AGORA ATHENIAN

Volltext

(1)THE. ATHENIAN AGORA RESULTS OF EXCAVATIONS CONDUCTED. BY. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. THE AMERICAN SCHOOL OF CLASSICAL STUDIES AT ATHENS. VOLUME. THE THE. XX. CHURCH HOLY. OF. APOSTLES BY. ALISON. FRANTZ. THE AMERICAN SCHOOL OF CLASSICAL STUDIES AT ATHENS PRINCETON,. NEW JERSEY. 1971. American School of Classical Studies at Athens is collaborating with JSTOR to digitize, preserve, and extend access to The Athenian Agora ® www.jstor.org.

(2) © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. A,M. '_';. .,..'. ,. .. ". ,'-. ' ". ". '. " ..'. -. The Holy Apostles.. The Holy Apostles..

(3) WITH THE AID OF A GRANT. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. PUBLISHED. ALL RIGHTS. PRINTED. IN GERMANY. FROM MR. JOHN D. ROCKEFELLER,. RESERVED. at J. J. AUGUSTIN,. GLUCKSTADT. JR..

(4) © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. FOR. ANASTASIOS K. ORLANDOS.

(5) © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. PREFACE The investigationandrestorationof the Churchof the HolyApostleswereundertakenby theAmerican Schoolof ClassicalStudieson the invitationof theDepartmentof Restorationin theArchaeological Serviceof the GreekGovernment,underthe Directorshipof AnastasiosK. Orlandos.Warmestthanks are here expressedon behalfof the School to ProfessorOrlandosand his associate,EustathiosStikas, for invaluableassistancethroughall phasesof the undertaking. Theprojectwasmadepossibleby generousgrantsfromthe SamuelH. KressFoundationof New York whichmet all the expensesof the preliminarywork, the restoration,and the preparationof the final publication.In expressingour gratitudeto the Foundationit is a pleasureto recallthe personalinterest shownby the late RushKresswho, withMrs.Kress,visitedAthenstwicewhilethe workwasin progress. We are also indebtedto Miss Mary Davis, Vice Presidentof the KressFoundation,for continuedinterestand help, and to Mrs.MurrayDanforth,Mrs.HenrySharpe,and Mrs.C. AlexanderRobinson,all of Providence,Rhode Island,for theirassistancein the landscapingof the area. Many scholarsvisitedthe churchduringthe courseof the work and gave the benefitof theirexpert knowledge.I profitedespeciallyfrom discussionon the spot with GeorgeH. Forsyth,RichardKrautheimer,A. H. S. Megaw,RichardStillwell,and the late PaulA. Underwood. A numberof colleaguesin the Agorahavehelpedin the publicationof the church.NicholasRestakis circumstances. Poly Demproducedthe prints,sometimesfrom negativesmade underdisadvantageous oulini smoothedthe way in endlessmattersof long-distancecollaborationin the final stages.I am indebtedto MargaretCrosbyfor her patiencein allowingan intruderto encroachon her own areaof excavation,and for crowningher kindnessby makingpossiblethe color platefor the frontispiece. The restoreddrawingsof the church(Pls.29-37, 40) arethe ornamentof this book. Theyarethe work of WilliamB. Dinsmoor,Jr., to whomI expressmy gratitudeand admirationfor his patienceand skill. Specialthanksandappreciationgo to HomerThompson,classicistpar excellence,on whoseinitiativethe studyand restorationof this mediaevalmonumentwereundertakenand completed,and who couldconsiderthe vagariesof Byzantinebuilderswiththe same care that he bestowson the precisionof the architects of the 5th century B.C.. To recordadequatelymy debt to JohnTravloswouldrequirean acknowledgement on everypage. He has generouslygiven his counsel and sharedhis knowledge,and his mastery of both the practical and the theoreticalaspects of Byzantinearchitecturehas facilitatedthe excavation,accomplishedthe restoration,and enrichedthe publication. To ProfessorAnastasiosOrlandosI offerthis opusculum in token of affectionategratitudefor many kindnessesovera quarterof a century,and in recognitionof his incomparablecontributionto the study of Byzantinearchitecturein Greece. ALISON FRANTZ. Princeton,N.J. June20, 1971.

(6) © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. TABLE OF CONTENTS PAGES vii. PREFACE.................................................................................... ... LIST OF ILLUSTRATIONS. ..... .... ..... .... ... ... .................................................... FIGURES ................................................................................... X. ...... PLATES ............................................................................... xiii. ..........................................................................i... ABBREVIATIONS INTRODUCTION..................... 1. ......... ....................................... ......... REMAINS................................................................ THEEARLIER. 3 4. THE FIRSTPERIOD ............................................................................ THEPLAN.............. .. .................................................... .4............ 4. .. 5 7 7 9 10. . . . . THE FOUNDATIONS ..................................5........................... THE MASONRY............................................................................. BRICKWORK THE ORNAMENTAL ............................................................ THE COLUMNS............................................................................. .................................. THE DOMEAND VAULTS............................... THE ROOF ....................................................................... THEWINDOWS.............................................................................................. ...... .................................. THED OORS. ...................................... THEPAVING........................................................ ... ................... ....................... THEWALLDECORATION THE ARCOSOLIUM .. ................ THE ICONOSTASIS. THEFONT ........................ TYPE .................. ARCHITECTURAL DATE ........................ THETOMBS.....27.......... ...... 10 12 12. ... .......13. 13 14. ....................................... ........................................... .... ......................................... ..... ..... ..... ........................................... ................... .......................................... MATERIALFROMTHECHURCH..3.......0................. HUMANSKELETAL. THELATERPERIODS .......................................................................... II ............................................................................ PERIOD PERIOD III................................................................................. PERIOD IV .......................................................................... X. 17 18. .24. ... ..... 27. ............ 30. .. .. . .. 32 32 35 38. CONSERVATION ANDRESTORATION ............................................................... 40. INDEX ..................................... 43. PLATES.

(7) © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. LIST. 1. 2. 3. 4. 5. 6. 7. 8.. 9. 10. 11. 12.. OF. ILLUSTRATIONS FIGURES. Masonryon East Side of East Apse (J. Travlos). KufesqueDesignsin Masonry(A. H. S. Megaw,preparedby H. Besi). WindowDetails. A. East Apse. B. Dome, SoutheastFace (W. B. Dinsmoor,Jr.). Windows. a. EastApse. b. NorthApse. c. Diaconicon. d. Dome, SoutheastFace (W.B. Dinsmoor,Jr.). Door Frameof PeriodI (J. Travlos). SarcophagusFront, RestoredDrawing(J. Travlos). Iconostasis,RestoredDrawing(J. Travlos). Churchesin Greece. a. Methana, St. Nicholas. b. Platani, St. Nicholas. c. Kastoria,Koumbelidiki. d. Varasova,St. Demetrios. e. Athens, Moni Petraki. f. Manolada,Palaiopanagia. g. Athens, Holy Apostles. h. Gavrolimni,Panaxiotissa. i. Arta, St. Nicholas'PoSi&as.(1:200) SectionsthroughChurchProperand Narthex,showingTombs(J. Travlos). The Holy Apostles.Plan by A. Lenoir(1836). The Holy Apostles.Plan of 1854. Positionof Orthostatesin WestWall (J. Travlos).. PLATES. The Holy Apostles. FRONTISPIECE 1. The Holy ApostlesbeforeRestoration(1953). a. East End. b. West End. c. North Side. d. South Side. 2. a. Foundationsof Churchon Floor Beddingof Nymphaeum. b. Foundationsof Mint (?) underWestApse. c. EarlyByzantineHouse WallsunderEast Apse. d. The Churchca. 1890(Phot. G. Lambakis). 3. a. Foundationsof West Apse from East. b. Crosswallof PeriodsI-III, from West. c. Junctionof WestApse and AdjoiningWall. d. OriginalStep Blockin Place. 4. WallsStrippedof LaterMasonry. a. North Side. b. OriginalEnd of North Wall. c. South Side. 5. a. Foundationsof NorthwestAngle Chamber,from Southeast. b. Brickwith Sketchfor KufesqueDesign. c. SouthwestCapital. d. NorthwestCapital. e. SouthwestColumnBase. f. NortheastColumn..

(8) PLATES. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. 6. a. Detail of Dome afterRemovalof Stucco. b. OriginalWindowArchesin Dome. c. SoutheastFace of Dome with OriginalMullion. d. SouthwestPendentive. e. NortheastSegmentof Interiorof Dome. 7. a. WestVaultbeforeRepairs. b. Semi-domeof SouthwestAngle Chamber. c. Vaultingof SouthwestAngle Chamber. d. TriangularVault of SouthwestAngle Chamber. 8. Dome and Roofs afterRestoration. 9. a. PumicePackingover East Vault. b. PlasterBeddingfor Roof Tiles. c. Junctionof East Apse and BarrelVault. d. KufesqueBrickworkin South Apse. e. Windowsof East Apse and Diaconicon. f. RestoredWindowof Narthex. 10. a. OriginalPavingin NorthwestAngle Chamber,from South. b. Arcosolium,from North. c. Restoringthe Omphalos. d. Font in ByzantineMuseum(photographby EugeneVanderpool,Jr.). e. SarcophagusFront. 11. Fragmentsof Iconostasisof Holy Apostles. a. EpistyleBlock D, Front. b. EpistyleBlock D, Underside. c. EpistyleBlock E, Front. d. EpistyleBlock E, Underside. e. ClosurePanel. f. ColumnB.. 12.. 13.. 14.. 15.. 16.. 17.. B. Epistyle Block in Byzantine Museum.. h. EpistyleBlock found in Asklepieion. i. ClosurePanelfrom Moni Petraki. j. ClosurePanelin ByzantineMuseum. a. Manolada,Palaiopanagia,from Northwest. b. Manolada,Palaiopanagia,from Southeast. c. Manolada,Palaiopanagia,Narthex,North End. d. Hosios Loukas,Iconostasisin Katholikon(CourtesyE. Stikas). a. Athens, Moni Petraki,from East (CourtesyM. Sotiriou). b. Gavrolimni,Panaxiotissa,from Northeast. c. Hosios Loukas,Katholikonand Theotokos,from East. d. Hosios Loukas,Katholikon,from Southwest(CourtesyE. Stikas). a. Holy Apostles,West End beforeRestoration. b. From Northwest,RestorationCompleted. a. Tombs 1-4. b,c,d. Jugsfrom Tomb 2. e. Glass Bottlefrom Tomb 3. f. Lintelof SmallDoor of PeriodI. g. Tombsin Narthex. a. Tombs 12 and 13. b. Tombs 14 and 13. c. Tomb 11. d. Cornerof Arcosoliumwith Late XcovEurrpilov. e,f. SouthWall of Narthexin PeriodII. WallPaintingsof PeriodII. a. Angel of Trinityand Sacrificeof Isaacin Prothesis. b. Aaronin Dome. c,d. Saintsin West Angle Chambers.. xi.

(9) © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. xii. LIST OF ILLUSTRATIONS. 18. a. Gable of Later Periods. b. Window of Period II in North Wall. c. Altar and Floor of Period IV. d. Pantokrator in Dome, Period II. e. Pantokrator in Dome, Period IV. f. Blocked Doorway in North Apse, from Outside. 19. a. Narthex, with Threshold Adjusted for Period III, from East. b. West Wall, Orthostates Replaced in Original Positions. c. Interior, Period IV, Looking East. d. Interior, Period IV, Looking West. 20. a,b. Interior Bracing. c,d. Piers of West Arch after Destruction of Apse. e. Inner Face of Wall before Restoration. 21. Replacing Columns. a. Southeast Column in Process of Removal. b. Reinforcing Rods in Place. c. Cement Bases for New Columns. d. Half of Mould for Shaft. e. Mould for Core of Capital. f. Mould for Column in Place. 22. Reconstruction of Narthex. a. Centering. b. Reinforcing Rods. c. Roof Half Finished. d. Rebuilding the Fagade. 23. a. Laying the Roof Tiles. b. Carving the Lunette. 24. After Restoration, from Southeast 25. After Restoration and Landscaping, from Southeast (1959). 26. The Interior, Restored. a. Looking East. b. Looking Southeast. 27. The Interior, Restored. a Looking Southwest into Narthex. b. Southeast Bay Narthex, looking into Church Proper. 28. Development of Plan (J. Travlos). 29. Restored Plan, Period I (W. B. Dinsmoor, Jr.). 30. East Elevation (W. B. Dinsmoor, Jr.) 31. West Elevation (W. B. Dinsmoor, Jr.) 32. North Elevation (W. B. Dinsmoor, Jr.) 33. South Elevation (W. B. Dinsmoor, Jr.). 34. Transverse Section, Church Proper (W. B. Dinsmoor, Jr.). 35. Longitudinal Section (W. B. Dinsmoor, Jr.). 36. Transverse Section, East End (W. B. Dinsmoor, Jr.). 37. Transverse Section, Narthex (W. B. Dinsmoor, Jr.). 38. Plan, Period III (J. Travlos). 39. Floor Plan, Period IV (J. Travlos). 40. Plan of Roof (W. B. Dinsmoor, Jr.)..

(10) © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. ABME 'ApX.'Eq. B.N.J. Bouras,Manolada AIEE D.O.P. AXAE Ebersolt EEBX EMME. ABBREVIATIONS A. K. Orlandos, 'ApXEiovrCOvBuLavrTv~$v MvriaEkcov -ris 'E?a&8os, Athens. Athens 'ApXatoXoyu<~ 'Eq:,rIEpis, Athens Jahrbiicher, Byzantinisch-neugriechische Ch. Bouras,'H TTaaioi-rravayi'a MavoX&Sos, 'E'mrralMovvK' 'E'rrelp'S-riq Ho\Xvl'laveirloriou 'AporooTEXdov esE*aaoov[nrS, IV, 1969, -rExvuIIS2XOXs too Salonica,pp. 233-266 Athens inal'ESvoAoyKijS AE?J'iov-rfs 'Ic-ropudiS 'ET-rapEiaS, DumbartonOaksPapers,Washington AEX-rTov"rfSXplc-riavtKis 'ApxatoXoyt<is 'E'ratpEaS,Athens. J. Ebersolt,Monumentsd'architecture byzantine,Paris, 1934. Athens XTrovuS&v, 'E-relpis 'E-racpeiasBvLav-rtvwAv -r&vMEcatcoviK&$v Mvr1PEco0v E'VpEsTlptov Trq 'E?Na'?oS,Athens D. Kambouroglou, '1o'-ropioa Kambouroglou, '!aTrop[a -rGv 'A.9i[vaicov. ToupKoxpa-ria, -[EpioSos TrpcTrj,. 1458-1687,II, Athens, 1889 PelicanHistoryof Art, R. Krautheimer,EarlyChristianand ByzantineArchitecture, Krautheimer 1965 G. Lambakis,M6moiresur les antiquitischrftiennesde la Grace,Athens, 1902 Lambakis,Mgmoire A. Lenoir,L'architecture Lenoir,L'architecture monastique,I, Paris, 1852 H. "The Chronologyof some Middle-ByzantineChurches,"Annualof the Megaw, Megaw,Chronology BritishSchoolat Athens,XXXII, 1931-32,pp. 90-130 byzantine,Paris, 1916 Millet, L'ecolegrecque GabrielMillet,L'ecolegrecquedansl'architecture M. Sotiriou,T6 KaSoX7K6v AXAE,1961,pp. 101-129 'riisMovfs VTETrpa<Kr Sotiriou,Petraki 'ASqv$-vv, T6 oK6v E. Movij TroJ 'Oai'ov AoVKa $COKi?oS,Athens, 0tKoSOI. Stikas, Stikas,H.L. XpoviK6v-rfj5. Travlos, 1O?'EoX8oPK'1. 1970 'E J. Travlos, 7OAEOA5OIIKI QEXt~ts'rLv 'A$rivov, Athens, 1960.

(11) © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. INTRODUCTION T he Byzantinechurchof the Holy Apostles(Frontispiece)is one of the two buildingsin the Athenian Agora to have remainedstandingfrom the time of its constructionto the present.The Templeof Hephaistos,whichlooks down over the Agorafrom Kolonos Agoraioson the west side, sufferedsome minorvicissitudesbetweenthe Herulianinvasionof A.D. 267 and its conversioninto a Christianchurch in the 7th century.From that time on its new status ensuredthe templeprotectionand maintenance throughtwelvecenturies,afterwhichit was retiredfromthis pre-eminentpositionto the safe statusof a nationalmonument.The Churchof the Holy Apostles,althoughdamaged,survivedoccupationand invasionby Franks,Turksand Venetianslargelythroughthe accidentof havingbeen built overthe solid foundationsof a Nymphaeumof the 2nd centuryafterChrist.Thus the two buildings,one pagan,the otherChristian,owedtheirsurvivalto the ironicandunwittingagencyof the opposingreligion,to which each was implacablyopposed.' At the time of the constructionof our churchAthenshad only recentlyemergedfroma long periodof decline and debility which began in 529 with the closing of the schools by edict of Justinian. Threat of. invasionby Slavictribeson land and Arabpiratesby sea furtherdiscouragedeconomicgrowthas many of the inhabitants retreatedto the safer inland areas. By the end of the 9th century, however, the Slavs had been brought under control and the recapture of Crete from the Arabs in 961 reduced the danger. fromthat direction.Withthe pressuresrelaxedon both sidesAthensgraduallyrecoveredand by the beginning of the 1nthcentury had entered on a period of relative prosperity which was to last until the establishment of the Latin Kingdom of Constantinople after the Fourth Crusadeof 1204. These two centuries saw the erection of the major Byzantine churches in Athens, of which the Holy Apostles is one of the two earliest,2and in 1018 the city received an imperial visit when Basil II, Bulgaroktonos,passed. throughafterhis successfulcampaignagainstthe Bulgarsto give thanksfor his victoryin the Churchof the Virgin, theonce the Parthenon.3. The Churchof the Holy Apostles standsover the southeastcornerof the Agora at the side of an. important crossroads of both classical and Byzantine times, just west of the Panathenaic Way and the Post-HerulianWall which had protected the city in times ofstres s since the end of the 3rd century. 1 The churchis mentioned or describedbrieflyin the following sources: K. S. Pittakys, L'ancienneAthnes, 1835, pp. 43, 46,48, 70; Lenoir, L'architecturemonastique,I, 1852, p. 252; A. Mommsen,Athenaechristianae,1868, pp. 24-25, 91-92; A. Choisy, L'art de batir chez les byzantins,1883, pp. 132-133; idem,Histoirede l'architecture,II, 1889, pp. 34-35; G. Lambakis,ZTO&,12 August de la Grice, 1902,p. 12; Neroutsos, AIEE, 1884,p. 3; idem,'Ep5opas,1884,p. 189, note 1P;idem,Memoiresurles antiquiteschliretiennes III, 1889, p. 74; Kambouroglou,'lOropia, II, 1889, p. 293; idem,TTcxXaial 'ASfivat, 1922, p. 160; 0. Wulff, Die Baukunst,II, 1903, pp. 395,481; A. Struck,AthenundAttika,1911,p. 141; Th. Philadelpheus,'lo-ropia-rv 'ASrvwvvTrrl 1902,1, p.276; TovpKOKparias, J. Strzygowski,Die Baukunstder Armenier,1918, II, p. 798; G. T. Rivoira, LombardicArchitecture:Its Origin,DevelopmentandDerivatives(tr. G.McN. Rushforth),1910, I, p. 187; A. Xyngopoulos, EMME,I, pp. 77-79; Megaw, Chronology,passim; Ebersolt,Chap. X; Orlandos,ABME,VII, 1951,p. 154, and XI, 1969, pp. 65-66, 81-82; A. Frantz,Byzantion,XXIV, 1954, pp. 513-520; eadem,The Middle Ages in the AthenianAgora, Picture Book No. 7, 1961, figs. 46-59; J. Travlos, HOEMoSovtK1, 1960, pp. 151, 155; idem, XplcaavtKal. 'ASfivat,. in eplcKEUTtriK' Kal 'HIK9 'EyKUKXoTraiSeia, 1962, p. 738; Krautheimer, 1965, pp. 275-277; Propylaen Kunst-. geschichte,III, Byzanz(Chatzidakis), 1968, p. 225 and fig. 148; Stikas, H.L, 1970, pp. 193. 2 The katholikon of the Moni Petraki is now shown to be the first; both churchesare probablyto be dated in the 10th century (below, p. 23). 3 Kedrenos,II, 475 (Bonn); Glykas, IV, 578-579 (Bonn); Zonaras,XVII, 9. Basil was the firstemperorto visit the city since Constans II winteredthere in 662/3..

(12) 2. INTRODUCTION. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. The epithet Solaki, which has long been attachedto the church,derivesfrom the quarterin which it stands.4 The earliestknowndescriptionof the churchis by the Frencharchitect,AlbertLenoir,who visited Greeceand Constantinoplein 1836for the purposeof studyingthe historyof Christianarchitecture.5 Althoughthe majorpartof the churchhad remainedstanding,it had beendisfiguredby wellintentioned effortsto preserve,enlarge,and embellishit. As long as it was surroundedby the housesthat madethe areaa slumuntil 1931the inartisticadditionsof the late 19thcenturywerehardlynoticeable.But when the shacksthat clusteredaroundthe buildingweregraduallyclearedawayit becamean eyesorefromall directionsexceptthe east, whichhad remainedrelativelyunmolested.Furthermore, closerexamination revealedthat the fabric,althoughapparentlysound on a superficialview, was actuallyin a precarious statedue to the crumblingof the mortar.Sincesubstantialmeasuresof conservationwereobviouslyessential,it seemeddesirableat the sametimeto investigatethe buildingthoroughlyfromthe archaeological standpointin orderto recoverthe originalplan,whichwasconcealedby the late additions,to solidify whateverremainedof the originalbuilding,and to restoreit as far as possibleto its originalappearance by meansof whatevernew constructionmightbe necessary. At the beginningof workthe areahadalreadybeenfreedof the squalidhousessurrounding the church, one meterabovethe original but the pavedcourtyardhad beenleft undisturbedat a levelapproximately groundlevel (P1.1,a-d).At that time the aspectof the buildingwas that of the finalremodelingand enlargementcarriedout in 1876-1882,as recordedin an inscriptionon the bell tower.The mainfeatureof thisremodelingwas a largewesternextensionforminga navewhichwas saddleroofedat a heightgreater thanthatof the mainvaultsof the originalbuilding,thusobscuringfromthewestalmostthewholebuilding exceptthe dome. Through careful demolition of the late walls and excavation inside the building it was discoveredthat instead of the two visible phases, the first and the last, there were four main building periods in the history of the church: the original construction in the late 10th or early 1lth century; a first remodeling necessitated by damage to the west end, probably in the late 17th century; an enlargementsoon after the War. of Independence,between1836and 1854;and the more substantialrenovationof 1876-1882.6 The archaeologicalinvestigation was begun on February 12, 1954, and the restoration was completed. in timefor the churchto sharein the dedicationceremoniesof the Stoaof Attaloson September3, 1956.7 4. Mommsen, Athenaechristianae,pp. 24-25. The origin of the name is uncertain,but Pittakys' conjecture(L'ancienneAthenes, olKol,on the ground that here was the house of the legislator,need not be regardedany p. 70) that it was derivedfrom 2X6Acovos more seriouslythan the even more enticingbut equallyunsubstantiatednotion that the churchwas erectedon the site of the Altar of the Twelve Gods. A more probableexplanationis that of Kambouroglou('lo-ropiac, p. 293) that the epithetcame from the name of a "greatAthenianfamily" living in the vicinity. 5 L'architecturemonastique,I, p. 252, fig. 163. The account of his trip is found in "Rapportfait par M. Albert Lenoir, architecte, sur son voyage dans le Levantpendantl'ann6e1836," Annalesde la Societi libredes Beaux-Arts,1837.The resultsof this trip formed the basis of L'architecturemonastique,which was publishedin 1852. In that same year he issued a pamphlet, Instructionsa l'usage des voyageursen Orient:Monumentsde l'ere chritienne,in which he comparedthe Holy Apostles with Eusebius'descriptionof Constantine's Golden Octagonin Antioch. For the relevanceof Lenoir's descriptionof the Holy Apostles to the history of the church, see below, pp. 32-34. 6 For the later periods see below, pp. 32-39. 7 Accounts of the work as it progressedmay be found in Homer A. Thompson, "Activitiesin the Athenian Agora," Hesperia, XXIV, 1955, pp. 55-57; XXV, 1956, pp. 65-66; XXVI, 1957,pp. 101-103..

(13) © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. THE. EARLIER. REMAINS. C ontraryto a local tradition'there is no evidenceof a cryptbeneaththe churchnor of any earlier churchon the samespot. All structuralremainsunderthe churchwereunrelatedto it. The earliest of thesewas a shortstretchof the foundationfor the westwallof a buildingof the 5thcenturyB. C.,2 probablyto be identifiedas the Mint, whichwas foundjust insidethe westernapse (Pls. 2,b, 28). Most of the easternhalf of the churchrestedeitheron the bottomof the curvedwall trenchof the Nymphaeum3 or on the concretebeddingfor its floorslabs(Pls. 2, a, 28). Immediatelywest of the Mintwallweresome of the 4th centuryafterChrist.Any earlierremainsbeneaththe remainsof a metalworkingestablishment centralpart of the churchhad been eitherremovedor concealedby the tombswhichoccupiedthatarea. East of the iconostasis,however,wheretherewereno tombs,was a stretchof a roughwall foundation madeof rubblebeddedin clay. It restedon the foundationof the Nymphaeumand ran southwardfrom close to the northwall as far as the southeastcolumn,whereit was brokenaway.Anothershortbit of the samemasonryled off eastwardnot far fromthe preservedsouthend. Thesefoundationspresumably belonged to a house of earlierByzantinetimes, perhapsdemolishedto make room for the church (Pls. 2,c, 28). 1 Referredto by A. Xyngopoulos in EMME,p. 79, and also by Kambouroglou ('lo-ropia, pp. 293-294), who includeda sketch of a crypt wrongly attributedto the Holy Apostles. The sketch was originally publishedby E. Breton, Athenes2,1868, p. 182, and correctlyidentifiedas the wellhouse of the Klepsydra,on the north slope of the Acropolis. The wellhousewas consecrated,perhaps as early as the 10th century, as the chapel of the Holy Apostles (cf. A. W. Parsons, Hesperia,XII, 1943, pp. 250-251 and fig. 21, p. 222); hence the confusion. The chapel is mentioned in EMME,p. 103 under the name Ayioi 'ArooroToto"T-ra Ip&ppapa."The same sketch appears in TTaAaia'ASiva, Nb6ocr 'AaoAovTXvirs, 1931, p. 65, fig. 44 as "the crypt of the church of SS. Theodore, near the Tower of the Winds." 2 Hesperia,XXIV, 1955, p. 59. 3 Ibid., pp. 57-59..

(14) THE FIRST PERIOD © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. Before the workof restorationbeganthe churchpresentedthe appearanceof a triconch,withapseson thenorthandsouthsidesas wellas theeast.Thewestendwascompletelyobscuredby thelongmodem addition.A photographtakenby G. Lambakisca. 1890(PI.2,d), afterthe constructionof the addition but beforeit had receivedits finalcoat of stucco,showedcloisonnemasonry,apparentlycontemporary withthe earliestpartsof the building,extendingas farwestas the door in the northwall.Thissuggesteda triconch with a nave, after the "Hagioritic" plan, so called because of its frequency in the churches of. Mt. Athos.1Thosewho acceptedthis plan as the originalexcludedthe possibilityof a fourthapse. But according to an old tradition the church had originally been a baptistery2and therefore a tetra-. conchand,in fact,an apseof a remarkable shapewas shownby Lenoir(Fig. 10).3Choisy,also, described the church as having had four apses, one of which had been destroyed.4In this view the prolongation of. the wallswestwardwas regardedas a lateraddition,eventhoughthis wouldleaveunexplainedthe lack of symmetrybetweenthe obtuse angles of the prothesisand diaconiconand the rightangledwestern anglechambers.Ourfirstobjective,then,was to confirmor disproveone or the otherof thesetheories by removing all of the modem masonry and by excavation inside the church, and to discover where and how the building had originally terminatedat the west.. THE PLAN. The problemof the groundplanwas quicklysolvedby excavation(Pls. 28, 29), whichprovedboth of the opposingschoolsof thoughtrightin somedegree.Not far belowthe pavingwereuncoveredthe founto the threealreadyvisible,butprovidedwitha doorway(PI.3,a). dationsof a fourthapsecorresponding western that the At the sametimeit was clear apsehad alwaysbeensurroundedby a narthex,the extent of whichwas fixed by the foundationfor a crosswallconnectingthe northand southwalls,whichcame to light2.65m. west of thewestface of the apse(PI.3,b). Two coursesof masonrywerepreservedabove the groundlevel of the apse; they were of cloisonne,but simplerand rougherthanthat usedelsewhere in the building.The foundationswerewell bondedinto the adjacentwalls(P1.3,c). Whenthe wallsof the modernadditionwerestrippedof all latermasonrythe originalconstructionwas foundto reachas far as the doorwaysin the northand southwalls(PI.4, a). At this pointthe northwall showeda finishedendface,exactlyat the line of the newlydiscoveredcrosswall(PI.4, b). The foundations of all the wallsruncontinuouslywithoutany breakor changein construction.Pierswerebuiltinto both endsof eachof the wallsof the westernapse,withrespondswellintegratedinto the masonryof the three wallsof the narthex,showingthatthe narthexwas vaultedin threebaysat the westandtwo, flankingthe 1EMME, I, p. 77. Cf. Lambakisin 'EpSouds,1884, p. 189, note 1. 3 L'architecture,I, p. 252, fig. 163, whence our Fig. 10. 4 Auguste Choisy, Histoirede l'architecture,II, 1899,p. 33. Choisy had alreadypublishedthe buildingin some detail in his L'Art de batir chez les byzantins,pp. 132-133, which appeared in 1883,just one year after the completion of the final restoration.Given even an averagelapse of time betweenstudy and publication,he must have seen the churchwell before its latest phase, when any traces of an apse were concealedby the new marblepavement. 2.

(15) THE FOUNDATIONS, THE MASONRY. 5. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. apse,at the east (Pis.28,29). The northwall was preservedoverits entirelengthup to the heightwhere the slope of the gablebegan.The southwall remainedfor its wholelengthonly in the lowercourse,descendingin a jaggedline from its full heightat the east end to only four coursesat the west (PI.4,c). with a dome on pendentivescarriedon arches The plan is thus revealedas basicallya cross-in-square supportedby four free-standingcolumns.It is, however,elaboratedinto a tetraconchby the additionof a three-sidedapse, similarto that at the east end, on each of the otherthreesides.The plan is unusual in thatthe apsepenetratedinto the narthex,whichenclosedits lowerpartcompletely,leavinga trapezoidal spaceon eachsideof the apse.Thus,only the easternelementsof the building,the threeapsesto east, northand south,and the prothesisand diaconicon,with theirwallsformingobtuseangles,stood out on the groundplan.The west apse and the westernanglechamberswere visible only abovethe roof of the narthex.The anglechambersat this end wereright-angled,to take accountof the prolongationof their wallsinto those of the narthex,and theirwest wallswerepiercedby archeddoorwaysto providecirculation betweenthe side bays of the narthexand the mainbody of the church. In the interiorall apsesand anglechambersare semicircular,and a semicircular niche, subsequently blockedup, was sunk into the opposingwall of each of the westernangle chambers.Entranceto the churchwas providedby a largecentraldoorwayin the westwall of the narthex,flankedby two narrower openings.The originalstep block was found in placein front of the middledoor (P1.3,d). The buildingis not quiteregular.No two walls are exactlyparalleland the anglesof the outerwalls of all the apsesdifferin somedegree.The asymmetryis less pronouncedin the interior,wherethe major apsesand also the anglechambersare all a little morethan a semicircle(ca. 200?). The architectof the HolyApostleswas ableto use the eastwardrisinggroundline to good effectby introducinga two-stepchangeof level withinthe church,at the entranceto the westernapse, in addition to thosedictatedby commonpractice,i.e., fromthe outsideinto the narthexandagain,viathe solea,into the sanctuary.The climacticeffectis now very apparent,afterrestoration,as the visitorwalksinto the narthexfrom outdoors,and againas he mountsthe two intermediatesteps into the churchproper.A comparableeffectwas achievedalso on the outside,wherethe lowerlevelwas usedto diminishthe apparent heightof the narthexin relationto the westernapse and thus allow the latterto be seen to full advantage. THE FOUNDATIONS. The foundationsconsist of rubblemasonryset in firm lime mortar(PI.5,a) and vary in depthfrom 0.75 to 1.10m., the highestpoint beingat the east.Thisis accountedfor by the graduallyrisingground level in this directionwhichprevailedat the time of the constructionof the churchand was adheredto in the junctionof the foundationswith the cloisonnemasonry,whereasthe bottomline was dictatedby the level surfacesof the ancientmonumentson whichthe buildingwas bedded(Pls. 32, 33). The foundationswere laid exactlyon the linesto be followedby the walls,withboth innerand outersurfacesalready clearlydefined,and with no extraneousconnectinglines in the interior.The easternhalf of the of its building,as noted above,was built overthe foundationsof the Nymphaeum,on the underpinning massivesemicircularwall and on the heavyconcretebeddingfor its marblefloor slabs,whichhad been plunderedin antiquity(PI.2, a). This circumstance may accountin partfor the betterabilityof this section of the buildingto withstandthe damagewhichresultedin the destructionof the west end. THE MASONRY In the churchproperthe wallsof the originalbuildingweresubstantially intactwithonlyminorrepairs and alterations,chieflyon the south side.5The masonrythroughoutthe churchis a carefullylaid cloiBelow, p.5 39..

(16) THE FIRST PERIOD. 6. sonneconsistingof poroslimestone,chieflythecharacteristic Megarianstonewhichcontainsa largequantity of sea shells.The coursesare separatedby a double(veryrarelysingle)layerof bricksset in a firm whitemortarmadeof riversandand lime,with a considerableamountof grogmadeof groundup tile. The exposedsurfacesof the mortarhave weatheredto a pinkishbrown,but it was obviousduringthe repairworkthat whereit was protectedit was a startlingwhite,brokenup only by bits of grog. It was obviousthatthis was useddeliberatelyto givea strongcontrastwiththe stoneandbrick.Thesameeffect may be seenin churcheswheresomeof the masonryhas beenprotectedby laterconstruction,e.g. at the southwest comer of the Theotokos at Hosios Loukas, where the Katholikon abutted against the wall and. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. coveredits surface.6 Aboveeye levelalmostall of the verticaljointsarefilledwithbricksset in ornamentalpatterns(Fig.I Pls. 8, 9,d).7 In the lowercoursesthe simplestcloisonnemasonrywas used.. Fig. 1. Masonryon EastSideof EastApse Largeancientblocks were set on end at all key points of the building,in the lowest course:at the outer corners of all the apses and the narthex, and flanking thethhreedoorways in the west wall. In ad-. dition,two otherswereplacedat irregularintervalsin the southwallof the narthexand at leastone, and probablytwo, in the north(Pls. 4,c, 32, 33).8Behindthe cloisonnefacingthe thicknessof the wall was filledout witha core of rubblemasonry,givingthe wallsa total thicknessof 0.70-0.80m. 6 Stikas, H.L., p. 146, fig. 63, p. 151, fig. 68. 7 Below, pp. 7, 22.. 8 For the significance of these blocks in the dating of the church, see below, p. 25..

(17) THE ORNAMENTAL BRICKWORK, THE COLUMNS. 7. THE ORNAMENTAL BRICKWORK DENTILCOURSES. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. Five horizontalbricksawtoothfriezes,enclosedaboveand belowby a single course of bricks,break the monotonyof thewall faces(Pls. 1,a, c, d, 8). The uppermost,whichalso formsa corniceof threesuccessivelyprojectingrowsunderthe eaves of the majorapses,is separatedfrom the next belowby one course of cloisonnemasonry.From theredownwarda doublecourseintervenes.In addition,another dentilcornicerunsunderthe roof of the highestpartof the building,i.e. the fourbarrelvaultswiththeir gables.The lowestfrieze,whichis at the level of the springingof the archesof the windowsof the east apse, interruptsits courseto framethe brick archesof all the windows(Fig. 4). It also apparentlycontinuedon the same horizontalline to crownthe wallsof the narthexat the level of the beginningof the gable.. BRICKPATTERNS. Thepatternsin the verticaljoints of the cloisonn6masonryareof varyingdegreesof complexity(Fig. 2; P1.9,d).9Onlytwo can be consideredto havea Christological significance:No. 22, withalphaandomega, and No. 23, a leavedcrosswith IC XCin the upperangles.A few are simplegeometricdesigns;the rest are imitationKufic or Kufesque,10 all purelydecorative.n An interestingexampleof how the architecthandedon his ideas to the masonis providedby a half brickwitha Kufesquedesigndrawnin blackon its unevensurface(P1.5,b).12It was foundin the rubble core of the southwallof the southapseduringconservationoperations,havingapparentlybeen discarded after servingits purposeas a model and pickedup on the spot to be used as buildingmaterial.It cannotbe matchedexactlyon the churchbut Megaw'sNo. 6 (Fig. 2) is a close approximation.13. THE COLUMNS. The domewas supportedby archescarriedon fourfree-standing columns,4.07m. high,threeof which were still all had been taken from ancient still All four shaftsweremonolithic,of bluestanding; buildings. gray Hymettianmarble,with ancientcapitals.Those at the northwestand southeastwere late Corinthianin type(PI.5,d); the capitalof the southwestcolumnwas of the so-calledTowerof the Windstype, withlotusleavesspringingfrombehinda singlerowof acanthus(Pl.5,c).The northeastcolumnhad been replacedin some intermediateperiodby a makeshiftbuilt up of twenty-onedrumsof poros limestone rangingin heightfrom0.10to 0.20m.,threadedon an uprightironrod.Thedrumsweresecuredby melted lead pouredaroundthe rod. Unevenbeds of mortarwere laid betweenthe drumsand some effortwas madeto evenup the surfacesby drivingnails into the edgesof thejoints and evenoccasionallyinserting wholehorseshoes.An invertedIonic base servedas a capital(P1.5,f). In the finalperiodof the churchall the columnswerepaintedblackand the capitalswerepaintedin brightcolorstrimmedwith gilt. The two westerncolumnsrested,below floor level, on companionancientmarblebases, ca. 0.75 m. squareandca. 0.50m. high,withplainmouldingsat top andbottom.Thetop surfaceof eachwaschipped 9 The brick patternsof the. Holy Apostles form an importantpart of Megaw'sanalyticalstudy of brick patternsin general(Chtronology, pp. 102-115). Special thanks are here expressedto Mr. Megaw for permissionto reproducehis drawings,which include all the decorativeelementsfound on the church; our Figure 2 was preparedby Helen Besi from Mr. Megaw's drawings. 10 I have adopted GeorgeMiles' term 'Kufesque' for the "meaninglesssimulation of ornamentalKufic" (D.O.P., XVIII, 1964, p. 20). 11For the bearingof the ornamentalbrickworkon the date of the church,see below, pp. 24-26. 12 Inv. A 2523. 13 It is clear that the brick is to be regardedas a convenientpiece of scratchpaper ratherthan as the first stage of a design to be completedby the champleveprocess (Chronology,pp. 105-106). This latter possibilityis ruled out by the uneven surfaceand coarse consistencyof the brick and, even more, by the absenceof the champlevetechniqueelsewherein the church..

(18) 8. THE FIRST PERIOD. away to a depth of 0.05-0.06 m., except where the column rested, probably to receive the marble floor of the latest period.That they were originally usedin the Nymphaeumis suggestedby the presenceof a roughly cut channel, semicircularin section, ca. 0.04 m. wide and 0.02 m. deep, running vertically down the face. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. of the southwestbase. It showssome signsof waterwearand mightpossiblyhavebeen madeto receive a lead pipe (PI. 5,e).. t s. , I 2. 1. , 5. 4. 3. I. BS^im. \. 6. 7. A. 1t. ?. 1 -. A_-=. gL _A. 8. 2In. ""ie. 9. _^J. 12. 11. 10. IB. 1 15. 14. 21. 20. I1. 16. 0I. -- , - -. Fig. 2.. 19. 25. *. 31. 30. 29. I. /. ,,. 21. 37. 36. 35. 34. T - - - -. ;. 24. R 28. 13. 1B. 23. 33. 32. I. 17. 22. 27. 26. 1. 1"". t< sY. 1. __. _A. 1. 50 I. Kufesque Designs in Masonry. I. IM..

(19) THE DOME AND VAULTS. 9. Thefoundationsof the northeastcolumnconsistedof a largeblockof marblerestingon severalsmaller blocks of conglomerate,all bondedwith strongTherancement.The southeastcolumnrestedon an invertedIonic base.14Thefoundationsof all fourcolumnswerelaid directlyon the concretebeddingof the Nymphaeum. THE DOME AND VAULTS. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. Thedomeis of the traditionalAttictype: octagonal,withan archedcornicepushingwell up abovethe druminto the roof. The corniceis of Aeginetanporos limestonewith a broad,shallow,concavemoulding.Attachedmarbleshaftsat the eightcornersare surmountedby plainflaringcapitals(PI.6, a). The masonryis cloisonnewithsinglebricksin bothhorizontalandverticaljoints.Theappearanceof the dome had beengreatlymarredin the courseof timeby changesmadein the windows.Originallyeightin number, four had beenblockedup in the thirdor fourthphaseof the church,leavingopen only those at the cardinalpointsof thecompass.In the secondperiodall eighthadbeenremodeledby reducingtheirheight and toppingthemwith a low, flat arch(P1.6,b,c).15 The removalof the stuccowhichcoveredall the masonryof the domerevealedthe tops of the original windowframesstill in place.Theywerein varyingstatesof preservationand in all casesthe mortarhad crumbledbeyondall usefulness,buttheschemecouldbe recoveredwithcertainty(Figs.3, B,4, d; P1.6,b,c).. /. L-A. B. I. A-A. B-B. B4 I. 0 I. 1. .I I- . ,. W.B.D.,JR.-1969. 4 1. 3 1. M.. Fig. 3. WindowDetails. A. EastApse. B. Dome, SoutheastFace 14 Inv. A 4203. This might also be thought to be a late replacement,since the top of the base as found bore the impressionof the column in a bedding of Theran cement, but the contractorwho carriedout the restoration,Stratos Phergadiotes,reportshaving seen the same cement under the churchof the Katapolianion Paros. 15 For the sequence,see below, pp. 34-35..

(20) 10. THE FIRST PERIOD. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. Thisconsistedof a two-lightwindow,eachlightarchedseparatelyin brickand the wholeenclosedunder anotherarchwhichwas separatedfrom the corniceby a singlerow of bricks.The archesfor the lights sprangfrom a commonpoint in the middle,wherethey weresupportedby a singlemullion.Onlytwo mullionswerediscoveredin place, one in the south,the other in the southeastwindow.Probablyonly the latteris original;it is a slim stele-likerectanglewithits top matchingthe long narrowrestingsurface of thecapitalwhichit supports(Fig.3, B;P. 6, c). Thesouthwindowis nowdividedby an octagonalcolumn whichin no wayfitsits capital.Thetwo survivingcapitalsaredecoratedwithanincisedrosetteat eachend. The decorativeschemein the tympanumof the arches(Fig.4, d;P1. 8) was composedof a fairlylarge blockof poroslimestone,roughlytriangular,in the center,its loweredgechiseledto a point to fit in the two smallarches.Theremainingspacewasfilledwitha simplepatternof brickwork spandrelbetweenthe laid in a heavybeddingof mortar.Evenwherethe originalschemehad completelydisappearedon the outsideface, as in the southwindow,the archeswerefoundto be preservedhalfwaythroughthethickness of the wall, includingeven the facingof the soffit,whichwas a coatingof creamywhiteplaster,0.010.02 m. thick,with muchstraw(PI.6,b,c). The pendentivesand the ringof masonryon whichthe drumof the domerestedwereof brickslaid in mortarup to almosttwicetheirown thickness(P1.6,d).The drum,witha heightof 1.60m. and an inner diameterof 3.00 m., and the dome itself wereof well cut blocks of soft poros limestoneset in regular coursesin thin beds of mortar(P. 6,e). Barrelvaultson the mainaxesled to the four apsesand wereof similarconstructionto thatof the dome(PI.7,a). Thefouranglechamberswerecoveredwithsemi-domes, the remainingtriangularspacesbeingcoveredwithbrickvaults(PI.7,c,d). The lowerpartsof the semidomeswereof rubble;the upperwereof bricksset in thickbedsof mortarof less regularcostruction he adjacenttriangles(PI.7,b). The trapezoidalspacesflaking the westernapseweregrinaulted thanth but the threebays on the west side of the narthexwereprobablycoveredwith saucerdomescarriedon archeswhichbridgedthe gaps betweenthe piersin the walls and the cornersofthe apse.16 THE ROOF. The roofsof all partsof the churchexceptthe narthexwerewellpreserved(Pls. 8, 40). Broadpan tiles, cover tiles werelaid in a bed of mortarover a packinglargelycom0.51X 0.38 m., with semicircular posedof pumice(PI.9, a,b). The existingtilesappearto belongto the firstperiodof the churchor, if not, to a time when tiles of the same size and shapewerein use, as theirdimensionsexactlymatchthe impressionsin the mortarbedding.The tiles projected0.13 m. beyondthe two dentilcoursescrowningthe wallwhichin turnprojecteda total of 0.08m. beyondthe wall,andwiththesemakean attractivecornice. Junctionsbetweenthe differentroof levelswereachievedby a singlecourseof cloisonnemasonrywith Kufesqueelements(PI.9,c). The lowestof the dentilcoursessurroundingthe church,whichcorresponds to the pieseivedtop of the northwallOf the narthex,determinsthhight of the narthexbo the start overed of the gable.Betweenthis level ad thetir dentilcourseoiga cisonne masonryw d point at undermodernplasteron eachside the h of the westeraps, the lowercoursemar* whichthe westernanglechambersand apse becamevisiblefrom the outside(P1.14, a,b). THE WINDOWS The windowsin the mainbody of the churchareof the "arcade"type,17i.e. eachlightis archedseparatelyin brickand all lightsare of equalheight.The windowof the east apse is triple;thosein the north 16 Below, p. 20. 17 This is the name given. to the type by Megaw (Chronology,pp. 120ff., q.v. for the most useful discussionof the development of window design in Byzantinearchitecture)..

(21) 11. THE WINDOWS. and south apses and the dome are double (Figs. 3, 4; Pl. 9,e).18The mullions are elongated in section, on flaring bases and with flaring capitals decorated with incised crosses or rosettes. There is insufficientevidence for an accurate restoration of the windows of the narthex. The north wall had been pulled out at that point for the construction of an arcosolium (P1. 10,b), leaving only a small bit. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. -A. ,* .. A.. ..A.-. .. .l. A~~~~~. * ' ^'. *'. _. *-. J. #. ... '. v. t. b. a. e.. I. o - - --. .1a , .,C,~. ? /-. '. -!--:--:,. i.. .--. 'C. ,,e%. .o,,-. m-I,=-%M. ' ". '"1a-i' ' :;|*. - ' ~ .... ?6, .-.,.. . . . ' 2 -- . -?- .. '.'. ^ ;',_. ,. . . ,r >. . r. ... ,^,. \. ,. :. ~ . ..... .-... mat .. -. ..r ^- ^> , ^. ^-. Q i. Ii. c. Fig. 4.. 3i. I. W.B.D.,JR.-19. M.. 69. d. Windows. a. EastApse. b. North Apse. c. Diaconicon. d. Dome, SoutheastFace. 18 See also below, pp. 24-25..

(22) 12. THE FIRST PERIOD. of the crownof a brickarchwitha spanwideenoughto enclosea two-lightwindow.The corresponding part of the south windowis missing.Failingmorepreciseevidencethe windowshave been restoredon the analogyof those in the dome (P1.9,f). THE DOORS. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. Thedoorswereframedby mouldedjambsandlintelsof Pentelicmarble,of whichmanyfragmentswere foundin and aboutthe church.A lintel,foundintact,builtinto the later masonry,correspondedin size to the indicationsof openingsin the foundationsand could be assignedto one of the smallerdoorways (P1.15,f). Severalfragmentsjoinedto makeone completejamb,chiselledat the upperend to fit the cuttingsin the undersideof the lintel.It was thuspossibleto establishthe height of the small doorwaysat 1.875m. andtheirwidthas 0.655m. (Fig. 5). Thewidthof the centraldoorwaywasfixedby the threshold at 1.30m. In the absenceof definiteevidenceits heightwas restoredby analogyat 2.10 m.. Fig. 5. Door Frameof PeriodI. THE PAVING The originalpavingwas preservedin a numberof places along the walls,especiallyin the east apse, whereit was 0.19m. belowthe levelof themodernfloorin theeast apse,andin the northwestanglechamber,whereit had escapedthe later tombdiggers.The floorconsistedof flagsof irregularsize and shape,.

(23) THE WALL DECORATION, THE ARCOSOLIUM. 13. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. both graystone and whitemarble,jointedwith graylime mortar(PI. 0,a). The actualpavingwas confinedto the churchproper,but numerousindicationsof the floorlevelof the narthexestablishedthe fact that therewas a differenceof 0.44 m. betweenthe two partsof the church.19 Therewas no traceof a mortar bed for the floor. The remainingslabshad probablybeen relaidat a latertime. In the loose earthdirectlybeneaththe domewerefoundseveralfragmentsof opussectile,formingparts of rosettessuchas areusedin decorativepanelsnormallyfoundin this position.20 Consideringtheirfindit seems certain that these must have of formed the ing place, pieces part originalomphalosof the church, and they have,therefore,been incorporatedinto the restoreddesign(P1.10,c).2 THE WALL DECORATION. Nothing remainedto show the characterof the wall decorationof the firstperiod.The possibilityof mosaicsanywherein the churchis apparentlyruledout by the total absenceof remainsof the iron pins customarilyused to securethe backingof a mosaicto the wall itself.22 A few scrapsof frescowerefoundin the east apse,in an earlierlayerthanthe remainsof the painting of PeriodII. It is unlikelythat these werepart of the originaldecorationof the churchalthoughthey mightstill datefromwithinthe firstperiod.In all probabilitythe churchwas coveredfromthe beginning with paintingswhichwould have followedthe usualiconographicschemeand whichwould have been constantlyeitherfreshenedup or completelyrenewedas they crumbledaway.. THE ARCOSOLIUM. At some time after the constructionof the churchbut stillwithinits firstperiod,the northend of the narthexwas extendedto includean arcosolium.Thiswasaccomplishedby removingthe lowerpartof the wall betweenthe two westernmostpiersand rebuildingit 1.60m. beyondits originalface (P1.10,b).The gap was spannedby a brickarcha few centimeterslowerthanthat of the originalwindow,leavingonly the crownof the lattervisible(P1.32). The masonrywas cloisonnesimilarto, but not identicalwith,that of the originalbuilding.It was accentedby largeancientblocksstandingon end, similarto those used for the samepurposeelsewherein the church:one at each of the outercornersand one in the middleof the north side. The east and west walls werebuilt againstthe narthexwith no bonding.Moreover,the masonryof the narthexgives no indicationof havingbeen laid with an openingin mind;it has the appearanceratherof havingbeencarelesslytornout. Theopeningwasblockedonlyin the finalphase,after the destructionof the upperpartof the arcosolium,in orderto makea windowto fit the new schemeof fenestration. It is clearthat the arcosoliumis an afterthought,but how muchlateris an open question.The ground levelaroundit had not risenappreciablyand its masonryputsit well withinthe Byzantineperiod.Arcosolia in churcheswerenormallyused only for foundersor importantecclesiasticalpersonages.The fact that the arcosoliumof the Holy Apostleswas addedafter,but not long after,the erectionof the church 19. Above, p. 5. 20E.g. at Hosios Meletios, ABME,V, 1939-40, pp. 67-68 and figs. 19, 20. 21 Below p. 41. 22 It might be argued that any pins might have been removedin later times in the course of applyingsuccessivelayers of plaster as the paintingswere periodicallyrenewed,but the late Paul A. Underwood,whose experiencein this field was extensive,examined the walls with care and concludedthat their original surfaceswere sufficientlywell preservedto justify the assumptionthat no such pins ever existed. A handful of mosaic tesserae,found wrappedin a bit of paper in the blocking of the doorway in the north apse (below, p. 39) could hardlyhave been a survivalfrom the earliest period of the Holy Apostles since they were put there two centuries after any mosaics in the church must have been destroyed.They are undoubtedlya relic from a crumblingmosaic from some other church which a pious monk pressedinto the hands of a travelerin returnfor a few lepta..

(24) 14. THE FIRST PERIOD. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. permits speculation that it was made for the founder, whose name was perhaps Solakis, thus accounting for the persistence of the name as an epithet for the church. An elaborately worked sarcophagusfront of the Middle Byzantine period (Fig. 6 P1. 10,e) which was found not far from the church has now been placed in the position of the arcosolium.23It is well known. -- 1. IFt -. H--?. ?M.. Fig. 6.. 197'1. Sarcophagus Front, Restored Drawing. that the Byzantinemarbles of Athens traveledfar afield from their places of origin in Turkishand modern times as they were used as building material in fortifications and houses, so that attribution to specific buildings on the basis of their finding place is highly unsafe. But the good state of preservation for so fragile a piece as our sarcophagus front (its overall dimensions are 2.24 X 0.72 m., with a thickness of only 0.10 m.) suggests that it had not been moved far from its original position. We are probably therefore justified in attributing it to the Holy Apostles.24 THE ICONOSTASIS. Part of the foundation for the iconostasis was found between the two east columns. It consisted of a limestone block on a rubble bedding, on which was a narrowermarble block, badly broken. Many fragments of all the membersof the iconostasis itself: columns, epistyle blocks, and closure panels, were found built into the masonry of the two latest periods, in the loose fill in and around the church and in the building material taken from the demolition of modern houses in the vicinity. The screen apparently survived up to Period III, when the fragments first appear in the masonry of the church. The largest piece, the greaterpart of the closure panel (A; P1. 11,e), was built into the northwest corner of the foundations of Period IV together with a piece of the epistyle (E; P1. 11,c, d); a piece of the epistyle (D, 1) was built into the later phase of the northwest pier (Period III) and a large joining piece was used as part of the cover of Tomb 2 (D, 3). Another came from debris inside the church (D, 2) and a small fragment of the braided cross (D, 4) was found in the earthjust behind the iconostasis where it fell as it splinteredoff when the rest was dismantled (P1. 11,a). The largest piece of one of the columns (B) was built into the foundations of the modern iconostasis (P1. 1 , f) while a smaller piece of a column (C) came from the demolition of neighboring houses. Both pieces of the epistyle (P1. 11,a-d) were made from an Ionic architrave of the Roman period, at least 2.20 m. long. The three fasciae were visible from inside the sanctuary.All of the pieces have been built into the restored iconostasis (Fig. 7; P1. 26).. Inv. S 511. For otherarcosolia,cf. the Monasteryof the 'AyicovTTv&rcov EMME,pp. 128-129 and figs. 161-163, including ('Ovo7XoyiTorav), sarcophagusfront; also the monasterychurchof Hosios Meletios on Mt. Kithairon, of the last quarterof the 11th or the beginning of the 12thcentury,ABME,V, 1939-40,p. 59, fig. 12 andp. 62. For the date of the church,ibid.,p. 65. The carving on our sarcophagus front bears a fairly close resemblance,although not in all details, to some of the panels from the iconostasis at Hosios Meletios (ibid.,p. 105, fig. 52). 23 24.

(25) © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. THE ICONOSTASIS. Fig. 7.. 15. Iconostasis, Restored Drawing. FRAGMENTSOF THE ICONOSTASIS. A. Closure Panel. PI. ll,e. (A 4201) PH. 0.67; restored H. 0.835; W. 0.89; T. 0.09-0.10. Built into the northwest corner of the foundations of Period IV. The bottom (or top?) edge and two opposite corners are missing. Mended from three pieces. The whole surface much worn, as if from use as a paving block. Within a rectangularpanel, a large central rhomboid enclosing a circle, with circles in the corners of the rectangle, all interlaced. In the central circle, a Maltese cross. A broad flat band borders the panel at top and right; a narrowband at left. Back very roughly dressedwith a rectangularcutting in the middle. Pentelic marble. For the general scheme, cf. the iconostasis of the Katholikon at Hosios Loukas (PI. 12,d); also a panel from near the Bema church at Corinth (Scranton, Corinth,XVI, pl. 19, 10) and another in the Byzantine Museum, No. 104 (P1. 1i,j)..

(26) 16. THE FIRST PERIOD. B. Column. P1.11,f. (A 2561) PH. 0.55; diam.0.19. Builtinto the foundationsof the modem iconostasis. Thepieceincludespartof the octagonalshaftandpartsof two sidesof its squarecapital.Capitaldecoratedon one sidewitha rosettein a lyre-shapedframe;on the adjacentsidewitha rosettein a circlewith four (?)loops formingcorners.. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. C. Column. Not illustrated. (A 2562) PH. 0.27; PW. 0.10; PT. 0.19. From straymarblesin the vicinityof the church.Possiblyfrom the iconostasis. Similarto the precedingbut with only one side (as preserved)decoratedwith a rosettein a looped circle;adjacentside plainwith an attachmenthole.. D. EpistyleBlock. Pl. ll,a,b. (A 2492a) PL. 0.80; H. 0.235; T. 0.295(bottom);0.37 (top). Sevenjoiningfragments,builtinto late masonryof the church,in debrisinsideor foundamongstraymarblesin the vicinity. Bothendsmissing.Cutfroman ancientIonicarchitrave.Frontand bottomforman obtuseangle.On front,a central(?)braidedcross,projectingslightlybeyondthe face, flankedby interlacingdoublerectangles,eachcontaininga rosettein a doublecircle.A triangularleaf in each corner.On underside,a lozengeenclosingtwo palmettes,root to root; palmettesin the corners.All unitsmuch triple-bordered elongated.On back,the fasciaeof the originalarchitrave. For the braidedcrossreliefcf. the crosswhichoriginallystoodon the domeof the Theotokosat Hosios Loukas(Stikas,H.L., p. 212, fig. 105).. E. EpistyleBlock. Pl.ll,c,d. (A 2492b) PL. 1.07;H. 0.235; T.0.295(bottom);0.37 (top). Foundtogetherwith the closureslab A. Both endsand the top of the frontface brokenaway.Fromthe sameIonicarchitraveas D. On front, betweentwo convex-concaverosettebossesare threeinterlacedcircles,the centralenclosinga Maltese cross,the othersa palmettewithina palmette.To the rightof therighthandbossis the startof a palmettefilled cross. On underside,to left, a rectangularrestingsurface,followedby threeinterlacedcirclesenclosinga Maltesecross(center)andtwo plainrosettes.Onthe back,the fasciaeof the originalarchitrave. The schemeis closelyparalleledon two fragmentsof an epistylewhichundoubtedlybelong together and probablyjoin, one in the ByzantineMuseum,No. 197, the otherphotographedin the Asklepieion in 1961(P1.ll,g,h). The generalschemeof the decorationof the iconostasisis typicalof the MiddleByzantineperiod.Panels with a combinationof interlacedrectangle,rhomboid,and circleshavebeenfoundin manyplaces2 and of varyingdegreesof complexity,but these variationsare apparentlywithoutregionalsignificance and they havenot been sufficientlystudiedto affordprecisechronologicalcriteria.In the simplerforms the rhomboidis connectedto the enclosingrectangular frameonlyat its corners,whereit meetsthe frame at mid pointin each of the four sides.The circlesfillingthe comersare linkedto the rhomboidbut not to the frame,and the centerof the rhomboidis occupiedby an unconnectedrosette,as, e.g., in a panel from the Moni Petrakiin Athens(P1. l,i)26and some of the panelsin the windowsand galleryof the 25. Cf. Bulletinde correspondance hellenique,XXXIII, 1909, pp. 352 ff; also Sotiriou, Petraki,p. 111.. 26Now in the ByzantineMuseum,No. 166. Cf. Sotiriou, Petraki,pl. 49. I am indebtedto Mme. Sotiriou and to M. Michaelides. for permissionto include this and other architecturalmarblesin the ByzantineMuseum..

(27) THE FONT. 17. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. Katholikonat HosiosLoukas.In the Holy Apostlespaneleachof the cornercirclesis linkedto two adjacent sidesof the framebut not to the rhomboid,whichis connectedinsteadwiththe centralcircleas well as withthe rectangular frame.A stillmorecomplexformis represented by anotherpanelin the Byzantine Museum(No. 104)in whichall the elementsareconnectedwitheachother(PI. 1 ,j). Rosettesof assorted typesare the normalfillingornamentfor the circles.Pinwheels,as in the panelfrom the Moni Petraki, seemto be confinedto the earlierexamples.The Maltesecrossin the centerof the Holy Apostlespanel is rare,but it occursin a very similarpanelin the ambonof the basilicain Kalambaka27 and a related variantin the lunetteover the west door of the LittleMetropolisin Athens. Fragmentsof a numberof panelsof thistypecameto lightduringexcavationsmadepriorto rebuilding the refectoryat Hosios Loukas,28 whichStikasattributesto an earlierbuildingon the spot, eitherthe Theotokosor the small oratorychapelsaid to have been built by the followersof the saintsoon after his death.So far as can be madeout in theirfragmentary conditiontheirinterlaceis relativelysimpleand they differfromthoseunderdiscussionin that theyareborderedwithwidebandsof Kufesqueornament in whichStikassees a close resemblanceto the brickfriezessurroundingthe Theotokos.29. THE FONT. A marblefont, now in the ByzantineMuseum,was foundin the courtyardof the churchand can almost certainlybe attributedto the firstperiodof the church.30It is decoratedwith an interlacedcross (P1.10,d). 27. EEBZ,VI, 1929, p. 303, fig. 7. G. Sotiriou dates the churchin the 11th century. Stikas, H.L., pp. 17ff. and figs. 9, 10. 29 Cf. a similar panel with a Kufesqueband in the ByzantineMuseum,No. 323, reproducedin part in H.L., p. 22, fig. 15. 30 EMME,I, p. 78, fig. 74, and p. 79. 28.

(28) © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. THE ARCHITECTURAL TYPE T he churchof the Holy Apostlesis the workof someunknownarchitectwho combinedthe elements to createa uniquebuildingwhichstands of the centralizedplan, a tetraconchand a cross-in-square far above all the other extantchurchesof Athensin imaginationand sophistication.'His achievement with the tetraconchbut also in findinga happysolution lies not only in combiningthe cross-in-square to the problemof addinga narthexto this type. A seriesof churchesin Greecedemonstrateshow architectsweregrapplingwith the problemof combininga triconchor, morerarely,a tetraconchwitha narthexso thatthe resultwouldbe satisfactoryfrom a practicalpoint of view and at the sametime aestheticallypleasing.In none of the knownchurchesexThe simplestsolutionis represented by numerceptthe Holy Apostleswas the resultentirelysuccessful.2 the In ous diminutivethree-apsedchurchesfoundall overGreece,especiallyin opencountry. thesethere is usuallyno narthexin the senseof an articulatedpartof the building;the westvaultis merelyprolonged slightlyto give a little additionalspace,e.g. in the smallchurchof St.Nicholasin Methana(Fig.8,a).3A morecomplexexample,but stillwithoutanychangein the outlineof the groundplan,is to be foundin the PanagiaKoumbelidikiin Kastoria(Fig.8,c), in whichthe westwardextensionis set off by beingcovered witha transversebarrelvaultand is separatedfromtherestof thechurchby a wallwitha widedoorway.4 The extraspaceachievedby this devicewas necessarilysmall,beinglimitedto the widthof the vaults of the church;furtherprolongationwouldhaveturnedthe narthexinto a nave,whichwasliturgicallyundesirableat this time.5The nextstep,illustratedin the largerchurchof St. Nicholasat Platani,nearPatras, was to attacha much broadernarthex,allowingit to projectat both ends (Fig. 8,b).6That this is apparentlythe only survivingexampleof this typemaybe coincidenceor it mayhavebeenquitereasonably regardedas a failure,for the ends werevirtuallydead spacesand the clumsyappearancefrom the outsidewouldnot have encouragedemulation. A moresuccessfulapproachwas takenby the architectof the now ruinedchurchof St. Demetriosat Varasova,on the Gulfof Corinth,7a muchlargerbuilding(ca. 16.50 X 11m.; Fig. 8,d). As at St. Nicholas at Platani,a narthexwas addedagainstthe westernvault,but hereits connectionwasmadeless tenuous by continuing the north and south walls eastwardto merge with those of the north and south apses, thus avoiding the creation of awkward open spaces such as are found at Platani. The western vault of. the churchproperwaslongerthanthat at Plataniandthe largerspacesenclosedto eithersideweremade nicheon the chambersenteredfromthe narthex,eachhavinga smallsemicircular into smallrectangular east side. The nicheslookedinto the northand southapsesthroughsmallarchedwindows. 1 Combinationsof diverseplans in a single buildingare not uncommon.For a varietyof examplescf. Ebersolt,Chap. X. Ebersolt cites the Holy Apostles but of course without the then unknowncomplicationsof the west end. 2 This series forms part of Orlandos'illuminatingstudy of the triconchin Greecein ABME,I, 1935,pp. 105-120, whichincludes a section on the problem of the narthex. ProfessorOrlandoshas kindly given permission to include his plans, some of which are reproducedhere in Fig. 8. No claims are made for the actual dates of the individualbuildingsmentionedhere, but they representa logical sequence. 3 ABME,I, 1935, p. 113, fig. 7. 4 ABME,IV, 1938, p. 127, fig. 88, reproducedas our Fig. 8,c but without the later addition. 5Many of the churchesof this type have been so enlarged,but only in the 17th and 18th centuries. 6 ABME,I, 1935,pp. 112, 116 and fig. 12. 7 Ibid., pp. 105ff..

(29) © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. THE ARCHITECTURAL TYPE. 19. a c. f. d. g. 1h. i. Fig. 8. Churches in Greece. a. Methana, St. Nicholas. b. Platani, St. Nicholas. c. Kastoria, Koumbelidiki. d. Varasova, St. Demetrios. e. Athens, Moni Petraki. f. Manolada, Palaiopanagia. g. Athens, Holy Apostles. h. Gavrolimni, Panaxiotissa. i. Arta, St. Nicholas cPoba-g. (1:200).

(30) 20. THE ARCHITECTURAL TYPE. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. Althoughthe generaleffectof the churchat Varasovawas moreharmoniousthan that offeredby St. Nicholas,circulationbetweenthe narthexand the rest of the churchwas improvednot at all, for there was no communicationbetweenthe smalllateralchambers,whichperhapsservedas chapels,and the churchproper.8 An intermediate stagebetweenVarasovaandtheHolyApostlesmaybe recognizedin the churchof the Dormitionof the Virgin,popularlyknownas the Palaiopanagia,at Manolada,in Elis (Fig. 8,f; P1.12, a,b,c).9The basis of this churchis not a triconchor a tetraconchbut a free cross, with armsof equal length,the westernarmbeingsurroundedby a pi-shapednarthex.Thetwo baysflankingthe westernarm of the cross communicatewith the churchproperby doorways,this finallyprovidingfree accessfrom the narthexto the mainbody of the church.10 representsa transitionbetweenthe typesof churchesdiscussed Althoughthe planof the Palaiopanagia above, the actualbuildingmust be regardedas a late provincialexample.The large blocks forming crossesin the masonryof the west facade,for example,are characteristic of the secondhalf of the 11th andthe 12thcentury,andthe fact thattheyareoutlinedin brickmakesthe laterpartof thisperiodmore likely(PI. 12,a).11In fact, Bourasfindsso manysimilaritiesin buildingdetailsbetweenManoladaand the Naupliachurch,builtin 1143,that he believesthe two buildingsto be closelycontemporary.12 For the arrangementof the interior,Manoladaoffersthe closestparallelto the Holy Apostlesyet discovered.The generaleffect,however,is muchless pleasing.The parallelwallsof the westernarmof the crosspresentan uncompromising interruptionof the interiorspaceand createa tunnel-likeimpression (PI. 12,c), whereasin the Holy Apostlesthe diagonalwalls of the westernapse lead naturallyinto the angle chambers.The heavy walls supportingthe dome at Manolada add to the impressionof confinement. in contrastto one of lightnessandopenspacegivenby the freestandingcolumnsof the Athenianchurch, an impressionwhichis heightenedby the successivechangesin level.13 The narthexof the Palaiopanagiaprovidedthe analogyfor the restorationof the Holy Apostles,with saucerdomesoverthe threewesternbays.Theirregularshapeof the easternbaysin the Athenianbuilding called,however,for a groinvault and,in fact,a verysmallbit of the startof a groinvaultwas found in the northeastcornerof the northernbay. The similaritybetweenthe two churchesis confinedto the west end. Theheavyproportionsof Manoof the domesupportedby long vaults(P1.12,b),'4havelittlein comlada,withits forthrightarrangement of equilibrium, monwiththeharmonious"combinations ingeniousalmostto thepointof subtlety"which evokedthe admirationof Choisy15in the Holy Apostles.The obtuseanglesof all the apses and of the prothesisand diaconiconsoftenthe articulation,and the dentilcourseswhichrun aroundthe building unifyall the elements. 8 The breaksin the north and south walls of the westernvault shown on the plan apparentlyleave open the possibility of door-. ways into this part of the church,but Orlandosspecificallyexcludedthis (ibid., p. 115). 9 This interestingbuildingwas first noticed by G. Lambakis,who publishedbrief accounts in AXAE,ser. 1,II, 1894, p. 14, and again in his Memoire, p. 19. A corrected plan was published by Orlandosin ABME,I, 1935, p. 118, fig. 15, in the article under discussion(here reproducedas Fig. 8,f). Most recentlythe churchhas receivedthe full publicationit deserves from Ch. Bouras in EXOXis ToO'ApiarroTesXEou naveTorlopiou 'ETcrraMovK1Kh 'ErreTilpISTrS TToXUTEXVXiKs. IV, 1969, pp. 233-266, with eescra?ovi{Knri,. earlier bibliography. 10Bouras (p. 235) regardsthese doorwaysas "probably"a late modification,but apparentlyon no other grounds than that the TrCO alcbvcov,Athens, 1924, Si& pacroou walls are representedas unbrokenon Lambakis'plan and that G. Papandreou,in 'H 'HAeia p. 203, stated that they had recentlybeen opened. But Lambakis'plan is inaccuratein many respectsand Bouras regardedPapandreou's descriptionin generalas "unimportant."To the observeron the spot, the doorwaysshow no sign of not being contemporary with the original building. 11E.g., the Kapnikareaand Daphni in the11th century,the Hagia Moni at Nauplia and the churchesat Chonikaand Amphissa in the 12th century(Megaw, Chronology,pp. 101-102). 12 Bouras, Manolada,p. 258. For the date of the Hagia Moni, cf. Megaw, Chronology,p. 94. 13 Above, p. 5. 14 Bouras notes (Manolada,pp. 236-237) that the architect of the Palaiopanagiawas not insensitiveto the heavy effect created by intrusionof the massive walls into the interiorspace and that he alleviatedit by cutting away the cornerswhere they met under the dome to give them a concave surface. The same device was used in the church of H. Photeini in Thebes(now ruined), a church of relatedplan which Orlandosdates in the second half of the 10thcentury(ABME,V, 1939-40, pp. 145-146). 15Choisy, L'art de batirchez les byzantins,pp. 132-133..

(31) THE ARCHITECTURAL TYPE. 21. © American School of Classical Studies at Athens For personal use only. License: CC-BY-NC-ND.. Varasovaand Manoladaareessentiallyhybrids,withthe easternhalf constructedon the planof a free The three crossbut thewestern,becauseof the enclosingnarthex,becomingvirtuallya cross-in-square. the with surrounded the fourth of the Athenian the narthex,bring by Holy Apostles, projectingapses churchinto this generalcategory,with the differencethat the angle chambersand the free-standing columnsease the transitionbetweenthe two architectural typesand open up the interiorspace.Furthermore,in the Holy Apostles,sincethe westernapse rose high abovethe narthex,it is only in the ground is evident,whereasin the two provincialchurchesthe westernarm of the plan that the cross-in-square cross is completelyswallowedup by the narthexand is invisiblefrom the outside. in which The essenceof the planof the Holy Apostlesis the simpleConstantinopolitan cross-in-square all armsof the crossareof equallength,the anglechambersconsequentlysquareandthe domesupported by four free-standingcolumns.The variationconsistedin the additionof apsesto all four armsof the cross,therebyopeningup spacein all directionsand emphasizingthe centralizedcharacterof the plan.16 The apseswereaddeddirectlyto the mainvaultswithoutthe intermediatevaultwhichin the composite form makesa transitionand providesextraspacefor the sanctuary.The squareanglechamberswhich arenormallycoveredwitheithergroinvaultsor saucerdomesarehereroofedin the combinationof half domes and triangularvaults noted above,17a variationdictatedby the semicircularinteriorof these chambers. The Constantinopolitan is rarein Greecebeforethe end of the 11thcentury,in concross-in-square trastwiththe provincialtypein whichthe anglechambersarebarrel-vaulted; the resultis thatin the provincialchurchessymmetryis less of a factor sincethese spacescould be, and often were,elongatedat will.18The evolutionof the Greektype can be tracedin actualmonumentsfromits originin the basilica andthe transitional throughSkripou,the firstsurvivingchurchin Greeceto presagethe cross-in-square,19 churchesof the 10thcentury,downto the fullydevelopedbuildingof the 1lth-12th centuries.Illustrative of the 10thcenturychurchesis the PanagiaPanaxiotissaat Gavrolimni,on a ruggedmountainslopebehindNaupaktos(Fig.8,h; PI. 13,b).Althougha developedcross-in-square, it has someof the archaicfeatures of Skripou, e.g., the semicircularapse and the short stretches of wall which will later become freestanding supports.20The Constantinopolitantype, on the other hand, having no roots in Greece,2'makes its first appearance already fully developed and so appears as an intrusion and a conscious importation. Its sudden appearancecan hardly be due to pure chance. It seems more likely that some external circumstance arose to open this new channel in the currentof Byzantinearchitecturein Greece, and it is possible that the answer lies in the Monastery of Hosios Loukas (PI. 13,c, d). Without doubt the most important building of the Constantinopolitanstyle ever built in Greece22is the Church of the Theotokos at Hosios Loukas. (PI. 13,c). The recently completed work of conservation and. 16 For the relation of the Holy Apostles to the tetraconch,cf. also Orlandos in ABME,XI, 1969,pp. 81, 82. An earlier but related in Salonica(ABME,VII, 1951, pp. 146exampleof the type may be seeinin the 10th centurychurchof H. AndreasTCv epio-repcov 167), a tetraconchin which a prothesisand diaconicon,added for liturgicalratherthan architecturalreasons,are almost completely sealed off from the main body of the church.The four free-standingcolumns are so close to the inner cornersthat the buildingcan hardlybe classifiedas a cross-in-squareat all, and domes cover all arms of the cross. Orlandosdescribesthe churchas being a mixture of Early Christianand proto-Byzantineelements.. 17 Above, p. 10.. 18 The. differentiationof the Constantinopolitanand Greek types of cross-in-squarewas first enunciatedby Millet in his L'ecole grecquedans l'architecturebyzantine,1916. This pioneer work formed the basis for the study of Byzantinearchitecturein Greece. More detailedexaminationand in many cases removalof later accretionshave made possible greaterprecisionin applyinghis principles, which by now have become almost axiomatic. At the same time, they have made the distinctionsless clear-cut.The subject has been more fully explored and elaboratedby Orlandos,as summarizedin ABME,V, 1939-40, pp. 3-10. For recentdiscussionscf. Sotiriou, Petraki,pp. 101-129 and Krautheimer,pp. 275-280. The basic differencesare seen as going back to the ultimatederivation of the two types: the Greek from the Easternbasilica and the Constantinopolitanfrom the cruciformchurchbroughtto perfection in the capital. Only the elements directlyapplicableto the presentsubjectare touched on here. 19M. Sotiriou, 'Apx. 'Ep., 1931, pp. 119-157. 20 ABME,I, 1935, pp. 121-124. For other 10th centurychurchescf. Sotiriou, Petraki, pp. 107-108. 21 The cruciform church of the Katapolianion Paros is no more native to Greece than are the later churches of Constantinopolitan origin. 22 The churchesof Salonica and Mt. Athos are excludedfrom considerationas being in the sphere of the capital, not of the provinces (cf. Orlandos,ABME,V, 1939-40, p. 6, note 1)..