\ %-^»\ ^:jvj\ o\ THELEPTA 1I3 top of a little glen above them stands the citadel, a square enclosure of about six acres. At the west end stood the forum or chief public buildings, and at the east end the church. The site of the church is undisturbed, and the shafts of the pillars still stand aslant at various angles in the debris of the roof, showing that the building was not demolished but merely abandoned and left to perish away. It was a basilica with a single apse facing east and a nave with aisles separated by a row of pillars with corinthian caps. The large number of mosaic tesserae lying about came probably from the roof of the apse. Beside this, no doubt, the principal church in the city, I found the footings of another and much smaller church with a single apse on the south side of the citadel. Another interesting spot in the site of Thelepta are the stone quarries where the materials for constructing the city were obtained. These are situated just above the thermae and near the river cafion I have just mentioned. The process of stone cutting can be seen in all its stages; finished and unfinished blocks are lying about ; others still undetached have the wedge holes ready drilled for splitting them off the rock. This process was effected by boring holes and inserting dry wooden wedges that were afterwards made to expand by being soaked with water and so split the stone from the rock ; the face of the quarries is covered with the marks of the workmen's tools. Feriana is a little oasis village. In the public square a number of carved fragments were collected and put together by the officers stationed here while the line was in construction. The examples on Plate 71 of Volume I show that the native artists developed a style of their own in which the objects they represent are more naturally and less conventionally rendered than in the pagan models. The carving, of course, is not executed with the same mechanical precision but the result is a great improvement upon the stiff and lifeless work of the preceding period. Henchir Gebeul Henchir Gebeul is the Arab name for the site of an abandoned and ruined Eoman city situated about twelve miles west of 114 CENTEAL TUNIS Feriana. I gratefully record the kindness and hospitality of Dr. and Mme. Boricaud, of Feriana, who received us into their house, made us welcome, and enabled us to see Henchir Gebeul. We visited the place in company of the brigadier of gendarmes, a genial Corsican, who drove us across the steppe in a high two- wheeled dogcart. Archaeology apart, the chief items of interest on our adventurous drive were supplied by the brigadier's grey charger, who made a first appearance in harness on this occasion. The Koman name of this place is not known, nor are the circumstances or date of the foundation and ultimate destruction. The position shows that the city was built to guard one of the approaches leading from the desert up to Tebessa, and judging by the size of the public buildings, and the unusual care and skill bestowed on some of the capitals, consoles, and other carved fragments lying about, it must have been a place of no little importance. It is now a weird and ghostly place, and the stately ruins in this remote and forgotten spot are more impressive than any I have seen in the region. The dead silence was broken by pleasant shouts of welcome from a company of nomad Arabs who have taken possession of the ruins. The principal building stands at right angles to the river-bed on the west bank. It consists now of two parallel walls of fine masonry that apparently represent the sides of a hall or nave. The lower parts of these walls are hidden in debris of the roof that has fallen in and covered the floor, but sufficient of the upper parts of them is left to show that they were decorated with a row of eight niches on each side. All that remains of the upper stage of the building are the piers, or pilasters, on each wall, and a close inspection of them shows that the nave roof was a barrel vault. On the outer sides of both walls and these pilasters there are indications that the nave was flanked by lateral aisles, and that they too were covered by barrel vaults. And that, I am afraid, is all that can be said about this odd building until the debris have been removed and the site exca- vated ; the approximate dimensions will be seen by the little plan taken from M. Saladin's account of it. M. Saladin was the first archaeologist to visit this site, and he describes the ruins in his report to the Minister of Public Instruction made in 1882. He draws some comparison between HENCHIR GEBEUL 115 this ' hall ' and the so-called stables at Haidra and Tebessa ; but it seems more probable that the hall here was either a temple, a court of justice, or even a public bath. About 20 metres west of it are several Roman Corinthian capitals and some consoles of unusually good workmanship, executed in local stone ; these probably belonged to a group of public buildings that we might expect to find in a forum or citadel. The city covers a small area, but the importance of it should rather be judged by the excellence of the carved fragments lying about. No attempt has been made to excavate, and as there is no modern city for many miles around the materials have not been disturbed, but lie just where they fell as it would seem in some great earthquake. The country to the north of this place and in the direction of Tebessa is full of Roman and Christian remains. But the district is so inaccessible, it is so hot in summer and so bitterly cold in winter, that it has remained almost entirely unexplored, and but for the police and government officials whose duty obliges them to visit it, the existence of these remains would be unknown. EASTERN ALGERIA TEBESSA In describing some of the smaller unknown churches in North Africa and Egypt, I have frequently referred to the monastery at Tebessa, especially to the basilican church and the trefoil chapel attached to it, because the approximate date of their construc- tion is known, and being comparatively well preserved, so far as foundations and ground plan are concerned, they give some idea of the internal arrangements of a Christian church and memorial chapel in the Roman period of Church history between the reign of Constantine and the Vandal conquest of North Africa. The monastery, however, has been so often and so well described by French archaeologists that I shall not attempt to give a detailed survey of it here, or more than sufficient to enable the student, with the accompanying photographs and plans, to identify the various points referred to in the notes on the smaller churches and chapels elsewhere. I may first remind him that Tebessa was one of the chief military stations and centres of Eoman civilization in North Africa, and that all the great roads from Carthage and North and South Tunis to Timgad, Lambessa, and the Algerian plateau passed through it. It is evident from the age and importance of this monastery and the great basilican church that at a very early date Tebessa also became an important centre of Christianity. It is an open question whether the material of the ruined fabric of the basilica as we now see it belongs to some pagan building, be it a court of justice, a temple, or other public edifice erected before the reign of Constantine. M. Gsell reckons it among the edifices du culte Chretien, and it is difficult to under- TEBESSA 117 stand how any one acquainted with the usual arrangements of an African church coul^ suppose that it was intended for any- thing but Christian worship. The original purpose the trefoil chapel was built to serve, be it secular or Christian, is not quite so clear, but it is certain that at an early date, and before the Vandal conquest, it was used for Christian burial. The monastery stands a short distance from Tebessa and near the high road leading to Haidra and El Kef. This road leaves the city by the East gate and passes under a Boman triumphal arch dedicated to Septimius Severus, Caracalla, and Julia Domna in the third century. It was restored during the Byzantine occupation, and contains an early example of a squinch made of small stones built at that time to support the roof.* The buildings and surrounding grounds of the monastery were enclosed in a high curtain wall strengthened at intervals by square bastions on the inner side ; according to some archaeo- logists this peculiarity indicates that the wall was built before the Byzantine conquest. The basilica stands in the centre of this enclosure. On the west or front side of it there is a great square court with indica- tions of a paved cloister round it. On the south side of the basilica are the trefoil chapel, a baptistery, and small church of the Byzantine period. The so-called stables stand in the north-west corner of the enclosure and bound the great square court on the north side. Bound the walls of the basilica on the north, east, and part of the south sides, are a number of rooms built of large blocks of stone supposed to be cells of the monastery and built subsequently to the basilica itself. Passing from the El Kef road through the main entrance of the curtain wall, a fine gate of which only the arch remains, we enter the court by a paved avenue. Standing in the middle of the court and facing east, we obtain the first view of the basilica, taken by my father. This faces directly on to the west front of the basilica, showing the steps in front and the nave and the central apse * See Gsell, vol. i. p. 183. A.ccording to M. Gsell this squinch supported a dome and is of the same date as the arch itself. Examples of domed gates occur at Tripoli and Latakkia in Syria. I venture to differ as to the age of the squinch, and attribute it to the Byzantine restoration in Justinian's reign. 118 EASTEEN ALGEEIA in the distance. The massive vaulted roof in the distance on the right side of the basilica belongs to the trefoil chapel. Before v^^alking up the steps into the church, the visitor will notice a row of stone posts with grooved sides shown in the fore- ground of the pictures, and also two structural peculiarities in the construction of the church. These posts mark the border of the court and a broad paved avenue that led from the entrance gate along the front of the basilica. They were joined by panels fitted into the side grooves, and so formed a continuous balustrade. The same kind of post to support the altar screen will be seen inside the basilica, and was commonly used for that purpose in North African churches. These posts were often provided with fircone or acorn-shaped tops, and some in use may be seen in the choir of the Cappella Palatina at Palermo. The first structural peculiarity I have referred to can be seen in the photograph. It will be noticed that the floor of the basilica is raised some three metres above the ground level. This substructure appears, from excavations made in search of tombs and treasure, to be quite solid, and must have involved an immense amount of labour and material. This device of raising a building on a platform was commonly adopted by pagan architects, but it is not usually found in North African churches. All the other buildings of the monastery including, oddly enough, the trefoil chapel, are built on the ground level. The second peculiarity is not seen in the photograph, and is not noticeable on the spot, but the measured plan shows that the front of the basilica and the balustered front of the forecourt are not in alignment. It has been argued from this that the court and the basilica were built at different times. We now pass from the court into the church by fourteen steps about 20 metres broad, and arranged in three flights. The row of rough stones at the top marks the foundation of a porch or narthex flanked by two towers that formed the west fa9ade. Nothing of this now remains but the bases of a few pillars and some of the Roman Corinthian capitals. Passing through the porch we enter the square atrium or forecourt with a water stoup in the middle of it, and Roman pillars taken from some older building. BASILICA OF TEBESSA. l''i,re Court and steps : and Atnnv' and font. 33 Pla7i (M. Duprat) To face page Ij8 118 EASTERN ALGERIA ^^ < vttuued roof in the ui olongs to the trefoil chiw nji up the steps into the church, the visitor will ;e posts with grooved aides shown in the fore- ':^.*^^^^ ^f9 ^TS^^Ji^^^^^ peculiarities in I of the church. narij the border or the court and a broad paved I led from the entoiB^e gate along the They were joined by panels fitted iui.- and so formed a continuous balustrade. The Sii le altar screen will be seen inside the basilica, h African fircone or the choir be oor oi ijii^. level. This h of tombs nvolved an e of rn''^" " fcy i n churches, ding, oddly level. raph, and is ws that the brecourt are lat fi^^ '^'^'Tt fourteen hts. The u stones at the top marks the foundation of a porch ilanked b^fcw*^\C\?wWKd) thfi^formed the west facade. this now remains but the bases of a few pillars jii; ' -^ Roman Corinthian capitals. 1 <.w., M'h the porch we enter the square atrium or forecourt w >ater stoup in the middle of it, and Roman pillars taken from ttoTne older building. ?)\\ ^-^vi^ ^:ivi\ o\ TEBESSA 119 The stoup stood in the centre or open part of the court, and was covered by a canopy resting on four pillars ; these have dis- appeared, but their bases still remain in the bevelled recesses at the four outer corners of the stoup. The basin inside was scooped out into four lobes, giving it a cross shape. These atria are common enough in Christian churches, and there is nothing unusual or remarkable in this one, but it has been suggested that as the adjoining baptistery appears to be of much later date, possibly not older than the Byzantine occupation, this stoup may have been used as a font for much the same cere- mony as the Coptic Epiphany tanks were originally built to accommodate. A door leads from the atrium into the baptistery, and it was made after the wall was built showing that the baptistery is an addition. Three steps lead down from the floor of the atrium into the baptistery, so that the latter is considerably higher than the ground level, and consequently of the floor of the trefoil chapel. The baptistery is divided into two parts, an outer hall or narthex, and the baptistery proper with a round font sunk in the middle of the floor. The latter was covered by a canopy supported on pillars resting on stones taken from a classical building. The font itself consists of a shallow circular pit, built in three stages of equal depth and decreasing diameter, so as to form steps from the floor into the cuvette. The same pattern will be found in the font of a small church at Timgad, and in the great basilica at Timgad an octagonal font is built much in the same way ; but the pattern was not very common. From the atrium three doors lead into the church, the main door into the nave and the side doors from the peristyle of the court into the aisles. The next picture (my father's) is taken looking east from the main door along the nave and into the apse. The church is divided as usual into three parts, the nave, the two side aisles, and the sanctuary represented by a semi- circular apse and a prothesis and diaconicon on the north and south sides. The masonry is composed uniformly of large well cut blocks of local stone. The interesting point to notice in it are the numerous and varied mason's marks ; they occur chiefly in the church, but some will also be found in the trefoil chapel and 120 EASTEEN ALGEKIA in the avenue at the west front. I have not come across similar marks in Tunis or even at Haidra near by, but probably they exist, and they are certainly well worth careful study. Some of them, at any rate, appear to be Berber characters.* Numerous capitals of the Roman Corinthian variety are lying about, and on the north wall there are several carved brackets that appear to have supported the gallery in the aisles ; among them is a pretty design with dolphins that I did not see elsewhere in this district or in Tunis. The floor of the nave was covered with mosaic, and the eastern part of it was shut off by a screen to enclose the altar. The view shows a piece of the apse, the steps leading up to it, and the door of the chapel of the prothesis. The French engineers have restored a piece of the north colonnade of the nave, including the bases of the upper row of pillars in front of the gallery or triforium. The style is purely Eoman. In the foreground is one of the grooved posts to hold the screen I have already referred to. These screens were probably made of stone or marble open work,f and fitted into the grooves of the posts and also into the bases of the pillars between the nave and aisles. There is also some indication that part of the south aisle was shut off from the nave in the same way. The altar stood in the centre of the nave halfway between the little post and the apse, but only the foundations can now be seen. According to the usual practice of North African church architects the semi-circular apse and adjacent chambers or chapels were all enclosed in a square wall, so that the apse does not project and is not visible from the exterior. The occurrence of side chapels to the apse, the predecessors of the Byzantine prothesis and diaconicon, came as a surprise to me in a building of this early date. These side chapels are frequently found in North Africa. Sometimes there is one, and sometimes two ; some communicate directly with the central apse, and some do not. No general deductions can be made from these * Publications of the Archaeological Society of Constantine, vol. ix. 3rd. series (vol. XXX. of the collection), 1897 ; and vol. i. 3rd series (vol. xxii. of the collection), 1888. See also my drawings in the preceding volume on the fiysheet to Plate 56. t As in a panel in the Alexandria Museum. 34 BASILICA OF TEBESSA. Nave^ lookins^ tnto the chancel To face page 120 ^f. 120 KASTEKN MiGEKIA in the avenuf ai lik; wcm iiuui. I have not cuaic at marks in Tunis or even at Haidra near by, but pi' (ixist, and they »re certainly well worth careful study. Some of them, at ar-^y rate, appear to be Berber characters.* Nimierous capitn' ' 'e Roman Corinthian variety are lying about, and on tli' wall there are several carved brackets that appear to hav' supported the gallery in the aisles ; among them is a -ign with dolphins that I did not see elsewhere in this ! or in Tunis, e floor of the nave was covered with mosaic, and the va part of it was shut off by a screen to enclose the altar. The view shows a piece of the apse, the steps leading up to it, and the door of the chapel of the pro thesis. The French engineers have restored a piece of the north colonnade of the luve, including the bases of the upper row of pillars in front v«f the gallery or triforium. The style is purely Romai In the foreground is one of the grooved posts to liwm u.^ wjieen I have already referred to. These screens were probably made of stone or marble open work,! and fitted into the grooves oi the posts and also into the bases of the pillars between the nave and aisles. -^"^^^^"^aT^ sbi^^"^^lfc§tion that part of th*^ south aisle wd»:>«is&it ^M elroii^utlNs>\n,a'VQVin the same way. The altar stood in the centre of the nave halfway between the post and the apse, but only the foundations can now be ■^< '11. According to the usual practice of North African church architects the semi-circular apse and adjacent chambers or ipels were all enclosed in a square wall, so that the apse does ' project and is not visible from the exterior. The occurrence -le chapels to the apse, the predecessors of the Byzantine i< sis and diaconicon, came as a surprise to me in a building s early date. These side chapels are ly found in '^^■ica. Sometimes there is one, auv. -. . umes two; itmicate directly with the central apse, and some d No general deductions can be made from these .f the Archaeological Society of Constantine, vol. ix. 3rd series (v<.; : ■ \. 1897 ; and vol. i. 3rd series (vol. xxii. of the collection), 1 >5.HX. igs in the preceding volume on the flysheet to Plato 56. . a Alexandria Museum. 0^\ ^■^v»\ ^:>»\ o\ TEBESSA 121 peculiarities as to the service or purpose they were intended to accommodate. The substantial and important fact from the ritual point of view is that they should be found in this early building. It does not necessarily follow that they were intended or used for any ritual purpose, or for anything more than practical convenience ; but it seems reasonable to suppose that they are the predecessors of the side chapels that became essential for the due performance of a part of the later Byzantine Communion Service, and I hazard the conjecture that that part was founded upon some early and common practice in these African churches. The floor of the apse is raised about half of a metre above the floor of the nave, while the floors of the chapels are level with it. This elevation was negotiated by four flights of two steps, one flight in each chapel and one flight on each side of the apse. An arrangement of steps like this is not uncommon, and several examples will be found, as for instance, at Dougga and in the so-called basilica of Justinian at Carthage, to name two of many examples. The student of Church architecture will be reminded of the same sort of arrangement in the great screens in S. Paul's and at Winchester Cathedral, and the purpose they were intended to serve at the consecration of a bishop according to the Sarum use. In the middle of the apse floor, and against the wall, are the foundations of what was probably the bishop's throne, but this and the seats for the clergy have disappeared. The arrange- ment of seats for the clergy round the wall of the apse was common in these provinces, in Egypt, and in Nubia. It is also found in Eome, at Torcello, and in S. Eirene at Constantinople. It has often occurred to me to wonder whether the open space between the altar and the steps leading up to the apse was screened off by an open screen like that in S. Maria in Cosmedin at Kome, or in S. Mark's at Venice. The remains of the Tunis churches I have seen afford no evidence upon this point. To reach the trefoil chapel we have to return into the nave, and then into the south aisle of the church. A wide opening in the south wall opposite to the second and third bays from the west end of the nave leads to a broad flight of twelve steps. 122 EASTERN ALGERIA By these steps we descend three metres from the floor level of the basilica to that of the trefoil chapel. The reader should now refer to the illustrations and plans on Plates 54, 55, and 56 of the preceding volume. The pictures of the chapel at El Gebioui and the similar chapel at Maatria show how the central square chamber was roofed by a cross vault and the apses with semi-domes ; these were no doubt made of brick in the usual way, and traces of the mosaic they were covered with were found by M. Clarinval. The two cippolino pillars at the entrance of each of the apses were intended to support the arch of the apse in the same way that those at Haidra supported the chancel arch. The masonry of the chapel throughout is of large blocks of local stone, carefully cut and fitted in the same long and short work as those in the basilica itself. The floor deserves more careful notice, and here I must refer to the reports of MM. Seriziat and Clarinval, by whom it was excavated. The former, describing his exploration in 1868, says * that the work of clearing away the rubbish from the trefoil was not sufflciently far advanced for him to report more than the discovery of a large sarcophagus that he had transported to Tebessa. Two years later M. Commandant Clarinval, describing his work, says : t * Two years ago M. Seriziat commenced to clear away the rubbish that filled the trefoil, and found near the staircase a large white marble sarcophagus.' I need not describe this object. It is a Christian work of early date, and roughly executed. M. Clarinval employed a soldier artist to draw it, and it is reproduced in the proceedings of the Constantine Archaeological Society with his notes. ' In continuing the clearing we uncovered a staircase of twelve steps and the mutilated remains of the sarcophagus hidden in the debris. We found the floor was decorated with mosaics that have disappeared except in the east apse. Eight columns of green marble stood in pairs at the entrance of each apse, and * Publication of the Archaeological Society of Constantine, vol. ii. 2nd series (vol. xii. of the collection), 1868. t Id., vol. iv. 2nd series (vol. xiv. of the collection), 1870. TEBESSA 128 the walls were veneered with marble. Towards the bottom of the debris were layers of charred wood mixed up with tiles and mosaics from the roof, showing that the building had been destroyed by fire.' Then follows his description of the discovery of the boy's tomb that I have already described and illustrated in the preceding volume. ' In the course of excavating this tomb I discovered an ancient floor covered with mosaic at 1 m. 20 c. below the present floor level. In the middle of the chapel two large slabs seemed to indicate the presence of an altar, and as the ground sounded hollow I had an excavation made at this spot. A square hole was found beneath filled with rubbish, and at 1 m. 20 c. the ancient floor covered with mosaic was again found ; immediately below it in some cinder rubbish was a tombal stone covering two funeral urns, a lamp, and a little pot. Two sides of the hole were made of blocks similar to those used in building the basilica, while the side (on the plan) is made of small stones. . . . The excavations at the foot of the stair- case showed that the steps rested on no foundations, and again brought to light the ancient mosaic 1 m. 20 c. below the present level.' The rest of the report relates to the excavations of the adjoining chambers, and these are referred to presently. M. Duprat's account * of the trefoil chapel and the digging is taken from these reports, and includes the discovery of the mosaic and jewelled panel. This was first taken to the modern French church at Tebessa, subsequently moved to the museum, and eventually stolen from there. M. Gsell's account supplies some further particulars. After recapitulating what has been said above he proceeds : * The floor of the central apse was covered with mosaic representing intertwined vine branches issuing from chalices ; the mosaic floor of the left hand apse had round and lozenge patterns and the ' gamma ' cross ; the right hand apse had a design of birds with a stag in the centre. The roof was covered with mosaic cubes made of glass that were found in the rubbish. Four little pillars carved with vines, fish, &c., that probably belonged to an altar, were also found near by.' * Id., vol. ix. 3rd series (vol. xxx. of the collection), 1897. 124 EASTEKN ALGEEIA In the right and left apses, doors led to the adjoining chambers in the four angles of the square. They, too, had mosaic floors. Those on the south side had massive barrel vaults, one of which can be seen in the pictures ; those on either side of the stairs had an upper story supported on a cross vault, of which fragments made of a soft stone are still visible in the angles of the walls. There is a conflict of evidence concerning the construction of the two last-mentioned chambers. M. Gsell says that the outer walls were detached from the wall of the basilica, whereas those next the stairs were built into it, and he argues from this that the chapel and the basilica are of the same date. I hesitate to differ from him on a question of fact, but my observation agrees with that of M. Clarinval, and to me it seemed that all these walls were applied and not built into the wall of the basilica. There remain the rectangular chambers at the west side of the trefoil chapel, and the access to these was by a door cut in the centre of the western apse. The first and smaller chamber was apparently only a vestibule leading to a larger and more important square hall beyond. M. Duprat identifies this as the treasury of the monastery. Whether that is so or not, the excavations made in 1869 and 1870 by M. Seriziat show that it was used for burial. Four mosaic tombs were found ; they were of Bishop Palladius, probably the prelate of that name who suffered in the reign of Hunneric the Vandal king, Petro- nius, Quodvultdeus, and Marcella, a little girl three years old. The remains of the bishop were found resting on a bed of laurel leaves, and also those of the little girl, whose hair was fair. As in the trefoil chapel itself, so here, the original lioman floor was found at 1 m. '20 c. below the surface of these mosaic tombs. The little basilican church on the south side of the trefoil chapel is a building of much later date, and probably belongs to the period of the Byzantine occupation when the great church had been destroyed. It is made of old materials, and the altar stood out in the middle of the church. According to M. Duprat both the basilica and the trefoil chapel were secular Boman buildings, the former a court of TEBESSA 125 justice, and the latter a bath. There is a good deal to be said in favour of this opinion so far as the trefoil is concerned. M. Seriziat's excavation shows that there was an original building on the present site, with a mosaic floor in the common Roman style of decoration, in which, so far as I know, nothing distinctively Christian was found ; it may also be said that the eight cippolino columns were imported by the Romans before the Christian period, and they were apparently made to fit this building. I have also pointed out in the introduction that this trefoil plan was certainly in common use for the laconica of public baths.* But, on the other hand, it must be said that the adaptation by the Christians of a plan suitable for a bath to a baptistery would be reasonable and probable enough, and it is also certain that from very early times this trefoil plan was used for the sanctuary of a church or the cella memoria or funerary chapel, f M. Ballu is of opinion that the trefoil was a funerary chapel, but the tombs found there, which, as I have shown, are of later date, are the least strong evidence in support of it, for, as the floor level shows, they are certainly more recent than the rest of the fabric. The conclusion that I have come to is that this trefoil plan was used as well for a church as for a tombal chapel or a bap- tistery, and that the Christians derived it from some building at Jerusalem then existing but now destroyed, that was traditionally associated with our Lord's ministry, much in the same way that in building the round churches | the Crusaders conceived they were copying the church of the Holy Sepulchre. With the exception of the so-called stables, the other buildings in the enclosure of the monastery, though of interest in them- selves, and as associated with this site, afford no useful material for comparison with buildings elsewhere. The stables are one of the archaeological puzzles of Algeria, and as another similar building exists at Haidra I have described the two together in a short account of the principal Christian remains of that city. * As at Lambessa and Thelepta. t As in the churches at Sohag and in the cemetery of S. Callixtus at Rome. I I refer to the London Temple church, Northampton, Cambridge, and Little Maplestead in Essex. 126 EASTERN ALGERIA There are many other archaeological remains of great interest in and about Tebessa itself ; and, foremost among them, the walls of the modern town. These were built by Justinian's generals, and are still almost as perfect as the day they were erected. The little museum was a Roman temple also in very fair preservation, and it contains a number of classical and Byzantine fragments collected in the neighbourhood.* TIMGAD The little church illustrated on the plate opposite is situated in the northern quarter of the city and, as the foundations extend across one of the Roman streets, it was presumably built either during the Vandal occupation or after the Byzantine conquest. The church lies east and west and is of the usual simple plan ; there is first a little porch at the west end leading to an atrium or forecourt ; from this atrium three doors lead into the church beyond. The plan is basilican, with nave, side aisles, a semi- circular central apse at the east end, and a chamber on each side of it corresponding to the prothesis and diaconicon ; as usual in African churches these were contained in a square outer wall. The outer walls are made of large upright stones, probably taken from a Roman building. They are placed at intervals of eight to ten feet, the intervening space being filled with rubble and smaller stones. A square court on the north of the atrium contains the baptistery. The pillars in the nave have monolith shafts of stone and stone capitals carved in the usual early Roman Corinthian pattern ; these also probably came from a Roman building. The two points of interest in this church are, first, the double wall of the apse ; this either represents an ambulatory, a place for burial, or perhaps an arrangement for the seats of the clergy, similar to that seen in the apse at Announa ; and secondly, the circular font built almost exactly like the example in the basilica at Tebessa. The baptistery was a covered cloister of the same kind as that adjoining the so-called basilica of Justinian at Carthage and the basiHca at Sbeitla- M. Gsell describes and gives a plan of this church, but apparently it had not been com- pletely excavated when he saw it. * A tombstone from here is illustrated with others from Guelma on p. 130. ANNOUNA 127 The great basilica of Timgad stood outside the west wall of the city, but it has been completely razed to the ground, and all that remains of it are the floor, small pieces of the walls, and a few shafts and capitals of pillars. At the north end stood the baptistery, now completely gutted, but containing the very beau- tiful mosaic font that has fortunately been preserved. The only two architectural details in this basilica that need mention are, first, the absence of chapels of the prothesis and diaconicon on each side of the main apse ; secondly, the rectangular chapel terminating in a semicircular apse placed at right angles to the nave on the north-east side corresponding to the cella trichora at Tebessa. The same kind of large chapel, for that is what it no doubt was, also occurs in the so-called basilica of Justinian at Carthage and in the other large basilica on the north side of Timgad. There is no indication in any of these buildings of the liturgical purpose this chapel was intended to serve, but it probably had some importance,* for the same kind of apsed chamber at right angles to a nave will be found m the majority of Tunisian churches and in some of the Christian tombs in the cemetery of Khargeh in Egypt. There are one or two other churches in or near Timgad, and a reputed chapel in the Byzantine fort on the south side of the city, but they present no features of particular interest for purposes of comparison with churches elsewhere. ANNOUNA The city of Announa, the ancient Thibilis, is situated in the hills about six miles south-west of the famous hot springs of Hammam Meskoutine. It was close to the high road leading from Bone to Lambessa and other cities on the Algerian plateau, and from the extent of the remains was obviously a place of considerable importance both in the Roman and Byzantine periods. The basilica is one of the few churches in Algeria and Tunis of which a considerable part remains above ground. It is situated at the end of a main street, in a prominent position at the foot of * Les Ruines de Timgad, A. Ballu, 1911, p. 38. I had"not seen thi.s book when these notes were written. 128 EASTEKN ALGEEIA a little hill that may have been occupied by a pagan temple. The old materials taken from a Roman building and the form of a cross over the main door + P combined with A and o depending from the arms show that it was built by the Byzantines. There is some indication from the appearance of the debris, and shafts of pillars and caps that are lying about, that the city was destroyed in an earthquake. The fa9ade, as the picture shows, is almost perfect. It was pierced by the main door and a square window on each side. A single stone over the door was relieved by a discharging arch above. On the latter is a large Greek cross with A and u depend- ing from the arms. The general view is taken from above the apse, seen in the foreground, and the nave and the inside of tlie fa9ade are in the distance beyond. The side and front views of the apse are taken from the nave. The latter shows a block of stone in the centre that formed the foundation of the bishop's throne, and the bases of two detached pillars to support the apse arch. The arrangement of the seats for the clergy, built in tiers in the apse, may explain the peculiar construction of the apses seen in the small church at Timgad, in the large basilica at Sbeitla, and at Dougga. The front seats are ranged in a quarter circle, while those behind are in a semicircle. This is one of the few African churches where there are no side chapels for prothesis and diaconicon, and the aisles terminate in square walls abutting on to the hill side. These pictures also show the style of building walls with large upright stones and rubble of smaller stones, so commonly found in all the Byzantine buildings in these countries. The inside walls were covered with cement, and either painted or, in the richer churches, veneered with marble. Passing now to the nave, the font and several capitals of white marble of the usual late Roman and Corinthian variety are seen lying about ; the carving on some of the pillars is unfinished. On the pilasters in the east wall are two square caps in the same style that appear to have been taken from some Roman building. The general view of the interior taken from above the apse shows the common division of the nave. It was divided from the aisles by pillars on each side supporting a timber and tile roof. The aisles and the spaces between the pillars were paved with flagstones and GUELMA 129 the nave proper with mosaic. This mosaic was usually arranged in two patterns ; a more elaborate design being reserved for the sanctuary and a simpler one for the gangways and adjacent spaces. In this church a line of stones across the first bay of the nave marks the break in the pavement, and the altar, of which there is no trace, stood just beyond it, leaving, as usual, a consider- able space between the clergy seats in the apse and the altar. There is no indication that this space was used for any particular purpose, but it formed an open gangway and permitted of a perambulation of the altar and a passage from one aisle to the other. This arrangement is usual in African churches. The ruins of this city are very interesting for, like those at Timgad, they show just how the Roman town was laid out and how the Byzantines rebuilt the city and the houses over the Koman streets. Among the earlier ruins are the remains of a triumphal arch and some large edifice, probably the forum ; there is also the ruin of a little chapel with a single apse. GUELMA The site of the Roman city at Guelma is now occupied by a busy and thriving French colonial market town ; consequently nearly all the ancient buildings have vanished. There are, how- ever, a fine theatre and the thermae ; the latter are now enclosed in the French barrack square. Some tombstones preserved in the little public garden of the French quarter of the town deserve more than a passing notice. The first was made presumably for a pagan, as the following legend is in the usual Roman form, and there is no cross or sacred monogram about it. D.M.S, I JVLIVS . FAVS | TVS . PIVS . VISIT | ANIS . LXV . SE I D . VIXSIT. BEN | E | H.E.S. This legend is cut in poor shaped and irregular characters sur- rounded by a single circle line ; the small diagonal cuts on this circle show that it is intended to represent a wreath of foliage. Two saucer-shaped depressions on the dexter side and a third on the sinister represent the utensils or receptacles to contain food 130 EASTEKN ALGEEIA provided for the use of the departed in the nether world ; * the object on the sinister side is apparently intended to represent a kind of saucepan with a handle to it. The rest of the surface of the stone is incised with rude representations of leaves and flowers, and perhaps two cornucopia. The stone above it, of which only the lower half is visible in the photograph, is also a pagan tombstone ; in the lower corners are rude carvings of drinking cups. 2. The inscription on the next stone is unfortunately too much damaged to be legible in my photograph ; it begins with the p AM, showing that the person interred was a Christian, and the same kind of saucer incisions occur in two of the corners. On the other sides are a sprig of foliage and a jug ; the small sketch book at the side shows the size of the slab. 3. The next inscriptions are on stones that have now been placed one above the other. On the upper stone the legend is D.M.S. ^ JVLMECCENT | IN PACE FIDELIS | VIXIT ANNIS LXXII | HSE. This inscription is contained in a ring or conventional wreath and in the ' field ' are a number of small rubrics inclu- ding a star cross in a circle ; the same kind of saucer-like depressions are cut in the surface of the stone at the sides. 4. The inscription on the stone below reads : — J^^ flavivs. AVEN 1 TIVS 1 FIDELIS IN PACE | VIXIT ANNIS | . . . HSE. This legend is in a plain incised ring, and at the side of it are the usual saucer-like depressions. The effigies of utensils on the tombstones suggest to me that the Christians in the earliest times, and whilst the old faith sur- vived, purchased the ready-made slabs stocked for pagan use. The case bears, perhaps, some analogy to that of the silver * pot- house ' tankards so frequently found in collections of Church Communion plate. These ready-made tankards were purchased by the parishes after the Keformation in compliance with the injunctions of King James I ordering the churchwardens to procure ' stoups ' of metal to hold the communion wine. The parishes then went to the ' trade ' and purchased the most suitable ready-made article with the whistle in the tail of the handle. This whistle was used to call the pot-boy with when the flagon was empty and needed refilling. • Of. a slab in the narthex at Sta Sabina in Rome. 35 Early christian tombstones at Gnehna (Nos. 2 and 4 on p. I JO) and at Tebessa mnseum^ p. IJI, EASTEEN ALGEBIA ^^ provided for the use of the departed in the nether world ; * the object on the sinister side is apparently intended to represent a kind of saucepan with a handle to it. The rest of the surface of the stone is incised with rude representations of leaves and flowers, and perhaps two cornucopia. The stone above it, of which only the lower half is visible in the photograph, is also a pagan tombstone; in the lower corners are rude carvings of drinking cups. 2. Thr^ inscription on the next stone is unfortunately too much damaged to be legible in my photogi-aph ; it begins with the p All, showing that the person interred was a Christian, and the samf kind of saucer incisions occur in two of the corners. On ♦ u, .fher sides are a sprig of foliage and a jug ; the small sketch at the side shows the size of the slab. :l The next inscriptions are on stones that have now been 1 one above the other. On the upper stone the legend is P JVLMECCENT | IN PACE FIDELIS | VIXIT ANNIS LXXII | This inscription is contained in a ring or conventional vvroatb and in the * field ' ai:e a number of small rubrics inclu- ;i star %f^!^8'^A»f>^icitt4l«(^ tk^ie«toi6^\-4{jad of saucer-like .. ;. tensions are cu^^cJJX ^ei^<§»V'^*<<^ §f.^>^*one at the sides. 4. The inscri^^n^o^,|ii^^t^§^^^b^l(^w ^ca^ds :- Ji^ flavivs. VKN 1 TIVS I FIDELIS IN PACE | VIXIT ANNIS | . . . HSE. ■end is in a plain incised ring, and at the side of it are . ' saucer-like depressions. thgies of utensils on the tombstones suggest to me that ' tiristians in the earliest times, and whilst the old faith sur- ' purchased the ready-made slabs stocked for pagan use. ' bears, perhaps, some analogy to that of the silver *pot- 1 mkards so frequently found in collections of Church Comii! union plate. These ready-made tankards were purchased by t^ lies after the Reformation in r with the injuiK... >)f King James I ordering the dens to procure sioups' of metal to hold the communion wine. The parishes Mi-n went to the 'trade' and purchased the most suitable ! " article with the whistle in the tail of the handle. ^. . . . lie was used to call the pot-boy with when the flagon was empty and needed refilling. • Cf. a slab in the narthex at Sta Sabina in Rome. 0^\ -i^v^^^ <^:iv%A oT GUELMA 131 The next inscription is on a tombstone preserved in the Museum at Tebessa. The legends are: in the field, dexter side, SCI I Bi I N I CE I N I Ti I MAR | TiRis ; partly obHterated above i^, ori | a | . On the top and sinister side + arcan- GELV I S MIKA | EL ET [ GAB | RIE [ L \ SCE | .CRI | SPI | NE | MAR I TiRi I s. In the ring round the „^„, + posita ad mo PATRE FAVSTINO EPISCOPO VRBIS TEBESTINAE SUB DIE | V ID vs AB INC CNT XIII ; on the base of the tablet memoria SANCT^ MAXIM | ATILL^ ET SEGVND^ WESTERN TUNIS HA IDEA The Eoman name of this great city was Ammadsera, and the ruins, that owe their wonderful preservation to the isolated position, are among the best preserved and the most important in Tunis. I say Tunis, but in point of fact, the custom house for the frontier between Tunis and Algeria actually stands in the centre of the city. The ruins are well worth visiting, but unfortunately, as difficult of access by the phosphate railway from Tunis to Kalaa Djerda, as by an apology for a road from Tebessa. The Eoman city was first ruined either by the Vandals or more probably by the native Berbers, and when the Byzantines came upon the scene they made use of the old materials to build a large fort enclosed in a curtain wall, and a number of churches. By far the most important of the Byzantine remains is this fort which commanded the high road from Carthage to Tebessa. Beside this military structure the great gate or triumphal arch of the Boman city, erected in honour of the Emperor Septimius Severus, was walled up with Roman masonry and converted into a block house. The three buildings described in these notes are the Byzantine chapel in the fort, a large basilican church of the early Christian period lying to the north-west of the fort, and a group of buildings said to be a monastery that contains a large hall described by some antiquaries as a church, and by others as a stable. A similar edifice will be found in the enclosure of the great basilica at Tebessa. I saw three other small churches, one close to the south-west angle of the fort and the other two near the block house I have just mentioned. These little churches HAIDEA 133 were plain basilicas with single apses, and the best preserved * had a porch in front of the main west door and was built up on a plinth in which I noticed some hexagonal Roman tombstones or funeral stelae. The Church The great basilican church situated about two hundred yards north-west of the fort was a large building consisting of a nave separated from lateral aisles by a row of granite pillars about 4 met : 16 c. long with stone caps in the Corinthian style. The sanctuary at the west end has a semi-circular apse and at least one adjacent chamber. At the east end there was a narthex and a fa9ade flanked by two towers. As usual, the sanctuary is the interesting part of this church. The walls of the semi-circular apse and of the prothesis marked A on the plan, still stand about two metres above ground, and are made of well-cut stones taken from an older building. The diaconicon on the north side has disappeared, if indeed it ever existed. It is probable that the sanctuary of the church was enclosed in a square outer wall as usual in Africa and that there was an ambulatory between the outside wall, of which the base only remains, and the apse itself. This arrangement is as unusual in Africa as that of an apse projecting in a semi- circle beyond the outside wall. The picture shows the apse taken from the south-west corner, standing on the plinth or base of the square outer wall that enclosed it. The space between the corner in the foreground and the apse was occupied by the chapel of the prothesis ; this little apartment communicated with the sanctuary by a door (on the left) seen in the break of the masonry, and by another door (on the right) into the south aisle. The bolt catch on the pilaster of the door leading to the south aisle can be seen in the picture. Chapel in the Fort The garrison chapel at Haidra abuts on the middle of the west wall of the Byzantine fort. This great enclosure of about ten acres is surrounded by a high wall with bastions at the * A plan will be found on the fly-sheet. 134 WESTERN TUNIS corners and at intervals along the sides. The chapel stands at right angles to the wall, and the chancel is built up against one of the bastions. It seems to have been a little basilica with nave and aisles and a single apse. The site has not been excavated and the debris lie where they fell. The base of a tower at the angle, a portion of a pilaster, and the spring of one of the nave arches all on the north side, alone remain. The lower part of the apse, built against the bastion, has been preserved up to the spring of the semi-dome that covered it. The chancel arch was supported by two cippolino columns, one on either side. A few stones of the arch remain resting on the abacus of the pillar. The abacus, if I may for convenience use the term, was decorated, as the photograph shows, with conventional sunflowers or perhaps daisies. The corresponding pillar, also of cippolino, has fallen out of its place and is standing inclined at an angle half buried in the ground. The apse, full of the debris of the semi-dome and of the chancel arch, built and decorated on the same principle aS El Kef, is the interesting feature in this church. On a broken corbel stone lying in the apse I noticed portions of a Greek epitaph. The views of this part of the church will be found illustrated in the preceding volume on Plate 67. The Monastery I now come to the so-called monastery and the central edifice in it should be compared with the interior of a similar building in the precincts of the basilica of Tebessa. The latter is much the larger building of the two and the construction is not quite the same, but the distinctive peculiarity, the arrangement of troughs and arched apertures over them built into the walls dividing the nave from the aisles, occurs in both. The picture of this edifice at Haidra is taken from the point A on the plan, and the spectator is standing in the nave or hall looking in an easterly direction towards the wall dividing it from the south aisle at the point B. The reverse side of this wall is shown in the next picture. An examination of the masonry shows that the front or finished side of the wall faced the nave and was made of stones carefully dressed and laid; !;f« 30 HAIDRA. So-called stables with mangers I' O n d 7 TP n To face page 134 de comers and right angles of thp ' " *^'' and a; and ti angle, 'a t^ arches a) i apse, ii ■''■ sprin^f of WESTERN TUNIS- at intervals along the sides. The chapel stands at to the wall, and the chancel is built up against one ♦^s. It seems to have been a little basilica with nave a single apse. The site has not been excavated ; lie where they fell. The base of a tower at the ter, ab5V*feH^a-^ved up to the dome that covered it. The chancel arch was i'')lino columns, one on either side.— A te^ main g&a jjiug - oj a^the abadus of the pillar. term, was decorated: ^fl^wers or perhaps o, has fallen oui J inclined at ai of the debris of the se ft", ti id decorated on the same ptiJQ ^ 4ii'ture in this church. On ipse I noticed portions lit of tlie church will b I'lliiie on Plate 67. iple as El h broken cor 3( 1 Greek epita^l ,1 ii d illustratec i|ti e so-called monastery and ,i|e central edifice Ired with the interior dO ^ {sim ilar bulk i ig le basilica of Tebessa. Urn? lot te d is mu ;h g of the two aM the construction is not c©t.e 'ctator is st- p— distinctive peculiarity, the arrangemen 'Tf-* apertures over them 1 ni^ ilk^ fl'f' ^Hf m the aisles, occurs in b jtii Jib edifice at Haidra is t vljeu iroii tl: 10 !rom the south m»' vail is s^ masonry sau-Ab ; th»^ nave and w --]■■ ''"Mpfirm f.n^ ^ nt Ijall it ftt the point B. The reverse side of this M'Xt picture. An examination of the nt or finished side of the wall faced tones carefully dressed and laid; V^A n»^ ^tift\ o\ HAIDKA 135 on the top of this wall are a row of eight round-headed arches supported on little monolith pilasters, and above the arches there is a course of stones laid along the top. The under-side of these stones has evidently been cut away to relieve the pressure on the arch below. It will be understood that the wall on the opposite side was built in exactly the same way. At the points a a in the angles of the nave, are the bases of small pillars that were presumably introduced for decorative purposes. These pillars have disappeared and any part of the roof they may have supported is also gone. Of the roof nothing remains, but at each end of the hall is an arch of substantial masonry that probably supported a timber and tile roof. It is conceivable, but very unlikely, that the nave was covered through- out with a stone barrel vault. The south end of the nave was finished off by a semi-circular apse made of small stones and covered by a semi-circular semi-dome that abutted on to the stone arch I have just described. The floor of the apse was slightly raised above that of the nave and marked off by a row of flag stones. The aisles or gangways on each side of the nave led respectively to a small chamber on each side of the apse at the east end. The arrangement of the building at the west end is more confused, but I think the plan represents it fairly accurately. The entrances b and c led to a gangway d, and from it there was access to several apartments at the west end of the nave. On the south side there seems to have been a large courtyard. The interesting features in this edifice are the little rectangular troughs cut in the top of the wall between each pilaster to a depth of about 45 centimetres. A comparison of this building at Haidra with that at Tebessa, shows that the situation of the troughs was the same in both, but at Tebessa in place of the arches the apertures over them were square. The latter is a much larger building than the one at Haidra, and needed square piers placed at intervals in the nave to support the very wide span of the roof. Both at Haidra and Tebessa some of the troughs and the pilasters adjoining them are pierced with small holes of a kind frequently found cut in Roman ruins in North Africa, as 136 WESTERN TUNIS for example, in the steps of the principal temple of Dougga ; and the Arabs of to-day cut much the same kind of holes in walls to tether their cattle by. It has been argued from the occurrence of these holes that these troughs were for horses or cattle, and that the buildings were either stables or markets. On the other hand, it is argued that the form of the building at Haidra, especially the apse and adjoining chambers, show that it was a church. I may observe that it is difficult to believe that the building ait Tebessa was a church. First there is no evidence of it in the plan or details themselves, and secondly it is improbable that with the great basilica near by a second church of this odd shape should have been built in the same enclosure. The objection to applying the stable theory to the building at Haidra is a practical one. I calculated that the aisles were too narrow for horses, that eight aside could be accommodated in the nave, making sixteen animals altogether, and that they would be obliged to remain in a standing position. This practical difficulty and the improbability that so large and costly a building should have been erected for so small a number of animals seems to dispose of the stable theory in the case of Haidra unless we are to suppose, as M. Saladin does, that it was a fonduk or khan for temporary accommodation only. Having said that much, however, I do not doubt that the Arabs used the hall at Tebessa to stable their horses, or that the early Christians converted the nave at Haidra into a church, but these were not the original purposes either building was intended to serve, and whatever that purpose may have been, it was the same for both. It is easy to dispose of these conjectures and difficult to suggest even a plausible theory, let alone a satisfactory explan- ation concerning these troughs. As an alternative to M. Saladin 's proposal, I suggest that they were intended for the treatment of some mineral or agricultural product, excluding oil and wine, for we know what the presses commonly used for these commodities were like. Not the least curious thing about these troughs is that they occur so rarely,* and that the only known examples should be situated so close together. * There is some indication of the same kind of building at Sbeitla. D r" 37 3 |b— "^ j^::^ ^iear fe^lLcn. fn. j N 7b /wcg p(ig^ 13^ Wiii^irji^iN ^t Tor example, in the steps of the principal temple of Doagga; at ' ■ ' ■; of to-day cut much the same kind of holes in w r their cattle by. \t hoh been ai^ued from the occurrence of these holes that the.He were for horses or cattle, and that the buildings u, ' - '■-♦^.o.rQOrOtJje other hand, it is argued rl ^ at^Haidra, especially the apse and a\\ aside jcould be accommodated in the nave, Id be fficulty □ jiimals al |P^0^ > ^^^ ^1^^^ a stand!!gg=»SB«ion. This piltfe I . . • :v' that so^larg4 aod^oBtl^ a htiilfli^ig| should for so ^ma H a number of animalssaems to ;. theorf {(j)j the case of Haidr8J(unie8S we M. Sal^adin does, tlflW-ijt was a foiq[duk or i-y accom|pofV,tion onl|e=il rT'^l tat much'tJaflwever, I do not doi |UlJ| jbat tht till at Tqbessa to stable their W)rses, 'or that he aavp ,at H es ei hJBse. conjectures 1 06 w/ iit — rroug It they wferjs itural pr> ij .i'A comi]|i dnlv used for i... .. tfc.t curi j\ £^one a s Kaiira^terna inreQ^edZTo uct, exclu3iiv church, id . d ifl ] cult to toi y Bxpla n- ari maa thing ftbuut thoaaitraugl i 18 l .Si and that the cWy known examples • 'i^'i^ther. ■•• sumt' kind of building at Sboitla. ?i'^l s.-^Si<\ ^:ivi\ oT DOUGGA 137 EL KEF The principal place of importance between Teboursouk and the Algerian frontier at Haidra is the town still known by this Arab name. The Roman city was called Sicca Veneria, and must have been a place of considerable military importance. It stands at an altitude of 700 metres above the sea-level, and on the occasion of my visit was covered with snow and bitterly cold. The church is described and illustrated in the preceding volume on Plate 66. DOUGGA This miserable collection of Arab hovels occupies the site of the once flourishing Roman town known as the Civitas Thug- gensis, first a municipium in the reign of Septimius Severus, and later on a colony known as Colonia Licinia Septimia Alexandriana Thugga. The chief object of interest at Dougga is the Punic tomb erected for a local prince and dating from the second century before our Lord. This celebrated structure and the Roman remains are too well known to need more than a passing reference. Owing to the isolated position in the hills, about 5 kilom. south-west of Teboursouk, the more important ruins have escaped destruction in a wonderful degree ; the theatre in particular is perhaps the best preserved monument of its kind in existence. At the entrance of the town, and a few yards north of the road to Tebbursouk, are the remains of a little church built, like the Byzantine fort near by, of Roman materials. It was discovered in 1907, and has recently been excavated and explored. The small letters on the accompanying ground plan indicate the places the pictures were taken from. It is a little basilica with a square nave, side aisles, a sanctuary with a single semi-circular apse at the north end, a narthex, a crypt under the chancel, and some tombs and a chapel on the south-west side. There were three doors in the south front leading to the nave and the aisles respectively. Entering by the main door, we pass into the narthex, and thence into the nave. This was separated from the aisles by ten pillars, five on each side, the first two nearest the door being in pairs 138 WESTERN TUNIS to mark the narthex. The Hme-stone shafts and caps obtained from a Roman building have nearly all been taken away b}^ the Arabs. The walls were made on the usual Byzantine principle, as in the fort at Ain Tounga, of large vertical stones, braced here and there with horizontal ones, and the interstices filled in with small cut stones or rough rubble, according to the care bestowed on the construction. I am not at all sure that the nave of this church was covered in by a roof at all, and should not be more correctly described as an open atrium or forecourt ; the aisles would then form a kind of cloister, and this arrangement would account for the square drain-hole in the centre of the floor being provided to carry off the rain-water. The entrance to the chancel is by two small flights of three steps. It will be seen from my sketch that part of the floor of the chancel has fallen into the crypt below. The sanctuary was arranged with seats for the clergy on the usual semi-circular plan set in the square end of the church. There is no trace of the altar that, according to precedent, should have stood in the nave at a little distance from the steps ; it may have stood on the piece of the floor that has now fallen in, or between the steps themselves. In that case there would have been a window under it with perhaps a grille in front to light the crypt underneath. The little square hole in the middle of the floor of the nave is too small to have been the base of the altar. Access to the crypt is by two flights of steps, one in each chancel aisle. The ground plan of the crypt shows that it was arranged to accommodate several tombs. Of these, the most important was at the east end apse, where I noticed the cover of a stone coffin inscribed victoria sancti moniale (?) IN PACE. In the south side of the crypt are a number of other stone coffins, all desecrated and rifled. Passing across the crypt, and ascending the stairs on the west side, there is a door in the wall of the church immediately opposite the top of the steps leading into two little chambers shown in the photograph. The first is certainly a tomb. The other has an apse, and probably had a little chapel in it. There is hardly anything left of the decorative detail of the church. I noticed one shaft of coloured pink African marble lying in the aisle, a large limestone capital of the Roman period 9S nOUGGA. Chancel and crypt. To face page ij8 1H8 W\bM.J>iv^ iuMr. ^ to mark the narthex. The lime-stone shafts and caps obtained from, a Koman building have nearly all been taken away by the Ai«bs. The walls were made on the usual Byzantine principle, as in the fort at Ain Tounga, of large vertical stones, braced here and there with horizontal ones, and the interstices filled in with smn" or rough rubble, according to the care bestowed on ti-- lion. I am not at all sure that the nave of this church wad covered in by a roof at all, and should not be more correctly described as an open atrium or forecourt ; the aisles \ ' ' ' ' I a kind of cloister, and this arrangement would square drain-hole in the centre of the floor being pro\ arry off the rain-water. The entrance to the chancel is by two small flights of three steps. It will be seen from my ' ' ' ' ^ part of the floor of the chancel has fallen into the _■ A'. ■sanctuary was arranged with seats for the clergy on the •mi-circular plan set in the square end of the church. trace of the altar that, according to precedent, should n the nave at a little distance from the steps ; it may on the piece of the floor that has now fallen in, I lie steps thems^^^^Q^that case there would have )W under it\<^ii ^eKli\if»«^'^rille in front to Hght the meath. The little square hole in the middle of the I he nave is too small to have been the base of the altar. \-coss to the crypt is by two flights of steps, one in each ?e. The ground plan of the crypt shows that it was tj accommodate several tombs. Of these, the most inip<»ttant was at the east end apse, where I noticed the cover tone coffin inscribed victoria sancti moniale (?) m In the south side of the crypt are a number of other sto) ■>;, all desecrated and rifled. Passing across the crypt, and ing the stairs on the west side, there is a door in the ■■ he church immediately opposite the top of the steps 1. o little chambers shown in the photograph. The I a tomb. The other has an apse, and probably hft«i 'ol in it. re IB harfii', anything left of the decorative detail of the cn;ircii. I ' • e shaft of coloured pink African marble lymg in the > limestone capital of the Roman period AIN TOUNGA 139 with the cauHcole design hke the example in the monastery garden at Tibar, and a smaller one also of the same period. On an upright tombstone standing near the chancel steps is an inscription d. m. s. modia felicitas. The photographs show how roughly the nave was paved, and the irregularity of the plan and the general appearance of the work indicate that it was put together from materials obtained from the Koman city. When I first visited this church there was a piece of the cornice of the theatre with a Latin inscrip- tion on it lying in the nave, but I was unable to find it on my second visit. There can be little doubt that this church was rebuilt after the Byzantine conquest of the Vandals, in the later part of the sixth century, or the first part of the seventh.* The modern village of Teboursouk represents a Eoman and Byzantine fortress at the junction of the high road between Carthage and Tebessa with the cross-road from Beja and Pont de Trajan to central Tunis by the district of Gaffour. Near by were the two cities Ain Tounga and Dougga, with ruins of the same period containing many Roman and some early Christian buildings. Among these were the trefoil chapel at Henchir Maatria described in vol. 1, and another chapel like it near a place called El Aroussa, a station east of Gaffour on the Tunis — El Kef railway. After a long ride from Teboursouk through a mountain pass covered with trees, scrub and maquis and com- manding extensive views over the surrounding country I found the chapel at El Aroussa had been reduced to a shapeless heap of ruins. AIN TOUNGA The principal ruin at Ain Tounga on the road to Tunis is a well-preserved Byzantine fort built after the conquest of the Vandals from Eoman materials of the earlier city. The view of the inside of the fort shows how the Byzantines utilized the old masonry by the insertion of rubble into a framework of * Beside this little church the Christians used the Roman temples of Jupiter and Minerva and Ccelestis for Christian worship. These were no doubt appro- priated after the suppression of paganism in the reign of the Emperor Theodosius. 140 WESTERN TUNIS horizontal and vertical stones of large size taken from cornices, lintels, architraves, and the like. M. Guerin reproduces a long inscription from a Roman cornice now imbedded in the north wall of the fort. The long building with the apses at the extremities is said to be a bath ; but as there is some doubt about it I use the term here for convenience until the real purpose it was built to serve has been determined by excavation. It is a rectangular hall divided by two screens into three compartments communicating with one another by double doors in each screen. The two outer compartments of the hall are rounded off in the shape of semi- circular apses ; the former were covered with timber and tile roofs, and the latter by semi-domes.* The next two photographs can only be explained by reference to the plan on the fly-sheet ; the first is taken from a, and the second from b. I cannot help thinking that the idea of using engaged or detached pillars, so frequently found in early Christian buildings, to decorate the entrance to an apse was derived from some common Roman model of which this particular edifice supplies an example. The other buildings of interest are an odd little triumphal arch and the ruin of a building that bears a striking resemblance to the detached baptistery in the church of Servus at Sbeitla. The four pilasters at the angles of the square building, that was prob- ably the cella of a temple, are made of dressed and well-cut stones and certainly Roman. TIBAB About two miles west of the monastery of S. Joseph are some remains of a Roman country town. On the side of a hill stand the ruins of the sanctuary of a church interesting chiefly as affording an example of the favourite African plan of building a church apse with niches. Another example of this kind will be seen in the fragment of the apse in the building on the north-east corner of Sbeitla. The church was comparatively small; the principal apse is at the west end, and a porch or narthex at the east end. Beneath the nave is a crypt. * I am indebted to M. Merlin for the plan. TIBAE 141 From Tibar to Teboursouk there is a fine drive over the mountains by a zigzag road, and on the pass near the summit are the remains of another small town that had two triumphal arches of small size and plain design, and a church or chapel with an apse like that at Tibar. I saw no indications of Byzan- tine fortifications or other buildings either here or at Tibar, and infer that the remains belong to the early Christian period. In the monastery garden at Tibar there is a fine collection of Punic tombs, and some capitals of various dates. The monastery is, of course, quite modern. I gratefully record the hospitality shown to me by the White Fathers here under the presidency of my friend the Rev. Alexis Lemaitre, now Bishop of Setif and Vicar Apostohc of the Sudan, by Pere Cass, the mother superior and the staff of the little hostelry of the Sainte Famille where we lodged. EASTERN TUNIS UPPENA The ruined basilica at Uppena is one of the most important Christian monmnents in North Africa, for the original church, though not a large one, was destroyed at an early date, and a second and rather larger church was built over it. In the follow- ing account they will be called the first and second churches respectively. The ruins of the town of Uppena stand about two miles from the seashore of the Gulf of Hammamet, and quite close to the railway and the route nationale between Sousse and Tunis. The spot is now called Henchir Chegarnia, and about two miles north of Enfidaville, and six miles south-west of the large Roman fortress town Aphrodisium, now known as Henchir el Fradis.* After the Byzantine conquest Uppena was provided with a small fort made of materials taken from the ancient city to guard the road from the south to Carthage. This road was followed by Justinian's army after the successful landing at Caput Vada, the Ras Capoudia of to-day. From the materials used, it is evident that the second church was rebuilt about the same time as the fort. Nothing is known of the date or circumstances of the building, or of the destruction of the first church, or of the condi- tion the Greeks found it in when they came, but judging by the mosaics it could not have been altogether destroyed, and the old tombs were not disturbed, but the pavement of the second church was built at varying heights above the original floor. The first church was built to lie east and west, with the sanctuary at the west end. There was a nave with aisles on each * Possibly a corruption of the Roman name. It is also known as Ain Hallouf, the ' well of the boar,' and has been identified as one of the Vandal strongholds mentioned by Procopius. V>'v'^> 'Vi \^^^(^•^>^?. Q£ OJ, vt^iiifne^ ( AHM3SSU O 3MM3ITJHM3 3UPMt8A8 AJ 30 HAJ^ ^^"^ '^■^''^'^ ^'^^ '^"^ 39 UPPENA. Mosaics in the apse of two periods. Mosaic tombs at Enfida church. To face page 14J UPPENA 143 side, a semicircular apse at the west end, and another and smaller semicircular apse at the east end with a door on either side of it. On the north side there was a transept with a semicircular apse, and some square chambers adjoined it including a square tank or font. The nave was divided from the aisles by single pillars of brown sandstone with very plain caps to match ; there were six on each side, and also two detached pillars placed at the entrance of the chancel apse as in the garrison chapel at Haidra. The accompanying plan shows that while the centre of the second church was shifted a little to the north, some of the walls and doors built for the first were also used for the newer building ; so the same south wall served for both, but the north wall of the first became the division of the nave from the north aisle in the second. In consequence of the increased width of the second church the arc of the sanctuary apse and its position were en- larged and altered, and the small eastern apse was changed into a square one with a porch in front of it. The old chambers on the north side were also retained, and a new lobed basin was annexed to the square tank that did duty for font in the first church. The levels of the two churches can be best seen in the sanctuary and at the tank and font where the difference varies between a half and a quarter of a metre. The intervening space in the nave was filled up with the rubble and debris of the first church, including portions of the old stone columns, their bases, and capitals. The latter were used as substructures for the white marble pillars of the second church taken from a Eoman building that had been destroyed. I give two views of the sanctuary apse, one showing the old and the new apse walls, and the other the old pavement and a little fragment of the edging and the common vine pattern found in Roman mosaic work. The first view is taken from the point A on the plan looking into the two apses. The first apse is on the left and the second on the right. The second view is taken from the point B on the plan looking across the two apses. The second apse is in the foreground, and the white wall of the first apse is beyond it. There are no indication of bishop's throne or clergy seats in either of them, but the ground has been dis- turbed too much to distinguish the internal arrangements with any certainty. 144 EASTERN TUNIS A general view of a ruin of this kind is never satisfactory or of much practical use owing to the inevitable confusion of detail in the picture ; but with the assistance of the plan some of the features can be distinguished. The first view of the church is taken from the point C on the plan looking towards A, that is to say from the narthex to the sanctuary, showing the two apses in the distance. The more conspicuous objects, the white marble bases of the pillars on the left, a capital standing by itself on a block of stone in the middle of the picture, and the wall across the right side, mark the divisions of the old and new churches. The white bases belong to the pillars between the nave and the south aisle of the second church. The little capital marks the division of the nave and the north aisle of the first church, and the wall marks the north side of the second church. The second view is taken in the opposite direction from the point D looking towards C. On the right side is the south wall that was common to both churches. On the left are the same row of white marble pillar bases seen in the preceding view, but viewed in the opposite direction. Between them and the wall on the right some square stones in a row one behind the other mark the division of the nave and the south aisle of the old church. The next pictures are of a capital made of sandstone from the first church and the marble capital of a pillar of the second church. As there is no white marble in this part of Africa the latter were imported and perhaps obtained from a Roman building. But I have some doubt as to the provenance of these capitals, for it will be noticed that the acanthus foliage is cut in very low relief in distinction to the high relief of the earlier style. And for that reason it is quite likely that the capitals were not obtained from a Roman building but made new.* The next object of interest is the font, consisting of two parts ; f a square tank made for the first church in the old floor level, and above it the font of the second church, made in the usual lobed shape, associated with churches rebuilt during the Byzantine occupation. This lobed design appears to be derived from a pattern sometimes found in Roman thermae. * Compare this style of low relief with that of the Roman capital at Tibar and that of a late capital from Ephesus now in our church at Lower Kingswood in Surrey. t This is illustrated on Plate 57, vol. i. I now conio t . were provide.! This part of t/ii vi>?ion of the ecci ^ any rehcs iiiai ■ *Viat the iu;it"ytw'\:i ^-Js&K Vs?.\Z \o nooV\ Ajr of th(; i,A«ov^S\. uowsi'^ ^ )riginal inscripi ition of the vl ted on the ti Th BQL National e. i Mgr. Toulotte. / de la basiliqite d' Iqyp'-!' I Prods Verbaux . there on the Plate to tn 7\t\ '^^'i^\ "i^^X oX 39a Floor of Sim Ahfch church. (Canon Raoul^i ^•,, , >.. %Jtiinb in the church at Enfida- from uppena. Tomb of the nuirlvrs. To face page 145 UPPENA 145 I now come to the apses at the east end of the basilica that were provided in both churches for the so-called martyrs tomb. This part of the church was carefully excavated under the super- vision of the ecclesiastical authorities for the purpose of pre- serving any relics that might be found. In the first place it was noticed that the mosaic pavement with an inscription (a) found on the floor of the square apse of the second church was a replica of the original inscription (6) in the semicircular apse of the first church, but the latter was too much mutilated for preservation. I do not hesitate to reproduce the photograph of (a) given in the publication of the ' Mission Scientiflques.' * This mosaic is now deposited on the floor of the chancel of the modern church at Enfidaville. The inscription is sufiiciently legible on the photograph to make a transcript of it unnecessary.! The first names, of the Apostles Peter and Paul, were probably inserted honoris causa. It has been suggested that the other persons whose names appear may be identified as follows. The two Saturninus were father and son who suffered martyrdom in 304. Vindemius was perhaps Althiburus who was condemned to exile with Honorius in 484 ; Donatus, a martyr in the persecution under the Emperor Decius ; Jacler and Cecilius, colleagues of S. Cyprian ; and Lucilla the Saint whose name appears in the Carthage Calendar under date February 14th. In the earlier inscription (b) the words from * Gloria ' to 'Omnibus' were omitted. All that remains of it is the following fragment, A T V R T V R N I N G V D V D A ORTVNIADER EIIIINONAVG VEMBRIVM and on the front of the step, TATIS SVAE BEATISSIMIS MARTVRIBVS t • Nouvelles Archives des Missions Scientiflques, Tome XV. Fascicule 4. Paris Imprimerie Nationale, 1907. t Mgr. Toulotte. Proces verbaux d'une double Mission Archeologique aux ruir.cs de la basilique d'Uppena. Pub. Tunis, 1906. I Proces Verbaux; a photograph of the last part of this inscription is given there on the Plate to face p. 22. 146 EASTEEN TUNIS No relics that could be identified with the persons com- memorated in the inscription were found. Under the earlier apse there was a jar and under the later one two children's tombs, parts of two tombs that had been made for the first church, a jar, and a small box containing some ashes. Before describing the tombal mosaics found on the floors of these two churches, I may conveniently mention here another small church at a spot called Sidi Abich between Uppena and Enfidaville. This basilica had a semicircular apse at the west end, and beyond it a baptistery. It is ascribed to the Byzantine period, and probably rightly, as the font that obviously formed part of the original design of the church is of the lobed variety usually associated with that period. The floor of this church was also covered with a fine set of mosaics including several tombal epitaphs that have now been moved into the new French church at Enfidaville and mixed up with those from Uppena. The floor was so wrecked in the process of moving these tombs that the interior arrangements are not now distinguishable, but a sketch made before this happened is of sufficient interest to be worth reproducing here, because with a strong glass the names and places of the mosaic tombs can be identified with those on my photographs.* The smaller tombal mosaics that were moved from these two churches are now on the walls of the new church at Enfidaville. The first three found on the south side of the chancel at Sidi Abich are very roughly executed and were fitted into the earlier mosaic pavement. They belong respectively to Vitalis Famulus, a hermit called Eenobatus, and Faustinus. The fourth to Bellator came also from Sidi Abich and was found on the north side of the Baptistery. The next three from Uppena belong to Pascasius Keparatus and Julius Honorius a flamen. The next three, also from Uppena, belong to Lucilianus Vernacla and to Crescentius a deacon and Brutannicus his son. The pascal lamb in the wreath is one of a series found on either side of the nave at Sidi Abich. The tomb to Faustina * The original watercolour picture of this floor was made by M. Demont, and the Rev. Canon Raoul photographed it for reproduction in the Proces Verbaux, p. 6, where it will be found. ch at Enfida Turn's. To face page 146 146 EASTERN TUNI8 ^^ So rolies that could be identified with the person** ooin- I in the inscription were found. Under the :ip- ' 111. le was a jar and under the later one two chiiuM ■ -. tonil>8, paits of two tombs that had been made for the fir-t churct , . jar, and a small box containing some ashes. Be{.>D^ describing the tombal mosaics found on the floors of th- ■ ^" . churches, I may conveniently mention here another rch at a spot called Sidi Abich between Uppena and This basilica had a semicircular apse at the we . i>d, and beyond it a baptistery. It is ascribed to the By;- penod, and probably rightly, as the font that obviously lumn - jmrt of the original design of the church is of the lobed variety uHually associated with that period. The floor of this church was also covered with a fine set of mosaics including several tombal epitaphs that have now been moved into the new F-'^v church at Enfidaville and mixed up with those from I ^ The floor was so wrecked in the process of moving these torn that the interior arrangements are not now distinguishable, b- a sketch made before this happened is of sufiicient interest ' < be worth reproducing here, because with a strong glass tli names and places .•6f»^lt^V»?v^<^iV«^ t^«»T«^a d^Vi ^(A^i^ed wri' those on my photographs.* The smaller tombal mo8ai«.> Lu.vt were moved from tbcn<. ..v churches are now on the walls of the new church at Enfidavillf. The first three found on the south side of the chancel at Sidi Abich are very roughly executed and were fitted into the earlier mosaic pavement. They belong respectively to Vitalis Famulus, a hermit called Renobatus, and Faustinus. The fourth to Bellator came also from Sidi Abich and was found on the north side of the Baptistery. The next three from Uppena belong to Pascasius Rtv '^'••" aiid Julius Houorius a flamen. The next three, also from i Ixiiong to Lucilianus Vernacla and to Crescentius a deacon and : annicus his son. ■ he pascal lamb in the wreath is one of a series found on «5ide of the nave at Sidi Abich. The tomb to Faustina ;;.M!ial watercolour picture of this floor was made by M. DemiMt. *t.d the K' i photographed it for reproduction in the Precis Vtri UPPENA 147 and the child spenden (spes in Deo) conies from Uppena.* In passing I must not omit to draw attention first to the similarity between the cross in the corner of Faustina's tomb and those on the altar slab in the church of S. Giovanni at Gaeta, and in S. Maria in Cosmedin and S. Sabina at Kome ; secondly to the Byzantine border pattern that so frequently occurs in this part of North Africa. Beside the large Byzantine mosaic of the saints the following mosaic inscriptions t on bishops' tombs also from the basilica at Uppena are now on the floor of the chancel in the church of Enfidaville : Baleriolus Episcopus vixit annis LXXXII depositus die Gil kal Octobres. Above : + P in monogram in a wreath of foliage. On another tomb : Honorius episcopus vixit annis XC depositus sub die Gil Idus Augustus. Above : + P in monogram with a w depending from the arms of the cross. It will be noticed that the crosses at the heads of the examples from Sidi Abich differ from one another and from the others. The crosses over the tombs in the second group are substantially alike with A and t5 suspended from the arms of the cross. Those on the next three have A and jj in the field with the combined XP that is usually called the Constantinian monogram. I doubt the value of these various forms of cross for fixing dates, and have something to say upon this question later on. There is, of course, no doubt as to the date of the By- zantine mosaic of the martyrs, but the two problems that have so far not been solved are, what was the date of the earlier mosaic, and did the persons commemorated suffer martyr- dom as Christians during the pagan persecutions, or as Catholics in the times of the Vandal occupation, more particu- larly in the reign of Hunneric. The information at present available is altogether insufficient to answer them. Included in this series t is the mosaic found on a priest's tomb at Thinna, the important predecessor of the modern city of Sfax. It is now in the municipal museum with a large collection of sacred or secular objects found in the neighbour- * Illustrated in vol. i. on Plate 72. t Photographs of these will be found in Nouvelles Archives quoted above. 148 EASTEKN TUNIS hood. Among them is a cedar- wood coffin, probably of a Eoman officer, which contained his body and a bronze sword lying by his side. Bir bou Bekba Near this station on the Tunis-Sousse railway, where a branch line runs to Hammamet, are the ruins of a Eoman town and of an early basilican church. The latter is built on an unusual plan, is correctly orientated, and belongs, I think, to the early Christian period. The fabric has been completely razed to the ground, and all that remains are some fragments of pavement, the foundations of the altar that stood in the semicircular apse, and a small part of the outer walls on the south side. The numerous fragments of mosaic and variety of coloured marbles show that the church must have been richly, decorated. The plan seems to in- dicate that, contrary to the usual practice, the apse extended to the whole width of the church so that the aisles were carried round in the form of an ambulatory. A door in the centre of the outer wall of the apse led to a large octagonal baptistery.* In the centre of this stood the font of the same early type as the example now in the Bardo Museum, made of white marble, with two flights of three steps leading down into a circular cuvette sunk into the floor. This is probably an example of the earliest type of font for immersion that preceded the lobed variety found in churches belonging to the Restoration period after the Byzantine conquest. I was unable to take any pictures or any measurements, in consequence of the proximity of the French fort, but during our visit we were kindly received by the officers who took us into the barrack square and showed us some stelae and one or two capitals of the Roman period. The more important Roman remains, consisting of a large group of farm buildings and a nymphaeum, testify to the importance of this place, and it was probably then, as now, a military station. From Bir bou Rekba the turnpike road follows the * I give a plan of this on the fly-sheet opposite. CAETHAGE 149 railway into low ground, and approaching the seashore passes by a large cylindrical tomb made of stone not unlike that of Cecilia Metella on the Appian way. The first village in the direction of Sousse is Bou Ficha an important agricultural centre belonging to the Enfida Domain. On the wall of the house inhabited by the agent or factor of this agricultural company are some crosses* and one or two capitals from the site of the ancient city of Aphrodisium the modern Henchir Fradis in the foot hills near by.f There are no signs of a Christian church in Henchir Fradis, but several Roman ruins, including a handsome little triumphal arch and a barrage at the modern well, known as Ain el Hallouf. The Arabs of this place were very friendly, and the headman insisted on presenting me with a well-preserved coin of Constantine I. The site of this Roman city has been very imperfectly explored, and this and several other towns in the neighbourhood seem to offer a promising field for future research. CABTHAGE The three principal churches in Carthage appear to have been the great Christian basilica known as the Damns el Karita, the church in the amphitheatre dedicated to the martyrs S. Perpetua and her companions, and the large basilica on the south-east of the chapel of S. Louis, known as the Byzantine church. | The last is the only one of the three where the original plan can now be easily distinguished and compared with other churches in North Africa. Nothing remains of the church of S. Perpetua, and the site of the Damns el Karita has been so completely destroyed in the search for relics and epitaphs that it is almost impossible to make out what belongs to the original church or were later additions and restorations, or indeed to distinguish what the ground plan was. It is clear, however, that there were two churches built at different dates on the same site. I shall • Illustrated in vol. i. Plate 73. t Two crosses and one capital are illustrated in vol. i. on Plates 68 and 73. I In these notes this church is called for convenience the basilica of Justinian, upon the assumption that it was built after the Byzantine conquest. It is quite as large as the Damus el Karita, though I have also called it the ' smaller ' basilica, teferring more to archaeological importance than to size. 150 EASTEEN TUNIS begin therefore with the so-called Byzantine basilica, premising that it was apparently built at one date and not piecemeal, nor as the restoration of an earlier church. The Byzantine Basilica This great basilican church has a nave with double aisles, a semi-circular apse at the south-east end, and a baptistery at the north-west corner. It must have been a fine church, and judging by the carvings on some of the fragments lying about, the font and the pattern of the mosaic floor, was built after the Byzantine con- quest. It is designed in the form and with the internal arrange- ments of the ordinary Latin basilica of Eome, where the altar stood out in the open, and not as in the later development of Byzantine architecture, behind a screen and concealed from the view of the congregation. The character of the church can be easily understood from the photographs and the plan opposite. The general view is taken from high ground at the north end looking up the nave towards the apse. The nave was divided off from the inner aisles by rows of marble pillars in pairs, and the inner aisles in their turn are separated from the outer aisles by rows of single pillars. From the floor, which was covered with mosaics in various patterns, we get perhaps a better idea than from almost any African church of the division of a nave into choir, altar space, and compartments to accommodate the women and different ranks of the worshippers. Nearly all the nave was occupied by these enclosures commencing with a narrow gangway in the middle that gradually broadened out till it reached the altar space where it was carried round the sanctuary enclosure. The arrangement can be easily understood from the plan, and the different divisions can also be recognized in the photo- graph. The narrow or west part of this gangway marked A on the plan, appearing in the foreground of my view, was enclosed on each side with white marble balusters, and the floor also seems to have been slightly raised above the rest of the floor of the church and either paved with marble or mosaic. The floor of the nave on either side of the gangway was also paved with mosaic. The screen then broadened out for two 4« .1 Q Q ■'u o o o o o o o o ^ol JL___Q !o^L &i3 o — v~v*r— "T-^ 7b /(S'c^ /^ji?'^ 150 150 EASTERN TUNIS i^ begin therefore with the so-called Byzantine basilica, premising that it was apparently.'3j£ltiic^.'S6'ifen]fe date and not piecemeal, nor as the re8totw(»»^^6f "^$in^^^ii«e (djisfecii^ semi-circulajr a^se north-west con ler, by the carvings on » the pattern of 1 he m^ quest. It is de signed i ments of the crdinary view of the for Th^. My ran tin e.Bas This great BflJBJit' ?an church has a n[i^ ith double aisles, a and a baptistery at the le c hurch, and judging nng about, the font and buillfc after the Byzantine con- th^form/dndywith tho internal arrange- btin basiliga of Ron;e, where the altar stood out in tlje open, and not as in t he lalj er development of Byzantine arcl re, behind a screen d»rn>cbncealed from the iti on. [V^i The characior of the church can be easily understood from the photographs and 't he plan op posite. The general view is taken from high grouiidr^t*fh'e north end looking up the nave towards the apse, o^^u rui^ was divided off from the inner aisles by rows of -jnarMei pillara iKpaica^ and the inner aisles in their turn are separated l trom thqf)i Fro ttji iVg floBRB wjiinh Thq f)uter aisles by rows of single pillars. pati ^rns, we get [perhaps Afr: oan ^arcfe^ rthc-dirifii conl^artments to Acc omm odate the womec and different ranks of lEfiie lOve Imsls occupied by these -Avilth a narrow gangway in the middle itSS it Inched the altar space where the enc ;or0hippSS )sures coram finding: 3vor(^^— vvith mosaics in various better idea ;han from almost any a n^e irvt 3 choir, altar space, and tha ; fera^all3^ld«tfine( it \ris carried rounc th4 sanctuary enclbjsure. fhe ^rrac^^iaen" ; ca q b^^asiljP und ( rstood from the plan, the different di^ifHis can also be rppognized in the photo- ^heffi row « : we^-%art ^f thl^z^Sjugwai^ims^ked A he plan, appei^jinlg^. in the foreground^ (^ giv vi i n, was )sed on eiron side with wmte marbleOaalgstgr^ ?^d ; le floor also, seems toJiave been slightly raised ab^v^,^e ^res ; of the riocD of the cKurch and eitner paved with^n^r^l^ or mosaic. Th(t iflo»f of j ^bh ^ nave on eij^r sicj^ of{ - )bhe gangway /i as also pav'e^ with mosaic. The scrfeen th6n broadened uirt or two 0?^\ ^-^^vi^ ^:)a\^ oT CAKTHAGE 151 bays till it reached the sanctuary enclosure where it was carried round the altar. The altar itself was a large square erection, and has, of course, disappeared. The foundation of cement that stands up above the floor level can still be distinguished, and was apparently hollow in the centre ; the space beyond it was open and covered with mosaic of the same pattern as the rest of the gangway. At this point the arrangement of the baluster screens is not clear, but it seems as if there were an open space at B between the end of the altar enclosure and the entrance to the apse. Lying close to the altar are two blocks of white marble, either capitals or portions of the bases of pillars, with crosses carved on them. These either belonged to the screen or to a canopy over the altar. Beside the screens round the altar there are indications on the bases of some of the nave pillars that screens shut off part of the aisles from the rest of the church. But as some of these bases have been replaced in the restoration I cannot be sure that they are in their original positions. It is, however, clear that there were screens round the altar fitted into the nave pillars as well as on little posts as at Tebessa. Only a few courses of the semi-circular structure of the apse at the south-east end remain. This part of the church seems to have been either lighted by windows, or decorated with niches like those at Tibar, and in the small churches at Sbeitla and elsewhere." The floor of the apse was raised like a dais above that of the nave and approached by two flights of steps, one on each side of the front of the apse. The space between them is now empty, but probably was decorated with an open screen, as in Sta. Maria in Cosmedin at Rome. There is some indication on the floor of the apse that seats for clergy were provided according to the usual practice. Passing down the nave to the north end of the church we come to the baptistery. This was a square chamber, or perhaps an open court or cloister with sixteen pillars supporting the roof. In the centre stood a canopy supported on four small pillars fixed * Saladin mentions another church with an apse like this one. See illustration in his report, p. 52, Archives des Missions Scientifiques, Tome xiii., B^me serie. The place is now called Henchir el Baroud, and is situated in Central Tunis in the Oued Djilma, near Hadjeb el Aioun. 152 EASTEKN TUNIS in the corners of the font beneath. The font * itself was of the usual variety so commonly found in Africa. The top is hexagonal, and from two sides two little flights of steps lead down into the circular cuvette which was lined with marble. In this case, as also in the baptistery of the large church at Sbeitla, the bases of the outer row of pillars were provided with grooves to hold low screens. One would like to know exactly how the ceremony was performed in practice. In the fonts at the Damus el Karita, at Bir Bou Kekba, and in the Bardo museum, the steps lead down to the cuvette, and the cuvettes themselves are large enough to admit of an adult walking down and standing in the font, but in this and many other smaller fonts made with lobes, the cuvette is hardly larger in diameter than an average size hand-basin, and the little steps look as if they were intended for ornament rather than use. In any case it must have been difficult for an adult to use them, or indeed to stand in the water at all, and submersion in the cuvette was cer- tainly impossible. Beyond the baptistery is a chapel with an apse, and some other buildings that communicated with the basilica. They are of no particular interest and are destroyed almost beyond recognition. The material of the outer walls is rubble made of old materials, supported at intervals by large upright stones. This was the usual Byzantine method of construction seen in the fort at Ain Tounga and the church at Dougga. The main entrance and porch are on the south side of the church, but nothing of them remains. The church which must have been a very handsome one, and among the largest in the country, was no doubt destroyed in the beginning of the eighth century, after the defeat of the Byzantine fleet and the capture of Carthage by the Saracens. The Damus el Karita The accompanying photograph of the Damus el Karita will give some idea of the complete condition of ruin that even the floor of this great basilica has been reduced to by the repeated excavations * Illustrated on Plate 57 in vol. i. CARTHAGE 153 in searcli of epitaphs and the remains of early Christian saints and martyrs. The ruin is so complete that it is difficult to identify the diiferent parts of the church, even with the assistance of Pere Delattre's plan. The view is taken from the west end of the church looking towards the main apse. The larger blocks of masonry on each side are the piers of the nave which extended to the spot where the distant figure is standing in a portion of the semicircular apse ; the latter is also shown in another view. The original church, which must have been a very large one, with double if not triple aisles, was built upon the ruins of a Roman building, and is reputed to be the earliest church built in North Africa in the reign of Constantine. At a later date, probably after the Vandal conquest, when the original church had perished, a second and smaller church was built over the site. It was much shorter, as portions of the semicircular apse belonging to it have been found in the middle of the earlier nave. The little trefoil building at the extremity of a large semicircular atrium or court on the south-east side of the church, and a large font on the south-west side are the chief points of interest for my present purpose. The font is a large circular tank with a flight of steps leading down into the saucer-shaped cuvette, and was apparently in an outbuilding or baptistery adjoining the church. In the picture, which is unfortunately taken in a bad light, my servant will be seen standing on the steps, while some debris below mark the bottom of the cuvette. The trefoil chapel at the end of the atrium, or nymphaeum, now completely gutted, is described in the preceding volume. The traveller who goes to Carthage will of course visit the museum in the monastery of the White Fathers where Pere Delattre has collected together numerous fragments from the ancient city. These include a large array of lamps, numerous coins, a terra cotta model of an organ, and lead sheets inscribed with prayers and imprecations used in chariot and horse- racing. EGYPT SMALL COUNTRY CHURCHES Before proceeding to describe one or two small churches in Upper Egypt and Lower Nubia I must trouble the reader to revert to the notes on the White monastery at Sohag. It will be observed that within the ' curtain ' or outer wall of the monas- tery there are the remains of no less than three churches of different dates : first the basilica of the early Christian foundation with the trefoil sanctuary ; secondly the present Coptic church, built in the ancient sanctuary ; and thirdly a small church belong- ing to an intermediate period, built in the outer corridor on the south side of the basilica. The difference in the ' ritual ' arrangement and position of the altar* between the first and third edifices on the one hand and the second on the other will be at once apparent ; in the latter the altar is concealed behind the iconostasis or screen according to the Eastern practice, while in the former it stood out in the open according to the Western practice ; the apse of the little church in the narthex was also fitted with seats for the clergy and a throne in the centre for the bishop as at Announa + and elsewhere in North Africa. Some of the churches now to be described fall into one of these two categories and some in the other. The two basilican churches in Nubia, the north church at Addendan and the south church at Faras have apses with tribunes and the altar stands in the open ; the use of pillars to support the chancel arch is another feature that makes them almost exact counterparts of the church at Announa in Algeria. The altar arrangement of the three domed churches at East Serreh and the east church at Adden- * In fact there are three altars in the present Coptic church. t The arrangement at Announa is ahnost exactly the same. EGYPT 155 dan, also a domed building, are more doubtful ; the chancel arches of the former are much too small to admit of an iconostasis, and if the altar was screened at all a curtain must have been used. The churches in Egypt at Medinet Habu, and Nakadeh, and the little chapel at Abd el Kader near Abu Seir are built on the com- partment system and they have altar screens made of stone or brick. The church of S. Simeon at Assouan is also included for the sake of a piece of vaulting in the entrance porch that happens to be well preserved and conveniently illustrates the Coptic method of cross vault building ; but the student who wishes to know how the Copts built their barrel vaults should consult two recent publications on Christian architecture in Egypt and Nubia.* The churches on the Nile below Wady Haifa were excavated by the Eckley Coxe expedition and have been described by Messrs. Mileham and Maciver. They were found at the villages of Serreh and Addendan on the east bank and at Debereh and Faras on the west bank. To these must now be added another church at Faras, discovered by Mr. Griffith and Mr. Woolley, and in some respects it is perhaps the most interesting of the series, for it contains a number of well preserved wall paintings that throw some light upon the date and origin of those at Abu Seir. • Churches in Lower Nubia, by Mileham and Maciver, for the Eckley B. Goxe, Junior, Expedition to Nubia. Published by the University Museum, Philadelphia, 1910. Christian Antiquities in tlie Nile Valley, by Somers Clarke, 1912. LOWER NUBIA ADDENDAN North Church The churches described by Messrs. Mileham and Maciver fall into two categories. First, the common plain basilica with a semi-circular apse, seats in the chancel, chancel arch pillars, altar in the apse, and in one case a pulpit ; and secondly, square churches covered with domes. An unusual plan was found in one of the churches at Addendan, that must have resembled S. Maria delle cinque torri, Monte Cassino, with a central hall and little towers covered by cupolas at the four angles. I give two pictures of this church. The North church at Addendan is a basilica, with nave, aisles, narthex, and a semicircular apse with square sacristies on each side of it all enclosed in a square outer wall. The apse is pro- vided as at Announa with seats for the clergy arranged in tiers ; small doors communicate with the adjoining sacristies, and the altar stands out in the open between the pilasters of the chancel arch ; each pilaster is recessed on the nave side to take a granite pillar surmounted by a stone cap ; these no doubt supported the chancel arch. Upon this arrangement Messrs. Mileham and Maciver say as follows : — ' The tribune occupies the greater part of the sanctuary, with the result that the altar has been placed more to the west than is usual. It is doubtful if this was the original arrangement, for the doors to the sacristies are almost meaningless at the present time as they are blocked up for nearly half their height by the brickwork of the tribune. The altar, which was built of crude 4a ADDEND AN NUBIA. North church. General view with the Nile and interior looking West (with my guide); and East with pillar of chancel arch. To face page 156 t». LOWER NUBIA ADDENDAN North Church The churches described by Messrs. Mileham and Maciver fall into two categories. First, the common plain basilica with a semi-circular apse, seat&in the chancel, chancel arch pillars, altar in the apse, and in one case a pulpit ; and secondly, square churches ■ 1 A^^'W^^I^Yl'^^cSR"^^^^ P^^^ ^^^ found in one of the ...,._. ;l- al AdaLiidan, that must have resembled S. Maria delle cinque torn, Afontp uassino, with a central hall and htlJ§^3\«^o^V.W:*>>w» V^;a^VV ?>i^ .^^^i(SnrsaVig49»^ I give two pic^Bs^^^fe^«i> The North church at Addendan is a basilica, with nave, aisles, narthex, and a semicircular apse with square sacristies on each side of it all enclosed in a square outer wall. The apse is pro- vided as at Announa with seats for the clergy arranged in tiers ; small doors communicate with the adjoining sacristies, and the altar stands out in the open between the pilasters of the chancel arch ; each pilaster is recessed on the nave side to take a granite pillar surmounted by a stone cap ; these no doubt supported the chancel arch. Upon this arrangement IVT.^ssrs AFileham and Maciver say as follows: — ' The tribune occupies the greater part of the sanctuary, with the result that the altar has been placed more to the west than is usual. It is doubtful if this was the original arrangement, for the doors to the sacristies are almost meaningless at the present time as they are blocked up for nearly half their height by the brickwork of the tribune. The altar, which was built of crude ^^\ ^-^v^^ ^:i»\ oT LOWEE NUBIA 157 brick (not of burnt brick as in the other examples), stands between the piers of the Arch of Triumph.* This position would not allow of a screen between the columns, yet that a screen existed at some period is almost certain, for there is a rough seating for a beam cut in each column a little below the necking.' They do not date this church, but to a church near Debereh a few miles distant they assign the tenth century, and as the chancel arch capitals in both are almost exactly alike we may, I suppose, infer that the two churches were built about the same time. It will be seen that the general and internal arrangements of this church resemble those of the churches in Algeria and Tunis, and so far as construction is concerned an earlier date than that given by the American expedition might be assigned to them. The evidence of an alteration in the ritual arrangement is even more apparent in the church at Debereh, where a portion of an iconostasis remains that did not form part of the original design. FABAS The plan of the South church at Faras is almost exactly the same, but the altar stands a little nearer into the apse. Besides the same seats for the clergy arranged in tiers, this apse is also provided with three niches in the walls. The suggestion t that this triple niche arrangement is symbolic of the Trinity is quite likely to be well founded. The purpose these niches were intended to serve is indicated by the lamp stand in one of them, and the photograph shows that the oil stain is still visible. For the interesting collection of objects found the reader is referred to the Eckley Coxe publication. The characters on a fragment of parchment found in this church are attributed to the sixth century or the seventh, and the similarity to the North African churches justifies a preference for that period. * The chancel arch. f Churches in Lower Nubia, by Mileham and Maciver. The arrangement of three niches will be found in the Rocelletta in Calabria and something of the same kind occurs in the apses of the church at Tibar and in the smaller basilica at Carthage in Tunis. 158 LOWER NUBIA EAST SEBBEH This ruined town stands on the east bank of the Nile about 12 miles north of the Wady Haifa. It contains three small churches that are interesting chiefly from the architectural point of view for they show that the squinch and pendentive were used concurrently to support a dome built over a square sub- structure. From the liturgical arrangement they do not differ substantially from the early African church model. There was but a single altar in a sanctuary at the east end, and a small chapel of the prothesis and the diaconicon communicating with it on each side. These chapels contain recesses in the wall that were no doubt used to keep the sacred vessels, books, and other church furniture. The plans are reproduced from Messrs. Mileham and Maciver's book, and as they have given a detailed survey of the churches I need do no more here than illustrate the method of supporting the domes and reproduce my photograph to give a general idea of the appearance of these little buildings. The first and smallest church stands just outside the town wall on the north side. It is made of sun-dried bricks, and seen at a distance looks like a box with a small cupola rising from about the middle of it. The first view is taken from the south- east corner and shows the south side, the door of the narthex, and the cupola above. Entering the church by this door we come into the narthex or corridor that occupies the whole of the western end of the church, and was provided with another corre- sponding door at the north end. The next view is taken from the narthex looking along the small nave and showing the chancel arch now blocked up and all that remains of the crown- less dome that covered the eastern bay of the nave ; the sanctuary is beyond the central arch, and the two arches on the right and left give access to the north and south aisles of the nave respec- tively. With the exception of the dome all the rest of the roof has a barrel-vault, built in the usual Egyptian way. The squinch can clearly be distinguished in the view of the dome that was made, like the rest of the church, of sun-dried bricks. 43 EAST SERREH NUBIA. North church. Dome over chancel. -ace page /5 quite modern '^fi^^^^l^^y ^he kiwaris and his tipted here stands borders of the i stance looks like the desert about ftr^ sin happened to "unding country f thf south-east r{)ken roof in the chamber. The are the apses of next three pro- ■ a M i tisuturj i rliwp e l, and a g ai * I thiuk thi« must be the building described by Mr. Somers Clarke (p. 121) as Dc^r I'i M.Vlak M'khail Kamula. o^\ ^■^v^\ ^-it^'X 0^ EGYPT 161 jections are the chancel apse in the centre, and the apses of the prothesis and the diaconicon of the principal church. The little domed building on the left is detached from the main group seen in the picture. The entrance of the monastery is round the corner where the donkeys are standing, and just beyond the detached building. The next view is taken from the courtyard on the west side, and shows the entrance to the narthex. It will be noticed from the plan that the church consists of a number of small squares. This plan was adopted as no wood was available, and the architect had to construct his building in such a way that it could be covered by brick domes. In the circular chancels, or haikals, the domes are merely built up over the substructure like beehives, but in the rest of the church, the nave, narthex, and outer apartments, they are supported by squinches resting on the angles of the square substructure. The church is divided into a narthex of three squares, a nave of six, a chancel of three, and three haikals or sanctuaries. The general view is taken from near the font looking towards the Epiphany tank, which was covered by boards, and at the time of my visit contained about 18 inches of water; it was used once a year. The view of the chancel, taken from square 9, shows on the right hand side the screens in front of the central apse and of the prothesis. Another picture shows the screen in front of the central apse, in this case a circular domed room, and a corner of the altar behind. The latter is as usual a mass of stone masonry, about one metre square, with a slab on the top of it. In the wall of the apses there are one or two recesses or niches intended to hold lights, to receive the sacred vessels, or a lavabo for the celebrant. The altar farniture consists of a little wooden box or tabernacle on the altar, a chalice, paten, handcross, a candlestick or two, and a number of little ceremonial serviettes which play an important part in the service in this hot country. The student will of course consult Mr. Butler's work, and he will find illustrations of these and other altar furniture, their significance and object. The other view of this church, looking down a passage into the adjoining chapel, shows a recess or niche cut in the wall containing the kind of jar that usually does duty for font in Coptic churches. 162 LOWEK NUBIA MEDINET HABTJ S. Theodore This church is situated in the desert about three hundred yards south of the mausoleum of Rameses and close to the valley containing the tombs of the Queens. In general appearance it is not unlike Nakadeh, but it is smaller, and being better preserved is still used for worship. The church is divided into squares covered with domes as at Nakadeh, making sanctuary, chancel, nave, and baptistery with a narthex. These are all numbered on the plan. 1 to 4 are the haikals for as many altars, of which the principal is 3.* 5 is the baptistery with the Epiphany tank, and 10 the narthex in front of it. 6 to 9 are the chancel, and 11 to 18 the nave. The apartments for the women 11 and 15. The annexes 19 to 21 are used apparently for the same purpose. The general view of the nave is taken from A on the plan. The screen facing the spectator at the end of the church is the women's portion. The thick wall with the pointed arch in it, seen in the middle of the picture, is the screen between the chancel and the nave. The only decoration on it is a panel with three dedication crosses on it that should be compared with examples at Sbeitla. Beyond the chancel stands the screen of the haikal concealing the main altar : this screen is hung with pictures like a Greek iconostasis. The pictures include S. George, and are poor and obviously modern. The domes in this building are made of brick supported on squinches, of the same construction as at Nakadeh, and they are pierced with circular holes to admit the light and air. * The only way of distinguishing the principal altar in this and other churches where there are several altars in a row, seems to be by the ' trappings.' In the older churches only one altar seemed to be furnished, but in the modern church at Nakadeh all three were provided with cloths, serviettes, tabernacles and candles. The vessels, made of glass or base metal, were tied up in a cloth and laid on the altar. 45 MEDINET HABU, EGYPT. S Theodore. Exterior and stone with crosses; and plan. ASSOUAN. S. Simeon. Detail of vault. To face page 162 z^ 1(32 LOWER NUBIA MEDINET HABU S. Theodore This church is situated in the desert about three hundred viu<)^ -(.xirh of the mausoleum of Rameses and close to the valley coi: ' tombs of the Queens. In general appearance it is noi. _ .LKadeh, but it is smaller, and 'being better preserved is still used for worship. church is divided into squares covered with domes as at h, making sanctuary, chancel, nave, and baptistery with a . . . K. These are all numbered on the plan.- 1 to 4 are the haikals for as many altars, of which the principal is 3.* the baptistery with the Epiphany tank, and 10 the narthex in of it. 6 to 9 are the chancel, and 11 to IB the nave, piirtments for the women 11 and 15. The annexes 19 to 'il are used .^^^^g^tl^'^^^ WmW . , The general view of the nave is taken from A on the plan. reen facing the'^s^ec^ator at the end of the church is the ir^?^^^bm^'^«PW? ^!«i^WiSl^^fl^fto^^"^inted arch in it, seen in the middle of the picture, is the screen between, the hancel and the nave'. TTir on^ decoration on it is a panel with ' hwe dedication crosa^S^I?' \^ that should be compared with ♦ xamples at SbeitlgSi\MsB^SfDfiA\*lie chancel stands the screen of the haikal conceaUng the main altar : this screen is hung with s like a Greek iconostasis. The pictures include S. Greorge, I' poor and obviously modem. '■ domes in this building are made of brick supported on Mfuiuches, of the same construction as at Nakadeh, and they aiv ''■ with circular holes to admit the light and air. only way of distinguishing the principal altar in this and other churclii wi, ir« several altars in a row, seems to be by the 'trappings.' In tho oldv 3 only one altar seemed to be furnished, but in the modem church at N ' ree were provided with cloths, serviettes, tabernacles and candles. The . f glass or base metal, were tied up in a cloih and laid on the aitar. s.?S\ ^-^ssA, v^vi\^ o\ i\ i-N I 1 )1^\ 1<>8 WADY HAIFA aaqaeniiy viBtieil by liiuiupeans, as the It stands ahout three hundred yards ^"S^^'K^Vl''^^'^^'^^^^^'' ^^ ^^ north of the . *iiiLhalf a mile further north A\'n«^'^ ..7. r, whence the -.e by. .\\:A '.^\ \u> i^^\v;u^s^^>vv^>^>nn>^\\e«?n^^m^'^!ofeV\^^%''J^ Wady Haifa I chartered a felucca to Tho steward of the government >|^'-'"r'rr!ssion to ■v"^'''»npany me, tliat -u ""TiT i>'4- M^ T?^ rf ■ )i' 1^ [had l>e«'| anif raeiof hft stl o Wady 1^ ■^■' jrnd !• mcasurii involved i :he north w .ui IS most ; ^^^;t4j^>4F\ iu tl V leave no"^?^^ •>'*'»*^^^ : . . ^ry, uave, D and E two ai^ ^nd G two outer aisles added at a later dat6. The r<> ^ed of \\ vaulting, with thr ■'■ dome 1)1 ci^jss \,i • The pUu ^ ^\ ^■^\ ^^li\ oT 46 ME DINE T HABU. S Theodore. JVave^ looking across the church. The chancel screens on the right and women's pews on the left. jr U) liX i^\ '/;l ^Z^ 5=. rl ^^--^1 fe^jfe=t% ' » ' ♦ i..-' Il l . ly • It ^' 7b yi?c^ /'''^^ '^'^J LOWEK NUBiA 163 WADY HALFA Chapel at Ahd el Kader, near Abu Seir This chapel has been frequently visited by Europeans, as the walls unfortunately testify. It stands about three hundred yards from the west bank of the Nile and three miles or so north of the second cataract at Abu Seir. On a hill half a mile further north stands an Arab turbe dedicated to Abd el Kader, whence the name of the hamlet on the banks of the stream close by. The illustrations and plans were made in the following cir- cumstances. On reaching Wady Haifa I chartered a felucca to visit the rock and cataract. The steward of the government steamer, an Austrian, asked my permission to accompany me, and on the way he pointed out this little building and said that it contained some curious wall paintings. In returning we visited it, and I then made some sketches of the interior, including a fresco of a Nubian king. I showed these drawings to my host, Mr. Griffith, who was then working at Faras, and judging that the chapel was worth more careful survey than I had been able to give to it, he proposed that Mr. Woolley and some of his staff including two native photographers should return to Wady Haifa with me, and that we should copy the paintings and make a plan.* We returned accordingly to Wady Haifa, where an agreeable week was spent with Mr. Woolley measuring the chapel and draw- ing the wall paintings. This involved a daily sail to Abd el Kader, and on one occasion the north wind blew so violently and the river became so rough that two of our native passengers in the felucca were seasick, and Mr. Woolley and I came also within measurable distance of what Americans call courting Neptune. The plan * is most peculiar, but the two little ritual squints in the wall A leave no room for doubt that B was the sanctuary, C the nave, D and E two aisles, and F and G two outer aisles added at a later date. The roof is entirely composed of waggon vaulting, with the exception of a small square space covered by a dome or cross vault that has crumbled away. The materials • The plan is by Mr. Woolley. 164 ^ EGYPT throughout are of sun-dried bricks. It is obvious that this miniature church was not intended for congregational worship, and the term votive chapel probably describes more correctly its real character. I give three views of the exterior. The third is taken on the roof to show that this consisted of a series of vaults made of bricks placed slantwise in the usual Coptic way, and a little square hole originally covered by a dome or vault. The fourth view is taken in the north door looking across the chapel from G to H, and the figure will give some idea of the diminutive proportions. There is nothing specially remarkable in the architecture to any one familiar with Coptic methods. The interest centres in a large and wonderfully preserved collection of paintings that cover almost all the available wall space inside. The subjects were first drawn in black outline and then painted in colours, chiefly red and yellow. The picture deserving particular notice is of a Nubian king with a peculiar crown on his head. There are four figures in this group : (1) God the Father, (2) our Lord on the extreme left presenting (3) the king, who appears to be either seated or kneeling. On his left hand is another figure (4) perhaps intended to represent the patron saint. Beside the peculiar crown with the two horns, notice the king's dress decorated with scroll work and what appear to be representations of the double-headed eagle. In his left hand he holds some object that I took to be a book of the Gospels, but, as Mr. Woolley points out, is more probably a model of the church. Another remarkable group represents Shadrach, Meshach, and Abednego in the fiery furnace, with the Spirit of God above protecting them from the surrounding flames. Compare their dress with that given in the Bible narrative and notice the check pattern trousers and the large cloaks. There are several eques- trian figures, including the Magi and three archangels. The photograph of one of the latter, whose identity we have not established, shows a horseman seated on a white horse holding a lance in his right hand with which he is spearing a scorpion. Notice the crown with three crosses on his head, his cloak flying out in the wind, his hessian boots, and the trappings of the horse. The details of some of the costumes, the trappings and the 47 Chapel Abu Setr. Wall paintings^ S T/ieodore on horseback. ^reseiifatviu of the church by a King with turban and cresent ornament o;/ ?'/, Shadrnch^ Mesach and Abedncgo in the fiery furnace. y^-^¥: To face page 164 IF. 5 . EGYi^T ^^ .iijujs^ui.'ui. are of bun-dned bricks. It lo .jL-yiw...- i,.,. i.no miniature church was not intended for congregational worship, and the term votive chapel probably describes more correctly its real character. I give three views of the exterior. The third is taken on the roof to show that this consisted of a series of vaults made of bncks placed slantwise in the usual Coptic way, and a little su illy covered bj' a dome or vault. The fourth vit;., -- ..«.»v .. .,. ;,iie north door looking across the chapel from G to F, and the figure will give some idea of the diminntive s nothing specially remarkable i a,r an) one familiar with Coptic methods. The iuL; . V -.. ....v-i in a large and wonderfully preserved collection of paintings that cover almost all the available wall space inside. The subjects were first drawn in black outline and tb- ' irs, chiefly red and yellow. The picture dc , ^ .;oti^^jjj^f a Nubian king with a peculiar crown on .1. ^^^r^^^^re four figures in this group : (1) Qod the Father, (2) our Lord on the extreme left presenting (8) the king, who -••':'^--" /- ^^-^»t seated or kneeling. On his left hand is ^^ .,_:,. j, .uUps intended to represent thej^irti^*S.«i^*'^*^^^^fiSS^^i';^«iMft^ife^ <^wo horns, noti^^^i^ kW*^\.^^;^RVtdS^r^te>'i^v^'^# \^§i^i \im^ a,nd what appear to be^^tjK-jji^. H^i* d the 1 ,r e rtaii) w; onf't'i' it buih -H, the bi.< >. if'\ li.sii.tlly are in tl l>uilding buildings, a curtain igh keep e church, ite of the we enter and it is the oasis together, t)tic way. I It up in Ite, but at je diAgram showing f h Mie Research Fund's i would be in a cross viiu t.' But '^ than a vault. Still, the »ith«x TAuitw on f». 5)7 of the ?.^\ 'i"o,.»\ V)r\^ oT 48 CHAPEL near Ahd el Kader. HALF A. Scale I : ^o. Mens, et del. C.L.W. igi2. O, Vi. ki l.fc \* >• \ r — I u .J-: V? '..V^M— ^ r^- -v » ■> "' : D To face page i6s ASSOUAN 165 harness and the general appearance of the horses are like those seen in Persian pictures. The fourth picture is a conventional representation of the Cherubim, with wings issuing from a circle ; in the centre was a representation of our Lord. The drawing of the bull's head and the human face rendered in a few deft strokes deserve notice. Among the minor subjects are several dedication crosses, and near one over the door is the legend Agios Stavros. It ia worth noticing that though there is nothing Byzantine about these paintings, the legends of the saints are all written in Greek characters. The numerous graffiti are with a few exceptions in the native Nubian language ; but an important one in Greek begins with the conventional Greek phrase, ' In Thy kingdom, Lord, remember thy servants . . .' ASSOUAN Monastery of S. Simeon This fortress monastery is typical of a class of building common to Egypt and the adjoining countries. The buildings, including a large basilican church, are surrounded by a curtain wall, and in the centre of the enclosure there is a high keep containing the dwelling-rooms, the refectory, and offices. The principal points of interest in it are the keep, the church, and the square porch or pavilion that masks the main gate of the monastery in the curtain wall on the east side. As we enter by this pavilion, I may conveniently descrihe it first, and it is one of the best examples of its class I have seen outside the oasis of Khargeh. It is made of stone rubble, roughly put together, and covered by a brick vault built in the usual Coptic way. I use the term vault because the bricks are not built up in rings or layers as they usually are in the Coptic dome, but at angles to one another, as they would be in a cross vault.* But the softness of the material, and the fact that the bricks were put in place wet, and eventually consolidated into a smooth surface, makes the roof look more like a dome than a vault. Still, the * See diagram showing the construction of the narthex vaults on p. 37 of the Byzantine Research Fund's publication on S. Eirene at Constantinople 166 EGYPT principle of construction is that of a vault, and in theory it is that adopted in building the roofs of the chapels at El Gebioui and Maatriain Tunis ; but though the principle is the same, the work was carried out in a different way owing to the difference in the materials available ; and the comparison is the more interesting, for the native architects in Egypt and Tunis of to-day retain the principle and still follow the different practices of their forefathers. The keep is built like an Egyptian temple with sloping walls, the lower half of stone, the upper part of sun-dried bricks, and a cornice at the top. The numerous rooms inside, in- cluding a long hall or gallery, cellars and storerooms, are roofed with barrel vaults made of bricks in the usual Coptic way. The church, which was a basilica, has been completely gutted, and all that remains of it are the side walls and the apses at each end of the nave. The sanctuary is at the east end, and as usual in these Coptic churches the apse is built within a square outer wall, the spaces in the angles being occupied by small rooms that represent the prothesis and the diaconicon respectively. The apse itself is square and roofed with a semi- circular semi-dome, supported on squinches.* The wall paintmgs date from the thirteenth century or later, and are of no particular merit. Our Lord seated in glory occupies the centre of the field ; His left hand holds a book of the Gospels and the right hand is raised in benediction according to the Greek way. On each side are the archangels, and beyond them two other angels. The altar stood in the usual Coptic way between a screen and the apse, and a recess in the apse wall was probably intended for the bishop's chair. At the west end of the nave is another and smaller semi- circular apse covered with a semi-dome, apparently a memorial chapel, and near it, in a cave on the north side, are a num- ber of wall paintings with diminutive figures of saints. This cave was presumably the place of interment of the founder or other notable person connected with the monastery. * The reader should study Mr. Somers Clarke's account of this church. :^)^ *^^ /iei*fe^l Kli 107 KHAh ^ Thr • often archaeolu- explqratio' worth a ' straction o! tfi> to the village ot . three parallel tow out in the middi- ji village with t graphs taken f give a general •■ In Roman tini place of exile, ar. across the Lil practically in Christian setti- ■ tombs, and judg^^^^^ community must ha afford no .- -i >-' >• and circuit whether, as some of Nestorius the Arab con part of Africa . . no difficulty in li beginning of the but the circumstai.; two provmces the ' the Arabs, the C- and still difficult • in a couir construct! I genera f anfl r Khargeh has been and an American iken a systematic tombs. It is well i y since the con- n line at Farshut i;Ted roughly in '■ hill standing they look like !>anying photo- i the cemetery was used as a ee days' march \ce escape was ■ '-c was a • i by these decoration the The tombs ; the origin M colony, or •i the followers it survived 'ity in this \ would be between the Vie seventh, le in those 1. „ KHARGEH. Oast's. Ftew of the cemetery and a chapel. Front and side Ti'rtv. To face KHAKGEH 167 KHARGEH The Christian cemetery in the oasis of Khargeh has been often visited and described by travellers, and an American archaeological expedition has now undertaken a systematic exploration of the site and excavation of the tombs. It is well worth a visit, and the journey is now quite easy since the con- struction of the oasis railway from the main line at Farshut to the village of Khargeh. The tombs are arranged roughly in three parallel rows on the spur of a limestone hill standing out in the middle of the oasis, and at a distance they look like a village with two main streets. The accompanying photo- graphs taken from the north and south ends of the cemetery give a general idea of the locality. In Roman times, as indeed to-day, the oasis was used as a place of exile, and since the Nile valley is a three days' march across the Libyan desert, it formed a prison whence escape was practically impossible. The tradition that there was a Christian settlement from early times is confirmed by these tombs, and judging by their number, size, and decoration the community must have been large and prosperous. The tombs afford no evidence in the way of dated inscriptions of the origin and circumstances of the foundation of the Christian colony, or whether, as some hold, it was at first composed of the followers of Nestorius who suffered exile there, or whether it survived the Arab conquest. Were the history of Christianity in this part of Africa the same as in Tunis and Algiers there would be no difficulty in fixing the date of these tombs between the beginning of the fourth century and the end of the seventh, but the circumstances here are quite different, for while in those two provinces the Christians were completely wiped out by the Arabs, the Copts in Egypt were able to hold their own and still continue to the present day. It is, moreover, very difficult to determine the age of primitive buildings like* these in a country where the same designs, patterns, and methods of construction were copied over and over again by succeeding generations, and the materials used were always the same and remained unaffected by the arid chmate. The general 168 EGYPT appearance of the tombs, the presence of certain classical details in construction and ornament, and the absence of certain others peculiar to the Arabs and eventually adopted by the Copts, justify a preference for an early date before the Saracen conquest in the seventh century, and that is all that can be safely said in the present state of knowledge. There is so much variety in elevation and ground plan that hardly two tombs are exactly alike. I was surprised to find that this cemetery does not contain a single example of a tomb built on the trefoil plan. The favourite design, adopted with many variations, seems to have been an octagonal cell covered with a vault, and a square hall or porch in front, so that as a general rule the more important tombs were composed of two chambers, the inner one serving for the interment of bodies placed in a small trench or pit in the floor, and the outer on^ as a porch. The tomb A is illustrated here to represent the usual combination of a porch in front and cella behind. The first two views show the south front and the west side, where the furrows or scars seen in the walls are caused by rain that falls, I believe, about once in a decade. The next is a picture of the interior of another octagonal cella, roofed by a vault ; the walls are ornamented with arch mouldings and small niches of classical design probably intended for lamps. In this particular tomb there are no wall paintings, but in some the roofs are decorated with representations of familiar subjects from the Old and New Testament. These are roughly drawn and crudely coloured in the style familiar to us from the catacombs at Rome. To this common pattern of tomb there are several exceptions. The most important situated at the north end of the cemetery is identified as the cathedral. This was probably a small monastery, and it contains a church, a hall, and a number of smaller chambers, of which some were cells and others tombs, grouped together within a curtain wall. The little plan copied from a sketch in my notebook must be taken merely to indicate the relative positions of the southern part of this group of buildings, and the places the photographs were taken from. The material used in this and all the other tombs is sun-dried bricks. The first view is taken on the south side to show the so ilr^:^^tit^. KHARGEH. Two circular tombs. To face page i68 li]s " ' ti|,p i . the piesence of certain classical dot **'S m c • ! lent, and the absence of certain of peci;. tnd eventually adopted by the Cop- justify A pr**^ 'f an early date before the Saracen conquest in the s> tutury, and that is all that can be safely said in th "te of knowledge. J; : .. h variety in elevation and ground plan that hardly two tombs are exactly alike. I was surprised to find that ry does not contain a single example of a tomb })■'■' i^iefoil plan. The favourite design, adopted with ! >U8, seems to have been an octagonal cell covered with and a square hall or porch in front, so that as a ; the more important tombs were composed of two ihe inner one serving for the interment of bodies : a stnall trench or pit in the floor, and the outer one i i.b A is illustrated here to represent the usual '• '> of a porch in front and cella behind. The first liow the south front and the west side, where the )r scars seen in the walls are caused by rain that falls, , about once in (B^?i\ ^"?,,'ft\ ^:iv\ d\. "itri'riiiiii'ir*7"' Mlifjiy^tl ^••■si.-^^. UUi facade of tti- supporting ! bay there w»^r» i'.in, and i"n u! angular oii. bases and little < to represent acu nlinth to r raised Oivy^ < two doors was The door in tltt -'-r south Wal through r hall and adjoinm*^ '■'•'od colonTi'id*^ ■ r au r t ^tuail niches a>ov« begin with the church wiuc-ii wji«i therefore paralle In length it corresptu and the square cast wall a' square east wal body of the chn according to th side chamber arjrai churches 'like T* I ' the! :-> no gi- found ii; that th cam froni the ' la • not att«> ing witl In it A !:«. » nave by thn four round i! 1 and siood r-rithe i.J .Miiai ^^ .i\i>Yi. t>nsiian flAito«}|^ ■ • hf^s of flAJJ!' ' i l-'aT iih shjows . tht ni are ■ ..uitiiyLuh .sV:itM*0 .iwft\\ Viu\>, ,\^»^ A"^"^\ V?,'i^ .^wo^ \»"\Vi%A^y>;i iX^-^^a^'A 51 Hall. Church. Khargeh cathedral tomb. West front, Hall, and plans. To face page i6g KHAKGEH 169 facade of the curtain wall decorated with an engaged colonnade supporting ten round-headed arches. In the first and sixth bay there were round-headed doors with small niches above them, and in the others small rectangular apertures above and triangular ones below for ventilation ; the pillars have square bases and little caps decorated with what is no doubt intended to represent acanthus leaves. This curtain wall stands on a plinth to correspond with the floor level of the interior raised above the ground-level outside. The approach to the two doors was by small flights of steps and little balusters. The door in the first bay leads to the church by a court (the south wall of the church is visible in the picture look- ing through the door), the door in the middle leads to the hall and adjoining chambers. I may conveniently begin with the church which was correctly orientated, and stood therefore parallel with the curtain wall shown in my picture. In length it corresponded with five bays of the arcading and the square east wall abutted on to the hall. Within the square east wall was a semi-circular apse built out into the body of the church, and flanked by two small side chambers according to the usual African system. The central apse and side chamber arrangement is found in the earliest Christian churches like Tebessa, and in the much later Coptic churches of the tenth or eleventh centuries in Northern Nubia ; it is therefore no guide to date. The apse contains three little niches also found in the Nubian churches. The church at Faras shows that these niches were used for lamps. Above them are three loop-headed crosses that arc certainly Egyptian. These crosses are of very early pattern, but as I have said before cannot be relied upon to fix a date. As the hall is shut off from the church, we must go outside again and re-enter the south front by the middle door. Passing through the vestibule we enter the hall. I shall not attempt to describe this peculiar and lopsided build- ing without referring to the accompanying pictures and plan. In it A is the nave, B is the north aisle, separated from the nave by three pillars with square caps and bases supporting four round arches, C an apsed recess * on the south side * See page 127. 170 EGYPT covered by a semi-dome decorated with paintings, J) the aisle on the south side. The shading on the plan indicates the barrel vaulting, and the cross shading the little dome or cross vault. E and F are the two entrances which are not opposite one another, and G is a sepulchral room covered like the nave partly with a barrel vault and with a small dome or cross vault. The small letters indicate the places the pictures are taken from, showing the east and north sides of the nave and part of the barrel and cross vault. The dimen* sions are approximately 10 metres long by 7 broad.* The other notable exceptions to the common plan are two circular tombs ; they are about the same size and measure about 3 metres in diameter and 2 metres from the ground to the spring of the dome. There is not much to be said about them beyond noticing that originally the sides were walled up and light and air admitted by little triangular apertures. The interment was effected in a shallow pit in the centre of the floor. The other photographs sufficiently explain themselves ; among them are a half-dome built of brick and a tomb with a porch that might be Roman. * The cathedral and the smaller building like it deserve more careful study than I was able to give to them. And I am not sure that my plan is quite correct. n-i- \,»i s fe«ir: H KHARGEH. Chapel^ totnhs and transept. Apse in cathedral towh with paintings on the half dovie. To face page ijo 170 EGYPT covered by a serai-dorae decorated with paintings, ]J th* on the south aide The shading on the plan indicatt ^Murrel vaolting, and the cross shading the little dome jr it, >j- , Mil It. t'' and F are the two entrances which are not lother, and G sepulchral room covered Jik • partly with a barrel vault and with a small vault. The small letters indicate the places ii<.c [■•iwi'^ii are taken from, 8ho^^^ng the east and north sides of the a»ve and part of the barrel and cross vault. The dimen- e approximately 10 metres long by 7 broad.* 'her notable exceptions to the common plan are twu libs ; they are about the same size and measure about I diameter and 2 metres from the ground to the spring There is not much to be said about them beyonti icu originally the sides were walled up and light and led by little triangular apertures. The interment '.^ted in a shallow pit in the centre of the floor. i i;<.^ other photographs sufficiently explain themselves ; among !'♦■;;' are a half-dome built of brick and a tomb with a porch r»>*r unaht he Koi^-^^^^-^^ • •^''^^'HHsiV^AK^^C'i'^^W^^M'iS^it deserve more careful study tiia:. . to give tb tnom. And I am not, crVi tnat my plan is qmto correct. 0\\ ^■^ft\ ^^vv\ o\ rFi ^, y. ■".^•^v'''^*>s> t tz The fc; by Count i notes on li in identifying i locality, and t! church wli'"' S. N'icolo .' rebuilt by Abl- ■ cr the ' Tibi Ir nun ^ «orae difficulty doubt as to the ' r the ruined ID chapel of hnrch i Placa .r\«.e^%^^:> m i\svsN<.\^ \»^«v.-.V«\:>^K''fvam vigesimam '■'■ andi templum ! residentiam (le, ad istius ( terras multas, et crista illin- oncf "- 'f usque ad Furnari II gpeluncam, ■ ) et ascendit iieridiei et et sicut luanifesta I h; \\\ '^■^ft\ ^-jsiy vn\ 53 KHARGEH. Architectural details in chapels. To face page lyi APPENDIX The following is a transcript of parts of the charter granted by Count Eoger to the Abbot Chremetes, referred to in the notes on the church at Castiglione. There is some difficulty in identifying the parcels, but there seems no doubt as to the locality, and the church of S. Domenico is either the ruined church which was replaced by the little Norman chapel of S. Nicolo near the Castiglione bridge, or the church to be rebuilt by Abbot Chremetes. Roger the Count grants to the Abbot Chremetes de Placa as follows : — * Tibi tradidi ad insulam istam talem in parvam vigesimam numerationem hominum Castrileonis, causa sublevandi teraplum Salvatoris ac ipsam reedificandi et ad monachorum residentiam ordinandi, velut ac reliquae ad insulam talem, rende, ad istius Monasterii servitium et consuetudinem divisi terras multas, nemora, loca deserta et campos sicut incipiunt.' Then follow the parcels. ' Ubi ascendit flumen magnum et conjungit usque ad Furnari fluviculum ac redit flumen ipsum usque ad magnam speluncam, et illinc usque ad collem magnum dictum Sterio et ascendit crista ipsius magni collis et Polemum (?) ex parte meridiei et illinc descendit flumen ipsum usque ad fluviculum et sicut ascendit monasterii crista et concludit deinde per manifesta concessione ipsius Monasterii.' I have not attempted to precise these boundaries, but assum- 172 APPENDIX ing that the two rivers are branches of the Alcantara, the terri- tory referred to seems to be that on the south of Francavilla and west of Castiglione where this church and the little Norman chapel I have referred to are situated. The grant proceeds : ' Inde principium confinium egimus per omnia concludentia in confine isto tali habeat monasterium de Placa absque uUo impedimento et rixa ; et cum ipso redidi ad ipsius monasterii servitium Agarenos Tauromenitos (i.e., the Saracens) quatuor cum eorum uxoribus et filiis .... (then follow the names) .... tali propter rationem reedificandi ac potentiabiliter sollicitato .... Similiter affirmo isto tali monasterio insula Sancti Stephani que est sub Tauromenio, cum omni regimine et navigabili lignum unum sol quod non sit a Mauris aliquibus consuetudine impeditum .... Messane hac prima prescripta Anno mundi sexmille sexcentum addito uno.' 1092. Compare this grant with the Girgenti charter referred to in Vol. I. p. 41. These Arabs no doubt were bondsmen. Garufi, I Dociimenti inediti etc, 1 n, 2, p. 7 and following. M.S. Bib : Com : Palermo Q 9, 49, f. Ill old numeration, 51 modem. I had intended to illustrate here a few examples of marble carved capitals and of a further series of gold coins in our collection from Leo VI to John Zimisces. But this volume is already quite long enough, and so I shall keep them for the third, which I foresee will be required to illustrate some more cellae trichorae at Querqueville, Spaccaforno (Val d'Ispica), Mon- majour and elsewhere, and the Byzantine churches of Southern Calabria, including Pentedattilo and S. Maria di Tridetti near Brancaleone. INDEX. ABD EL KADER, Chapel, 154, 162. Abbots : Aldebert, Lerins, 26 ; Aygulph, Lerins, 25, 26 ; Chremetes, S. Dome- nico, Castiglione, 52, 171 ; Maxi- mus, 9 ; Roffredo, 30 ; Teodemero Monte Cassino, 30, 82 ; Theoste- krites, Agro, 57. Abu Seir, 22, 154, 162-164. Abu Si Sifain, church at Cairo, 87. Abyssinia, 82, 158. Addendan, 15, 82, 153, 155. Adrian VI of Rome, 25. AFRicA,province8 of, 6. Arabs in, 9, 12, 13, 14, 48, 64, 85, 151, 165. Berbers in, 4, 7, 8, 45, 47, 85, 90, 120, 131. Byzantines in, 6, 8, 9, 13, 85, 86, 87, 91, 93. Romans in, 6, 7, 8,47, 48, 54, 90, 91, 109, 110, 113, 114, 115, 117, 127, 131, 136, 138, 139, 166. Vandals in, 7, 8, 9, 20, 47, 48, 85, 86, 90, 91, 103, 108, 116,117, 124, 131, 138, 141, 146, 152. The Church in, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 69, 85, 86. Christian architecture in, 8, 30, 31, 32, 85-152. Secular architecture in, 90. Agricultural, 111, 112. Baths, q.v. Civic (arches), 102, 108, 109, 117, 129, 131, 139, 140, 148. Military, 91, 108, 127, 132, 136, 138, 141. Pagan temples, 45, 59, 89, 90, 106, 126, 138. Stables, 107, 115, 125, 133, 134. Africa Proconsularis, 6, 85-152. Agro, 52, 54-58. Ain Hallouf, 141, 148. Ain Tounga, 106, 107, 137, 139, 151. Albenga, 16, 24, 93. Alcantara, River, 49, 50, 171, 172. Aldebert, Abbot of Lerins, 26. Alexandria, 13. 28, 30, 31, 32, 41, 47, 95, 98, 120. Algeria, 6, 12, 14, 32, 40, 41, 45, 47, 48, 91, 96, 100, 101, 116-130, 156, 166. Altars, 14, 15, 19, 22, 28, 39, 40, 88, 89, 98, 99, 107, 129, 137, 147, 150, 153, 155, 156, 157, 160. Coptic altar furniture, 160, 161, 165. See also Iconostasis, Baldacehino, and Cur- tains. Altar screens and posts, 8, 14, 78, 88, 98, 99, 118, 120, 121, 137, 158, 159, 160, 165. Altavilla, 59. Amalfi, 68, 70, 79, 80, 81. Amari, 79. Amico, Dictionary of topography, Sicily, 52. Ammadaera, see Haidra. Amr, Mosque of, in Cairo, 41. Anastasius, author, on Sorrento, 79. Anatolius, Duke of Gaeta, 68. Announa, 15, 35, 89, 91, 97, 99, 101, 126, 127, 153, 155. Antioch, 13. Aphrodisium, 141, 148. Appia, Via, 148. Apses, Tribunes, and their ritual sig- nificance, 2, 15, 39, 40, 45, 50-54, 56, 57, 58, 61 ; 87, 88, and Vol. I p. ix. On the triple apse at Con- stantinople, 99, 121, 126 ; 133, 142, 155, 156. Apulia, 70, 80. Arabs : Conquest of Africa, 9, 1^, 13, 14, 31, 48, 64, 85, 151, 165. 174 INDEX Arabs (continued) — in Sicily, 3, 53, 58, 70, 71, 172. in Italy and Provence, 25, 26, 68, 70, 71, 78, 79. Arians, 7, 86. Aries, 20, 24, 25, 26. Armenian, 8, 47, 88, 87, 95. Asia Minor, 9. Aspromonte, 3. Assouan, 154, 164, 165, 170. Athanasius, bishop of Naples, 79. Athos, Mount, 20. Atria, 98, 137. Augusta, Sicily, 55. Aygulph, Abbot of Lerins, 25, 26. BADIAZZA, 58. Bagnara, 4. Balata, 17. Baldacchino, 14, 15, 88, 106. Ballu, 125. Baptismal rite, 16, 24, 92, 93, 94. Baptisteries and fonts, 15, 17, 32, 33, 40, 45, 92, 93, 94, 95, 97, 102, 107, 119. 126, 143, 147, 151, 152,160, 161. Bardo Museum, 14, 88, 94, 147, 151. Bari, 71. Baronius, Eccl. Hist., 79. Barreca, Canon, 62, 64. Basilian Monastic order, 49-59, 52, 54, 58, 59. Basta Bey, 159. Baths, 32, 33; Ain Tounga, 107 Hammam Meskoutine, 127; Lam bessa, 32, 54 ; S. Menas, 17, 31, 32 Thelepta, 31, 32, 54. Beja, 138. Belisarius, 85, 87. Bellator, bishop of Sbeitla, 103, 102- 105. Beneventum, 70. Berbers, 4, 7, 8, 45, 47, 85, 90, 120, 131, Bethlehem, 17, 33, 46, 47, 54. Bir bou Rekba, 88, 94, 97, 99, 100, 147, 148, 151. Bishops, 11, 64 : — Baleriolus, Uppena, 146; Bellator, Sidi Abich, 148; Jucundus, Sheitla, 105 ; Honorius, Uppena, 146 ; Palladius, Tebessa, 124 ; Shenouda, Sohag, 45, 46, 48 ; Zosiixius, Syracuse, 59. Bishops of Rome : — Adrian VI, 25 Callixtus II, 25 ; Eugenius III, 25 ; Gregory the Great, 64 ; Gregory II, 68 ; Hadrian I, 64, 74, 81 ; John VIII, 11, 25, 71, 80, 81 ; John X, 71; Johh XII, 68, 69, 71; John XIV, 81; John XV, 80, 81; Innocent III, 30; Martin, 9, 11; Nicholas I, 81 ; Paul, 68, 69. Bishops' seats and thrones. See Apses. Bone, 127. Boricaud, Dr., 114. Bou Ficha, 148. Brancaleone, 172 Brick, use of, 50, 54-58 ; pottery tubes, 100, 112, 122; vaulting, 154, 157, 158, 159, 160. See Assouan and Kargeh. British occupation of Sicily, 55. Brolo, Lancia di. Archbishop of Monreale, 11. Busaidone, river in Sicily, 19. Butcher, Mrs., 6. Butler, A. J., 6, 159, 160. Byzacena, province, 6. Byzantine Period : in Africa, 6, 8, 9, 13, 85, 86, 87, 91, 93. in Calabria, 81. in Italy and Sicily, 29, 68 ; Navy, 70, 71. in Sardinia, 11. Capitals, 47, 108. Liturgy, 14, 54, 61, 88, 121. Churches in : Italy, 2, 54, 67, 76, 82. SicUy, 2, 27, 49-54, 58. Constantinople, 15, 17, 38, 89, 99, 121. Jerusalem, 33, 125. Salonica, 2, 16, 17, 52. CAGLIARI, 23. Cairo, 14, 15, 41, 47, 87, 88, 99, 100, 159, 169, 170, 171. Caius Durmio Quadratus, 30. Calabria, 3, 4, 11, 12, 19, 54, 55, 67, 69, 70, 80, 81, 82, 156, 172. Calhxtus, Cemetery of, 26-28, 125. Camera, Historian of Amalfi, 79. Camerina, 28. Cannes in Provence, 1, 20. Capitals, 2, 128. Cappella Palatina, Palermo, 67, 98, 118. Capri, 75-81. Capua, 70, 75. Caput Vada, 2, 141. Caracalla, the Emperor, 90, 102, 117, 170. Carthage, 6, 9, 15, 31, 47, 64, 85, 93, INDEX 175 94, 97, 98, 101, 121, 126, 127, 131, 141, 144 ; Basilica of Justinian, 149-151 ; Damus el Karita, 151, 152, 156. Casa Inglese, Etna, 55. Caserta, 76. Cassino, INIonte, 20, 25, 28, 29, 30, 32, 78, 82-84, 155, 158. Cassiodorus, 19. Castelvetrano, 27. Castiglione, 2, 49-54, 67, 171, 172. Catania, 2, 49, Catalani, church of the, 58. Catanzaro, 19. Cavaignac, General, 55. Cecilia, Metella, tomb, 148. Cefalu, 57 CeUae Trichor*, 1, 2, 17, 19-48. Origin of, 32, 33, 54, 125, 172. Cham and Ilea, 4. Chambi, 110. Charlemagne, the Eiuperor, 25, 26, 52, 70. Chremes, Abbot, 52. Chiocarelli, 79. Church. Africa, North, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 85-101, Anglican, 16, 87. Arian, 7, 86. Armenian. 8, 87. Donatist, 7, 85, 86. Egypt, 6, 11, 14, 15, 16, 40, 45, 80, 86, 87, 91, 93, 98, 99, 100, 153, 159, 160. Greek ; and the Constantinople Patri- archate, 2, 11, 12, 50, 52, 54, 61, 67, 69, 81, 86, 87, 121. Italy and Sicily, 10, 11, 15, 52, 53, 61, 79, 80, 81. Jacobite, 6, 13. Latin Church in North Africa, 6, 9, 10, 11, 12, 86, 87. Liturgical, 7, 87, 88, 89, 121, 158. Melkite. 13. Patrimonies, 7, 68, 69, 70, 71, 80, 81. Peace of Constantine, 6, 9, 85. Civvita Vecchia, 70. Clarinval, 122, 124. Clarke, Somers, 6, 154, 159, 165. Codex Cassinensis, 72, 78. Coelestis, 90. Constantine I, the Great, Emperor, 34, 89, 90, 116, 148, 152. Constans II (Constantine III), 9, 10, 11,69, 61. Constantine IV (Pogonatus), 9, 10, 11 Constantine Porphyrogennetos, 67, 68 Constantine, Algeria, 120, 122. Constantinople, 2, 8, 9, 12, 13, 15, 17, 38, 53, 61, 68, 69, 80, 88, 89, 92, 94, 121, 169. Coppanello, 19. Coptic art and ornaments, 31, 34, 95. Coptic Church, 6, 11, 13, 14, 15, 16, 88, 98/99, 100. Coptic liturgy and ritual, 14, 15 ; baptism, 16, 40, 45, 91, 93, 153, 159, 160. Corinthian capitals, 126. Corsica, 55, 66. Councils of the Church in Africa: at Aries, 20, 24 ; at Constantinople, 12; at Rome (the Lateran), 11, 68. Crescentius family, 81. Crocifisso, chapel of the, at Monte Cassino, 28. Crosses and monograms, 23, 39, 40, 46, 95, 128, 146, 148, 161, 164. Crowfoot, 45, 82. Crusaders, 33 ; round churches, 125. Curtains and altar screens, 87, 88, 98, 99. Cybele, 4. Cyi'ene, 6. DAMUS EL KARITA, Carthage, 47, 93, 94, 97, 148, 151, 152. Debereh, 15, 99, 154, 156. Decius, the Emperor, 144. Delattre, Pere, 152. Delia, 27, 52, 58. Deo Laudes, 86. Department of Antiquities, Tunis, 6, 102. Dernaia Pass, 110. Diaconica, 50, 54, 56, 61, 87, 88, 89, 119, 120, 121, 126, 128. Dickie, A. C, 33. Diehl,M. Ch., 6,9. Dimarzu, Mons, 52. Domes : Beehixe shaped, 29, 158, 160. On pendentives, 2, 22, 23, 31, 50, 51, 76, 78, 157. On sijuinches, 31, 37, 41, 44, 57, 58, 73, 78, 117, 118, 157, 160; in clusters, 56. Donatists, 7, 85, 86. 176 INDEX Doric architecture, 59. Dougga, 91, 97, 99, 101, 121, 136, 137, 138, 151. Duprat, M., 128, 124. EARTHQUAKE at Keggio and Messina (1908), 3, 4, 5, 19; in Sicily (1140 and 1169), 61, 65. Ecthesis of Heraclius, 10, 13. Edrisi, 79. Egypt, 1, 6, 12, 13, 14, 15, 17, 20, 31, 34, 48, 75, 89, 96, 100, 116, 121, 127, 151-169. El Aroussa, 138. El Gebioui, 31, 100, 112, 122, 164. El Kef, 30, 65, 89, 91, 97, 98, 100, 101, 117, 133, 136, Emilia, Duchess of Gaeta, 72. Emperoks and Empresses. Caracalla, 90, 102, 117, 170. Charlemagne, 25, 26, 70. Constans II (Constantine III), 9, 10, 11, 59, 61. Constantine the Great, 34, 89, 90, 116, 148, 152. Constantine IV (Pogonatus),9, 10, 11. Constantine Porphyrogennetos, 67, 68. Decius, 144. Eudocia, 39. Helena, 34. Heraclius, 8, 10, 13, 102. Isaac Komnenos, 53. Julia Domna, 117. Justinian, 7, 8, 13, 14, 34, 47, 85, 87, 88, 90, 91, 92, 93, 94, 97, 102, 108, 121, 126, 141, 148. Leo III, 12, 52, 64, 68, 69, 80. Louis II, 79, 80. Mahomet II, 13. Napoleon, 55. Nicephorus, 80, 81. Otto II, 71, 72, 80. Otto III, 71, 72. Septimius, Severus, 32, 90, 102, 117, 131, 136. Theodora, 13, 34. Theodosius, 48, 89, 138. Theophano, 72. Theophilus, 70. William II of Germany, 3. Enfidaville, 108, 141, 145, 148. Ephesus, 143. Epiphany tanks, 15, 92, 93, 119, 160. Erivani, 83. Esneh, 171. Ethiopia, 82. Etna, 50. Etruscan, 30, 55. Eudocia, the Empress, 39. Eugenius III, of Rome, 25. FABIO GIORDANO, 79. Faras, 41, 87, 99, 153, 154, 156 Faro, 4. Farshut, 166. Favara, the Castle, Palermo, 2, 53. Felix, the Governor of Caesarea, 30. Feriana, 101, 112, 114. Flavius Secundus, 110. Frescoes and wall paintings, 39 ; at Abu Seir, 162, 163, 164, 165, 167. Fondi, 71, 80. Fonts. See Baptisteries. Formia, 72. Forts, 132, 136, 140. Forzia d'Agro, 55. Francavilla, 49, 52, 172. Freccia, author, on Amalfi and Capri, 79. Frejus, 16, 24, 93. French occupation of Africa, 6 ; of Calabria, 55. GAETA, 29, 68-75, 78. Gaetani d'Aragona, 67. Gaflfour, 138. Gafsa, 31, 100, 108. Galla Placidia, 28, 29, 32. Gargano Mte, 72. Garigliano, River, 71. Gavino, S., 5, 100. Gebioui, chapel, 31, 100. Gelimer, 85, 91. Genoa, 25. Genseric 47, 85, 90, 91. German legion in Sicily, 55. Germany, the Emperor William of, 3. Ghardimaou, 100 Giammellaro, 55. Giampilieri, 55. Giants of Messina, 3, 4. Giardini, Taormina, 49. INDEX 177 Gog and Magog, 3, 4. Grasse, 25. Gregory the Viceroy at Sbeitla, 9, 102. Gregory I the Great, of Rome, 64. Gregory II, of Rome, 68. Greek Communities : in Africa, 6, 8, 9, 11, 12, 13, 14, 15, 87, 88. in Italy and in Sicily, 2, 70, 80, 81. Liturgy, 2, 50, 54, 61, 67, 121. Griffith, Prof., 18, 154, 162. Gsell, M., 32, 90, 95, 96, 98, 116, 123, 124, 126. Guelma, 126. Guerin, 110, 139. Guildford, 95, 106. Guiscard, 59, 95. HADGEB EL AIOUN, 150. Hadrian of Rome, 64, 70, 81. Haidra, 91, 97, 101, 107, 109, 115, 117, 120, 122, 125, 131 to 136. Haikals. See Apses. Hammamet, 114, 147. Hamman Meskoutine, 127. Hanson, Captain, 55. Haouch Khima Mta Darrouia, 110, 111. Haouch Taacha, 111. Hauteville family, 75. Helena, Empress, 34. Henchir, Chegarnia, 141. Henchir Choud el Battal, 111. Henchir el Baroud, 150. Henchir Fradis, 141, 148. Henchir Gebeul, 30, 109, 113. Henchir Maatria, 31, 122, 138, 164. Herachus, the Emperor, 8, 10, 13, 102. HoUzinger, 26. Holy Trinity Chapel. See Lerins and Cellae Trichore. Hunneric, 124, 146. ICONOCLASM, 68, 69. Iconostasis or screen, 14, 15, 16, 40, 87, 88, 89, 99, 107, 120, 161. Innocent III, of Rome, 30. Inscriptions : Arabic, 45. Berber, 120. Coptic, 39, 44, 45. Greek, 57, 95, 96, 133, 134. Latin, 64, 91, 95, 96, 105, 129, 130, 144, 145, 146. Nubian, 154, 162. Istria, 89. Italy, 68-84. JACOBITES, 6, 13. Jerusalem, the Holy Sepulchre, 33, 125. John, Duke of Gaeta, 68, 71, 72, 75, 78. John VIII, of Rome, 11, 25, 71, 80, 81. John X, of Rome, 71. John XII, of Rome, 68, 69, 71. John XIV, of Rome, 81. John XV, of Rome, 80, 81. Julia Domna, 117. Justinian, the Emperor, 7, 8, 13, 14, 34, 47, 86, 87, 88, 90, 91, 92, 93, 94, 97, 102, 108, 121, 126, 127, 141, 148. KAIROUAN, 109, 170. Kalaa Djerda, 131. Kamula. See Nakadeh. Kasserine, 109, 110. Khargeh, 13, 17, 127, 166-169. Kings, Queens, Princes, Dukes, and Governors. See also Emperors : Anatolius, of Gaeta, 68. Elena, H.M., of Italy, 5. Emilia, Duchess of Gaeta, 72. Gelimer, 85. Genseric, 85, 90. Gregory of Sbeitla, 9, 102. Guiscard, 59. Hunneric, 124, 146. John, of Gaeta, 68, 71, 72, 75, 78. Marino, of Gaeta, 72. Murat, of Naples, 55. Pandolph, of Capua, 75. Pepin, 25, 26, 69. Rameses, 161. Riddell, Geoffrey, 75. Roger the Great Count, of Sicily, 52, 59, 75, 171. Tancred, 75. William, Count of Provence, 26. Kings wood Church, 143. LACONICA, 125. Lambessa, 32, 54, 109, 116, 125, 127. Latin Church in Africa, 6, 9, 10, 11, 12, 86, 87. Latin Language and Inscriptions : in Africa, 7, 8, 9, 11, 64, 95, 96. in Sicily, 64. Lattakia, 117. 178 INDEX Lecce, 74, 75. Lemaitre, Bishop, 146. Leo d'Ostia, 83. Leprignano, 99. Lerins Islands, 1, 20, 21, 26, 47. Liturgical, 7, 87, 88, 89, 121, 153. Lombards, 68, 70, 72, 74. MAATRIA HENCHIR, 31, 122, 188. Maccari, 27, 31. Mahomet II, the Sultan, 13. Mahometans. See Arabs. Mainz, 100. Malvagna, 31, 49, 54. Mangoni, 79. Marbles, 100, 147. Marcella, 124. Marcellus, Tower of, 55. Martin, of Rome, 9, 11. Martorana, 67, 170. Masons' marks, 23, 120. Mata and Griffone, 3. Mauretania, 6. Maximus, Abbot, 9. Mazzara, 53. Medinet Habu, 154, 159. Mehedia, 169. Melkites, 13. Merimee, Prosper, 24. Merlin, M., 6, 102, 103, 104, 105, 139. Messina, 2, 4, 19, 55, 58, 59, 70, 71, 172. Metlaui, 108. Mileham and Maciver, 12, 154-8. Mojo, 49. Momnajour, 17, 26, 47, 172. Monograms, XP. See Crosses. Monophysites, 10, 13. Monothelites, 9, 10, 13. Monreale, 57. See also Brolo. Moris, Henri, 21, 24, 25. Mosaics, 14, 88, 91, 93, 99, 100, 101, 103, 107, 125, 127, 142, 145, 146. Pavements : Carthage, 149-151 ; Sbeitla, 103; Uppena, 145, 146. Tombs : Tebessa, 124 ; Enfidaville, 142, 145, 146. Mosques : Amr at Cairo, 41 ; Hodja, Moustapha (Balata) Constantino- ple, 17 ; Toklou Dede at Constanti- nople, 253 ; Kairouan, 169. Murat, King of Naples, 55. Muratori, 79. Museums : Alexandria, 28, 30, 32, 95, 98, 120; Bardo, Tunis, 14, 88, 94, 147, 151; Carthage, 152; Kirche- rian, Rome, 24 ; Sfax, 146 ; Syra- cuse, 19 ; Tebessa, 95, 126. NAKADEH, 154, 159, 160, 161. Naples, 4, 25, 68, 69, 70, 71, 72, 75, 78, 79. Narthex, 42, 55, 56, 76, 160. Napoleon, the Emperor, 55. Nea Taktika, 90, 91. Nebrodian Hills, 4. Nestorius, 13. Nicephorus Phocas, 80. Nicolosi, 55. Nil, S., 72. Nile Valley, 12, 14, 15, 20, 31, 34, 95, 151-169. NOBMANS : in Sicily, 2,44, 49, 52, 53, 65, 67. in South Italy, 2, 67-84. Architecture, examples of, 2, 51, 52, 57, 58, 59, 61, 62, 65, 67-84, 95, 170. Law and Charters, 52, 53, 171. Nubia, 12, 15, 22, 27, 41, 82, 87, 99, 121, 153, 154-164. Numidia, 6, 7, 76, 86. OIL PRESSES, 111. Optat, S., 86. Orientation of churches, 96, 97 Orsi, P., 18, 19, 49. Ostia, 70. Ostia, Leo d', 83. Otranto, 70. Otto II and Otto III, Emperors, 71, 72. Oued Djilma, 150. PAGAN : Deities: Cham, 4; Coelestis, 90, 138 Cybele, 4, 138 ; Jupiter, 90, 138 Minerva, 90 ; Neptune, 3 ; Rea, 4 Saturn, 4 ; Zanclos, 4. Temples, 48, 59, 89, 90, 106, 126, 138. Tombs, 136, 148, 161. Paintings on walls. See Frescoes. Palermo, 2, 44, 52, 53, 57, 59, 60, 67, 70, 73, 98, 118. Palestine, 33. See also Bethlehem, Jerusalem. Pandolph, of Capua, 75. Patriarchate of Constantinople, 11, 12, 61, 69, 81. Patrimonies of the Church, 7, 68, 69, 70, 80, 81. Paul of Rome, 68. Pavia, 81. Peace of the Church, 6, 9, 85. Pentapolis, the, 9. Pentedattilo, 172. Pepin, King, 26. Photius, Patriarch, of Constantinople, 81. INDEX 179 Pisa, 3. Placa, 52, 171, 172. Pliny, 30. Pont de Trajan, 138. Pontecorvo, 72. Porphyry marble, 100. Portotorres, 100. Priolo, 55. Procopius, 141. Prothesis, chapels of the, 50, 54, 56, 87, 88, 89, 119, 120, 126, 127, 128. Provence, 26, 47. Pulpits, 41. QUAKERS, the, 7. Quarries, 113. Querqueville, 172. Quodvultdeus, 124. RANDAZZO, 49, 50. Raoul, 145. Ras Capoudia. Ravenna, 28, 29, 32, 100. Rea, 4. Reggio, 5, 19, 55. Riddell, Geoffrey de. Governor of Gaeta, 75. Roccelletta, Calabria, 19, 156. Roffredo, of Monte Cassino, 30. Roger the Count, 59, 75, 171. Rome ; I. Classical Period. In relation to the occupation, and civil and military architecture of North Africa. See Africa. On deities, sepulchres, and temples. See Pagans. II. Christian and Mediceval Periods. In relation to the Byzantines at Con- stantinople and the Greek Settle- ments and Church in Italy and Sicily, 9, 10, 11, 68, 69, 70, 71, 72, 80, 81. Churches in, 22, 27, 28, 64, 74, 88, 89, 98, 99, 100, 121, 125, 129, 146, 150. Christianity in Africa, 7, 11, 12, 87. Language and inscriptions. See Latins. Latin Church. See Latin. Patrimonies, 7, 68, 69, 70, 80, 81. Pontiffs. See Bishops. Saracens. See Arabs. Sovereigns. See Emperors. Rometta, 53. Rossano, 30, 67, 76, 82, Round churches, 33, 52. Russia, ships of war at Messina, 4, 5. SACCARGIA, Sardinia, 1, 23. Saints and Dedications. Abu-s-Sifain, church at Cairo, 87. Agnese, church at Rome, 98, 100. Anba Shenouda, church at Cairo, 45. Andrew, church at Trani, 74. Annunziata dei Catalani, church at Messina, 2, 50, 58. Appollinare, church at Ravenna, 100. Benedict, rule of, 25. Callixtus, cemetery at Rome, 1, 22, 26-28, 125. Cattolica, church at Stilo, 76, 82. Celsa, church at Ravenna, 29. Clemente, church at Rome, 89. Costanzo, church at Capri, 75-81. Croix, chapel at Monmajour Aries, 17, 47. Domenico, church at Castiglione in Sicily, 2, 49-54. Eirene, church at Constantinople, 89, 99, 121. Elias, church at Salonica, 2, 16, 17, 52. Gavino, church at Porto Torres, 100. George, at Medinet Habu, 161. Germano, town of, 28. Giovanni al Mare, church at Gaeta, 68-75, 76, 78, 146. Giovanni degli Eremiti, church at Palermo, 53. Giovanni in Sinis, in Sardinia, 1, 22, 23. Giuseppe, church at Gaeta, 72. Holy Trinity, chapel at Lerins, 1. Honorat, monastery in Provence, 1, 17, 20, 26, 31, 47. Joseph, monastery at Tibar, 139. Lorenzo, church at Syracuse, 64, 66. Louis, chapel at Carthage, 93, 148. Marcian, chapel at Syracuse, 62-66. Maria, church ' delle cinque torri ' at St. Germano, 78, 82-84, 155, 158. Maria in Cosmedin, Rome, 64, 74, 88, 98, 99, 121, 146, 150. Maria la Porta, via in Messina, 5. 180 INDEX Maria Theotokos, cathedral at Syracuse, 59. Maria, Tridetti, 172. Marguerite, island at Cannes, 20, 21. Mark, church at Rossano, 67, 76, 82. Mark, Venice, 121. Mary, the Blessed Virgin, 3, 19. Menas, Monastery of, in Egypt, 17, 32, 47. Michael, Monastery of, at Kamula near Nakadeh, 159. Miniato, church at Florence, 99. Nazaire, church at Eavenna, 29. Nicholas and Cataldo, church at Lecce and Bari, 30, 74. Nicolo, chapel at Castiglione, 171. Nil, of Rossano, 72. Optat, 86. Pancras, chapel in Sicily, 19. Patrick, 20. Paul's Cathedral, 121. Perpetua, Martyr, 148. Peter, chapel at St. Germano, 30. Peter and Paul, church at Agro, 52, 54-58. Sabina, church at Rome, 74, 129, 146. Salvatore de Placa, 52, 171. Sarbana, chapel in Sardinia, 29. Saturnino, church at Cagliari, 23. Sauveur, chapel at Lerins, 20, 24. Shenouda, church at Cairo and founder of the monasteries at Sohag, 45. Simeon, monastery at Assouan, 164, 165. Sofia, at Constantinople, 15, 38, 88, 99. Spirito, church at Palermo, 44, 60. Stephen, island near Taormina, 172, Thecla, church at Constantinople, 2, 53. Theodore, church at Medinet Habu, 161. Theresa, chapel near Syracuse, 1, 19, 20, 31. Theresa la Riva, Sicily, 54. Trinity, church at Delia. See Delia, 27, 52, 58. Vincent, of Lerins, 20. Saladin, M., 6, 17, 110, 111, 112, 114, 135. Salerno, 70, 81. Salonica, 2, 51, 52. Saracens. See Arabs. Sardinia, 1, 11, 22, 23, 29, 68, 81, 100. Sarum Use, 121. Sassari, 23. Saturn, 4. Sbeitla or Sufetula, 9, 91, 92, 95, 97, 100, 101, 102-108, 109, 126, 135, 139, 150, 161. Schulz, 82, 83. Screens, 14, 15, 16. Septimius Severus, the Emperor, 32, 90, 102, 117, 131, 136. Seriziat, Commandant, 122, 124, 125. Seroux d'Agincourt, 29. Serreh, 27, 31, 153, 157, 158. Sfax, 95, 108, 146. Shenouda, 45, 46, 48. Sicca Veneria. See El Kef, 136. SicUy, 2, 11, 12, 19, 20, 27, 31, 49-66, 80. Sidi Abich, 88, 91, 94, 97, 101, 145, 146. Silanus, 29. Sinis, 1, 22, 23. Sofio, family, 4, 5. Sohag, 1, 31, 34-48, 54, 95, 125, 159. Sorrento, 79. Sousse, 92. 101, 109, 141, 147. Spain, 2, 21, 23, 30. Spaccaforno, 172. Squinches, use of, 31. Stables, 133, 134 : Haidra, 107, 125 ; Sbeitla, 107 ; Tebessa, 107, 115, 125. Staletti, 19. Stilo, 52, 67, 76, 82. Strzygowski, 44. Sufetula. See Sbeitla, 9. Suwaris Basta, 45, 159. Syracuse, 11, 19, 55, 56, 59, 61, 71. Syria, 1, 13, 17, 30. TABARKA, mosaic in the Bardo Museum from, 88, 94. Tancred, 75. Taormina, 49, 54, 57, 172. Tebessa, 15, 17, 31, 33, 47, 89, 94, 95, 98, 100, 101, 107, 109, 111, 114, 115, 116, 117, 122, 123, 126, 130, 131, 134, 135. Teboursouk, 136, 138, 139. Teodemaro, Abbot of Monte Cassino, 30, 82. Termini, 59. Terranuova, Sicily, 55. Thala, 109. Thelepta, 31, 32, 54, 110, 112, 113, 125. Theodora, the Empress, 13, 34. Theodosius, the Emperor, 48, 89, 138, INDEX 181 Theophano, the Empress, 72. Theophilus, the Emperor, 70. ThibiUs or Announa, 127. Thinna, 146. Tibar, 101, 107, 138, 139, 156. Timgad, 97, 98, 101, 109, 116, 119, 126, 127, 129. Tombs: Dougga, 136, 137; Haidra, 132 ; Guehiia, 95 ; Sidi Abich, 145 ; Tebessa, 95. Torcello, 89, 99, 121. Tosti, 83. Toulotte, Monsgr., 144. Tozeur, 31, 100. Trajecta, 71, 80. Trani, 74, 75. Tribunes or clergy seats. See Altars and Apses. Tridetti, 172. Tunis, 6, 9, 12, 14, 17, 31, 40, 41, 47, 48, 54, 85-152. See place names, 156, 166. Turkey, 13, 53. Type, the edict of the Emperor Constans II, 10, 11. Tyrrhenian, sea power in the, 70, 71. UMMIDIA QUADRATILLA, 30. Uppena, 91, 92, 93, 94, 101, 108, 141- 146. VAL D'ISPICA, 19, 172. Vandals, 7, 8, 9, 20, 47, 48, 85, 86, 90, 91, 103, 108, 116, 117, 124, 131, 138, 141, 146, 152. Vaults, construction of, 50, 54, 59, 60, 76, 77, 122, 154, 158, 164; with tubes of pottery, 100. Vento Raimondo, 68. Vinci, Coramendatore, 19. VioUet le Due, 21, 23. WADY HALFA, 31, 82, 154, 157, 162. Watson, Sir Charles, 33. "Woolley, 154, 162, 163. ZANCLOS, 4. Zisa, 2, 53. Zosimus, Bishop of Syracuse, 59. y^ ,^ i UNIVERSITY OF CALIFORNIA LIBRARY BERKELEY Return to desk from which borrowed. This book is DUE on the last date stamped below. 230cl5lDI? (jOcttSLU 260ct'S3Vl ^0-^^ ^^^^Z y^ NOV 11962 3\9^^ 9S REC'D DEC 30 '65 -11 AM FEB 2 5 1965 34 FEB16 19688S MARlU'QQ. 5fM G)-176 Freshfield, E.»H. _ Cellae trichorae DG865 F75