"1 give tie/* Books ... \fpK the. j 'diluting nf.a College m?iMs-J5oldg0. ¦YJkLE'WJMIT^EI&SflTrY" Bought with the income of the William C. Egleston Fund, 1913 THE ARCHITECTS' LIBRARY Editor: F. M. SIMPSON, F.R.LB.A. THE ARCHITECTS' LIBRARY Edited by F. M. Simpson, F.R.LB.A. Professor of Architecture in the University of London Medium 8vo. A History of Architectural Development. By F. M. Simpson, F.R.LB.A. Three Volumes. Vol. I. Ancient, Early Christian, and Byzantine. With 1 80 Illustrations, ioj. 6d. net. II. Mediaeval. With 257 Illustrations. 10s. 6d. net. III. Renaissance in Italy, France, and England. With 268 Illustrations. lew. 6d. net. Building Construction. Vol. I. By Beresford Pite, F.R.LB.A., F.T. Baggallay, F.R.LB.A., H. D. Searles-Wood, F.R.LB.A., E. Sprague, Assoc.M.Inst.C.E., etc. With 249 Illustrations. 10s. 6d. net. II. By J. H. Markham, A.R.I. B. A., Herbert A. Satchell, F.R.LB.A., Professor F. M. Simpson, F.R.LB.A., and others. [In preparation. LONGMANS, GREEN, AND CO. LONDON, NEW YORK, BOMBAY, AND CALCUTTA THE ARCHITECTS' LIBRARY A HISTORY OF ARCHITECTURAL DEVELOPMENT IN THREE VOLUMES VOL. II MEDIEVAL BY F. M. SIMPSON \>. architect PROFESSOR OF ARCHITECTURE IN THE UNIVERSITY OF LONDON ; FELLOW OF THE ROYAL INSTITUTE OF BRITISH ARCHITECTS \ SOMETIME PROFESSOR OF ARCHITECTURE IN THE UNIVERSITY OF LIVERPOOL ; ROYAL ACADEMY TRAVELLING STUDENT, 1884 WITH 257 ILLUSTRATIONS REISSUE LONGMANS, GREEN, AND CO. 39 PATERNOSTER ROW, LONDON NEW YORK, BOMBAY, AND CALCUTTA 1913 fill rights reserve^ PREFACE TO VOL. II. The aim of this volume, like that of the first, is to trace the development of architecture through the planning, construction, materials, and principles of design of the buildings described; reference being made to the influences which helped to shape that development. The volume deals only with mediaeval eccle siastical architecture. Secular buildings are not included, barely mentioned. The first half treats in detail of the parts of churches, walls, piers, arches, buttresses, windows, vaults, etc.; the second, of the churches themselves. In both, development is traced through all the centuries when Bomanesque and Gothic art flourished. My endeavour has been, whilst describing the examples in different countries, to hold the scales impartially between them all; pointing out, so far as lay in my power, the beauties and defects of each. For French Gothic art I have the greatest admiration — for the Southern even more than the Northern ; and I am the more ready to admit its superiority, in some respects, to our own, because the feeling formerly prevalent amongst French architects that Gothic was born and bred, lived and died, in France, and that English art was but the crumbs which fell from the Frenchman's table is rapidly undergoing a change. Amongst the most distinguished of recent French writers there is an increasing readiness to acknowledge the beauty of the art of this country and wherein it excels, which a quarter of a century ago was non-existent. For over thirty years I have been interested in mediaeval architecture. I have visited all the cathedrals in England, the more important ones two or three times, and most of the principal vi PREFACE TO VOL. II. churches. My first visit to France was in 1879, and I cannot remember a year since when I have not been abroad at least once, sometimes more often. Spain I do not know; but there are very few churches of importance in France, Germany, or Italy mentioned in this volume which I have not studied. I have, therefore, relied more on my notes and on my recollections of buildings, than on books ; turning to the latter mainly for dates. As regards the illustrations in the volume, several are from my own sketches, made at different times, which I have included, although conscious that they possess Little merit, because in each case I knew of nothing else which could take its place. About a dozen have been redrawn in ink for me by Mr. C. E. Power. Most of the plans, sections, etc., are necessarily from books, only one or two are from my own measurements, and some are from drawings belonging to University College. I am indebted to Messrs. J. Bilson, Talbot Brown, A. H. Kersey, and H. A. Paley, for kind permission to copy drawings made by them. All the diagrams, illustrations, etc., have been prepared specially for this volume, with the exception of half a dozen taken from Gwilt's Encyclopaedia. About one-half of the photographs reproduced are by myself, each being taken with special reference to some point which I wished to illustrate. F. M. SIMPSON. Univebsity College, Goweh Street, W.C, September, 1908. CONTENTS PAGE Preface . . . . . ... . v List op Illustrations . . . ix PABT I CHAPTEE I. Medieval Abchitectuee : Introduction 1 II. Abch Development 9 III. Columns, Piers, Capitals, and Bases ... . , . . 23 IV. Walls, Buttresses, Plinths, etc. . 43 V. Development oe Windows . . . ... ... 55 VI. Vaulting .... 70 VLT. Towebs and Spiees ... . 103 VIII. Mubal Decobation, Sculptube, Cabvinq, Stained Glass . . 116 PABT II IX. Early Chubches in Peance and Geemany, Plans, etc X. The Development of Chubch Planning . XI. Romanesque in Italy .... XII. Romanesque in Gebmany XIII. Romanesque in Southebn Peance . . . XIV. Eably Romanesque in England ... XV. Romanesque in England and Nobmandy . . . XVI. The Cathedrals of Northebn Pbance . XVII. Gothic Aechitbctube in Southebn France . . 135 143 165 190 203 232242 254 286 viii CONTENTS. CHAPTER PAGE XVIII. Gothic Abchitectube in England and Scotland .... 294 XIX. Gothic Abchitectuee in Gbbmany, Belgium, and Holland . 324 XX. Gothic Abchitectube in Italy . 338 XXI. Gothic Abchitectuee in Spain ... . . . . 349 XXII. English Parish Churches : Timber Roofs 358 Appendix. — Table of Dimensions of Typical Churches . . 373 Index . . . 379 LIST OF ILLUSTRATIONS INTRODUCTION. PIG. PAGE 1. Capital, S. Zeno, Verona 1 2. Sections through Romanesque Cathedrals 6 ARCH DEVELOPMENT. 3. Sections through Babbel-vaulted Roofs . 10 4. Intebsbcting Abches ... 11 5. Pblncipal Poems of Abches . 12 6. Teefoll and Foliated Abches 13 7. Pointed Hoeseshoe Aech, Coire Cathedral 14 8. Roman Methods, with Abchivolts ; Romanesque Method, with Label . . 15 9. Arches of One and Two Orders . 16 10. English Arch Mouldings, 1050-1170 17 11. English Arch Mouldings, 1150-1250 18 12. English Arch Mouldings, 1250-1500 19 13. French Aech Mouldings, 1050-1500 21 14. English Aech Labels, 1050-1500 . 22 COLUMNS, PIERS, CAPITALS, AND BASES. 15. Pieb and Abch Development 24 16. Peteeboro' Cathedral : Capital of Ohoib Aisle 25 17. English Columns and Piers, 1060-1500 ... 27 18. Capital oe Column surrounded by Detached Shafts, c. 1220 28 19. French Pibes and Columns, 1050-1500 31 20. Evolution of Typical Norman Capitals 35 21. Romanesque Capitals from Bologna, S. Zeno, Verona, Issoibe Cathedral (Cbypt), and S. Remi, Reims . . 36 22. Capital of Pieb fbom Choir of Tours Cathedral 38 23. English Moulded Capitals and Bases, 1150-1500 39 24. Spues 40 25. Much Wenlock Abbey Church, Shropshibe 41 26. Doorways, c. 1100-c 1400 42 LIST OF ILLUSTRATIONS. OUTSIDE OF CHURCHES, WALLS, BUTTRESSES, PLINTHS, ETC. fig. rA0E 27. The Pilgbims Chapel, Houghton-le-Dale, Nobfolk, Eeectbd 1350 44 28. Theusts and Abutments . ...... 47 29. Buttresses, 1060-1400 . . . . 48 30. Vitrei Lead Gubgoyle . . • . . 54 DEVELOPMENT OF WINDOWS. 31. Romanesque 2-Light Window; Plan, Elevations, and Section . 56 32. 3-Light Lancet Window, c. 1200: Plan, Elevations, and Detail. 57 33. Plate Teaceby Window, Chabtbes Cathedeal, and Window with Stilted Head, Reims Cathedeal 34. Head of Bab Teaceby Window, showing Jointing of Teaceby . 59 35. Window in Aisle, Stone Chubch, Kent : Plan, Elevation, and Section • • 61 36. Romanesque Wheel Window, Pathixbourne Chubch, Kent: Eleva tion and Section .... . 62 37. Development of Window Tracery ... ... 63 38. Bbede Church, East Window of South Aisle op Chancel ... 64 39. Development of Mullions . . . . .... 66 40. Cusps and Cusp Endings . . 67 41. Head of Traceried Window, showing Window Arch, Rear Aech, and scoinson arch . . .... 68 58 VAULTING. 42. Plans of Romanesque Vaulting, Continuous and in Bays ... 72 43. Ribbed Vaults, Theee Different Methods, with Main Arches Semicieculae 73 44. Winchestee Cathedral : Springing of Groined Vaults . 76 45. Oblong Vaults with Pointed Abches, etc ... 78 46. S. Geoeges de Bosohebville : One Bay of Choib Vault . . facing 78 47. Angebs Cathedral : Nave Vault facing 78 48. Sexpabtite Vaulting, Abbaye-aux-Dames, Caen 80 49. S. Pieeee, Lisieux : Quadbipabtite Vault op Choib . . . facing 80 50. Laon Cathbdbal : Sexpaetite Vault op Nave facing 80 51. Square and Oblong Vaulting . . . . . . . . 81 52. Plans of Ribbed Vaults : Durham Cathedeal Nave and Lincoln Cathedeal Choib 82 53. Methods of Laying Web ... . . 83 54. Hand-Centbe for Wbb ... . . 85 55. Exeteb Cathbdeal : Tiebcbbon Vault op S. John's Chapel, facing 86 56. Exeteb Cathedeal : Tibroeron Vaulting of Nave . . . facing 86 LIST OF ILLUSTRATIONS. xi FIG. PAGE 57. Diagrams of Rib Development 87 58. Pan Vaulting fbom Henry VTI.'s Chapel, Westminstee. ... 89 59. Ribs Detached at Speinging 92 60. Plan, Longitudinal and Teansvebse Sections of Bay of Vault ing, WITH TAS-DE-ChAEGE, ETC 94 61. Laon Cathedral: Longitudinal Section; Transverse Section . 96 62. Reims Cathedbal: Cross Section 99 63. Malmbsbuey Abbey Chubch: Flying Buttress and Pinnacle . . 100 64. Laon Cathedbal : Intebior facing 102 65. Sees Cathedbal: Interior facing 102 TOWERS AND SPIRES. 66. Coutances Cathedbal : Intbeioe of Lanteen . ... facing 104 67. Evebux Cathedbal : Intebior of Lantebn . .... facing 104 68. Bbaisne Chuech : Intebiob of Lantebn facing 104 69. S. Maclou, Rouen : Interior op Lantern facing 104 70. Towees of Theguiee Cathedral 105 71. S. Zeno, Vebona : West Front and Detached Tower . . facing 108 72. Tower of SS. Giovanni e Paolo, Rome . facing 108 73. Tower op S. Satibo, Milan 109 74. Glbculab Toweb, Pisa .... 110 75. Jumiegbs Abbey Church, fbom the Nobth-West .... facing 111 76. Coutances Cathedral : Central Lantbrn Tower .... facing 111 ¥7. Timber Spibes and Typical Gebman Spiee 112 78. Stone Spires ... 113 MURAL DECORATION. 79. Chichester Cathedeal : Tblforium, and Clerestory Passage above facing 118 80. Polebbook Chubch, Northants : Paintings on Screen .... 120 81. S. Vulfban, Abbeville: Vault ... facing 121 82. S. Jacques, Lisieux : Paintings, Vault facing 121 83. Autun Cathedbal : Open Narthex facing 122 84. S. Ouen, Rouen : Niches on Piebs facing 123 85. Chabtees Cathedbal: Two Figubes fbom Jambs of Centeal Western Doorway (fbom Casts in Tbocadebo Museum, Paris) facing 124 86. Capitals: S. Eusebe, Auxebbe; S. Vitale, Ravenna 126 87. English Carvings 130 xii LIST OF ILLUSTRATIONS. PAET II. PLANS OF EARLY CHURCHES IN FRANCE AND GERMANY. FIG. PAGE 88. Charlemagne's Church at Aix-la-Chapelle : Plans and Section 138 89. Gebmigny-les-Pees : Plan and Section 139 90. S. Cboix, Montmajoue: Plan 140 91. Chuech at Chaeboux : Plan ¦ ¦ • 140 92. Chubch of the Holy Sepulcheb, Jeeusalem: Plan 141 THE DEVELOPMENT OF CHURCH PLANNING. 93. Abbey Chubch op S. Gall : Plan 144 94. Cantebbuey Cathedral : Plan 146 95. Romsey Abbey Church : Plan .... 148 96. Church at Vignoby: Plan 148 97. Le Mans Cathedeal : East End facing 149 98. Le Mans Cathedeal: Plan 149 99. Noewich Cathedeal: Plan . 150 100. Lincoln and Salisbuey Cathedeals : Plans . 151 101. Ely Cathedbal: Intebior facing 152 102. Salisbury Cathedeal : Intebior facing 152 103. Laon Cathedral : Aisle Wall cut through for Chapels . facing 154 104. Coutances Cathedral : Chapels at Sides facing 154 105. Winchester Cathedbal: Chantby Chapel of Bishop William of Wykeham ... .... facing 155 106. N6tbe Dame, Paeis : Plan 156 107. Plan peom V. de Honnecouet's Sketch-book 158 108. Abbey Chubch op Maulbbonn : Plan 159 109. Abbey Chubch of S. Maey, Buildwas, Salop : Plan 160 110. Ebeach Abbey Chubch, Germany: Plan 160 ROMANESQUE IN ITALY. 111. S. Ambrogio, Milan : Plan 168 112. S. Ambrogio, Milan : Interior facing 168 113. S. Ambrogio, Milan : Atrium and West Front .... facing 168 114. S. Vincenzo-in-Peato, Milan : Easteen Apses 170 115. S. Michelb, Pa via : Plan and Elevation op East End .... 171 116. S. Zeno, Verona : Interior facing 172 117. S. Miniato, Florence : Intebiob facinq 172 118. S. Miniato, Plobbnce : Plan and Section 174 119. Marble Mosaics on Wall, Pisa Cathedbal 175 120. Tboja Cathedral: Lower Portion of West Fbont .... 176 121. Pisa Cathedbal: Plans yjj 122. Baei Cathedeal: Pierced Maeble Slab .181 LIST OF ILLUSTRATIONS. FIG. PAGE 123. S. Nicolo, Bari : Section and Plan 182 124. Molfetta Cathedbal: Longitudinal Section 184 125. Monbeale Cathedeal: Cloisters . .facing 184 126. Monbeale Cathedral : Intebiob facing 184 127. Moneeale Cathedral: Plan of Arch Mouldings, Abacus of Capital, and Shafts in Cloisters 185 128. Monreale Cathedral: Plan 187 129. S. Cataldo, Palermo: Interior 188 130. Cefalu Cathedral, from the North-East facing 188 181. S. Giovanni degli Erembti, Palermo, feom the South-East facing 188 ROMANESQUE IN GERMANY. 132. Laach Abbey Chubch: Plan 192 133. Wobms Cathedbal: Plan 193 134. Worms Cathedral : View of East End facing 194 135. Speier Cathedral: Transverse and Longitudinal Sections . . 196 136. Speier Cathedral : Plan 196 137. Worms Cathedral: Interior, looking West facing 196 138. Trieb Cathedral : Plan ... 199 139. S. Maria in Capitolio, Cologne : Plan 200 140. Tournai Cathedral from the South . 201 ROMANESQUE IN SOUTHERN FRANCE. 141. S. Front, Pehigueux: Plan 206 142. S. Front, Perigueux: Section across Transepts 206 143. Chubch at Souillac : Intebior 208 144. Angouleme Cathedbal: Plan 209 145. Angouleme Cathedeal: Section of One Bay op Nave .... 209 146. Fontevbault Abbey Church : Plan 210 147. S. Hllaire, Poitiees : Intebiob 211 148. Le Puy Cathedbal, Dome oveb One Bay of Nave . . . facing 212 149. S. Ours, Loches, peom the South 213 150. S. Philibebt, Toubnus, Plans and Sections 214 151. S. Nazalre, Carcassonne: Intebiob 216 152. Issoiee Cathedral: Plans and Sections 217 153. Issoire Cathedral: East Front facing 218 154. Issoire Cathedeal: Apses at East End facing 218 155. Issoibe Cathedbal: Intebiob facing 219 156. Issoiee Cathedral: Trefoil Horseshoe Abches in Tbifobium . 219 157. S. Sernln, Toulouse: Plan 220 158. Vezelay Abbey Church : Interior facing 220 159. Vezelay Abbey Church: Plan of Narthex and Nave .... 221 160. Autun Cathedral: Plan of Western Porch 222 161. Autun Cathedral : One Bay of Nobth Side of Nave and Details 224 xiv LIST OF ILLUSTRATIONS. FIG. PAGE 162. Angers Cathedbal: Plan . . 226 163. La Tbinite, Angbbs: Plan . . 227 164. S. Piebee, Saumub : Intebior . 229 165. Le Puy Cathedbal: West Front 230 166. S. Michel de l' Aiguille, Le Puy 230 EARLY ROMANESQUE IN ENGLAND. 167. Bbadpoed-on-Avon Church, Wilts: Plan 234 168. Bradfobd-on-Avon Chubch : East End . 234 169. Toweb op Eaels Babton Chubch . . . 235 170. Wall between Caen and Falaise . . 236 171. Woeth Chuech, Sussex : Plan .... 238 172. Wyckham Church: Window .... 239 ROMANESQUE IN ENGLAND AND NORMANDY. 173. Jumieges Abbey Chubch, fbom the South . . 242 174. S. Ettenne, Caen: Tbifoeium 249 175. Ouibtreham Church : Abcade and Clerestory 249 176. Durham Cathedeal : One Bay of Nave . . . 249 177. Durham Cathedral: Plan ... . 250 178. Winchester Cathedeal: Teansept . . . 250 179. Dueham Cathedeal: Inteeiob 250 180. Tewkesbury Abbey Chubch: Plan op West End 252 181. Tbwkesbuey Abbey Church: West Front . . 252 182. Durham Cathedeal : West Feont 252 THE CATHEDRALS OF NORTHERN FRANCE, 1120-1500. 183. S. Pierbb, Lisieux : Intebiob .... 258 184. Noyon Cathedral: Interior . . . 258 185. Soissons Cathedral: South Teansept . 260 186. Le Mans Cathedral : Two Bays of Nave 260 187. Reims Cathedeal : Plan 266 188. Amiens Cathedral: Plan 267 189. N6trb Dame, Paris: Interior .... 268 190. Much Wenlock Abbey Chubch, Sheopshiee : One Bay op Tbifoeium 270 191. S. Pibree, Chartres : Interior, showing Glazed Teiporium . facing 271 192. S. Aignan, Orleans : Interior .... 273 193. Bourges Cathedral : Section .... 274 194. Chartees Cathedeal : Plan 276 195. Amiens Cathedbal, pbom the South . . 276 196. Laon Cathedbal: West Feont .... 278 197. N6tee Dame, Paris : West Fbont . . . 278 198. N6tee Dame de Bon Sbcoues, Guingamp 283 199. S. Maclou, Rouen: West Feont . . . 284 PIG. LIST OF ILLUSTRATIONS. GOTHIC ARCHITECTURE IN SOUTHERN FRANCE. PAGE 200. Poitiers Cathedral : Interior 287 201. Albi Cathedral, fbom the South . . facing 289 202. Albi Cathedbal: Plan .... 289 203. Albi Cathedral: Cross Section 290 204. Chubch op the Jacobins, Toulouse: Plan, Sections, etc. . . 291 205. Chubch of the Coedeliebs, Toulouse: Plan 292 GOTHIC ARCHITECTURE IN ENGLAND. 206. Fountains Abbey Church, showing Aisle Vault restored . . 297 207. Plans op Wells Cathedral Pier and a Romanesque Pier . . 299 208. Amiens Cathedral, prom the West facing 300 209. Salisbuey Cathedbal, peom the West facing 302 210. Cboss Sections op Salisbury and Amiens Cathedrals .... 303 211. Wells Cathedral : West Front facing 304 212. Salisbury Cathedral, feom Noeth-East facing 304 213. Ely Cathedbal : East End facing 306 214. Westmlnsteb Abbey Choie: One Bay; and Lincoln Cathedral Choib : One Bay 308 215. Exeter Cathedeal: Plan 313 216. Yobk Cathedral : Choir .... facing 314 217. Ely Cathedbal : Interior op Lantern facing 315 218. Winchester Cathedral, as originally designed : Pne Bay . . 316 219. Winchester Cathbdeal, as altebed : One Bay 316 220. Bath Abbey Chubch : Cboss Section : and One Bay Longitudinal Section ... 318 221. King's College Chapel, Cambbidge : Plan 319 222. King's College Chapel, Cambbidge : One Bay Longitudinal Section ... .... 320 223. Jbdbubgh Abbey Church: Choir facing 322 224. Jedburgh Abbey Church: Nave facing 322 GOTHIC ARCHITECTURE IN GERMANY, BELGIUM, AND HOLLAND. 225. S. Yved, Braisne; and Liebfrauenkieohe, Teiee: Plans . . . 326 226. Cologne Cathedbal: Plan . . 327 227. S. Mary, Lubeck: Plan 329 228. S. Mary, Lubeck : Cross Section . 329 229. S. Elizabeth, Marburg: Cross Section 330 230. S. Peter, Lubeck : Cboss Section . . 331 231. Ulm Cathedral: Plan . . 332 232. Vienna Cathedral: Plan 334 233. The Cloth Hall, Ypbes, Belgium 336 xvi LIST OF ILLUSTRATIONS. GOTHIC ARCHITECTURE IN ITALY. PIG. PA6E 234. S. An astasia, Vbeona: Cross Section and One Bay Longitudinal Section • 340 235. Florence Cathedral : Cboss Section and One Bay Longitudinal Section 341 236. S. Anastasia, Vbeona : Interior . facing 342 237. Florence Cathedral: Plan. ... . . • • 343 238. S. Anastasia, Verona: Plan 344 239. Siena Cathedral : West Front facing 344 240. S. Ceoce, Florence : Plan ... 345 241. Milan Cathedral : Plan . . .... 346 242. Milan Cathedral : Detail op Upper Part facing 346 243. Milan Cathedral: Cross Section and One Bay Longitudinal Section 347 GOTHIC ARCHITECTURE IN SPAIN. 244. Salamanca Old Cathedeal: Intebiob and Exterior of Lantern 351 245. Gerona Cathedral : Plan . . 352 246. Barcelona Cathedeal and Cloisters : Plan .... 354 247. Barcelona Cathedeal : Intebiob . . . 356 ENGLISH PARISH CHURCHES. TIMBER ROOFS. 248. Iffley Chuech, Oxon : Plan ... 359 249. Warmington Church, Noethants : Plan .... . . 361 250. S. Pateick, Patbington: Plan . . 362 251. S. Patrick, Patbington: West Elevation .... ... 363 252. S. Peteb's Manceopt, Noewich : Cross Section and One Bay Longitudinal Section . . . 365 253. Heckington Church, Lincolnshire : Roof of South Pobch ; and Wimbotsham Chubch: Nave Roof . . 368 254. Constbuction of Hammee-beam .... . . . . 369 255. Bacton Church, Suffolk : Nave Roof . . ... 369 256. Westminster Hall : Cross Section . .... 370 257. Brampord Church, Sufpolk: Roof of South Aisle 371 A HISTORY OF ARCHITECTURAL DEVELOPMENT CAPITAL 5.ZENO, VERONA. MEDLEVAL ARCHITECTURE CHAPTBE I. INTRODUCTION. The beginning of the eleventh century marks a new era of development in architectural history. Christianity had by that time taken deep root throughout the whole of Europe, and a wave of church building was passing over all lands. The great monastic order of the Bene dictines was at first mainly responsible for the prodigious activity in ecclesi astical matters, and for many of the earlier phases of architectural develop ment. The monks of the order in the previous century naturally followed in their churches the building methods prevalent in Rome. The basilican plan, with its nave and aisles and roofs of timber, formed their model. But not for long. Before the eleventh century opened, a new movement had commenced which, once fairly under weigh, developed with great rapidity and produced lasting results. Where it originated exactly is difficult to state with any certainty. It can hardly be said to have come from the East, although church building had reached a very high level there five centuries before, and excellent work on traditional lines was still being done. Neither was it of Southern growth ; for, throughout Central and Southern Italy, Borne and Constantinople — the former through its basilican plan, the latter through its detail mainly — were still the two vol. h. b Fig. 1. Introduction. 2 A HISTORY OF ARCHITECTURAL DEVELOPMENT. paramount influences. It came from farther north ; and manifested itself partly in a symbolic treatment of ornament — based, perhaps, on Eastern traditions, but rendered with a robustness foreign to the East— and partly in a return to structural methods which had laid dormant for centuries. The intersecting vault reasserted itself, with the result that the timber roof in time disappeared entirely from large churches. The walls thickened ; columns were replaced by piers, bulky in proportion since the workmen feared the thrust of the vault, and complex in section because their functions were multiplied. The semi-barbaric symbolism, conspicuous in many of the carvings, forms a striking contrast to the refined, but, in the main, merely sensuous, beauty of Classic and Byzantine art. The west front of S. Michele, Pavia, to mention now one instance only, is covered with carved representations of figures and animals, the proportions of which, it must be admitted, are often very different from those of real life. But the execution of these and other contemporary carvings, if sometimes rude and unskilled, has a character of its own and is always forcible. Its rudeness may be forgiven, because of what its advent meant. The change indicated an infusion of vigorous new blood, and heralded the overthrow of the long paramount Classic traditions, and the approach of the great Northern art which was to prevail throughout Europe from the eleventh century to the end of the fifteenth. New The new movement found full expression first on the plains of Lombardy. Thence it spread through Germany, where the seed was already sown, leaving behind it, on the banks of the Bhine, some of its finest monuments, and westward through Southern France, where it had to contend at first with earlier traditions. Soon it reached Normandy, virgin soil almost, and crossed to England to lay the foundation of our national archi tecture, and to dot the country with fine cathedrals and abbeys and innumerable parish churches. Until the middle of the twelfth century the development of the new ideas proceeded slowly and methodically. Men were in no haste to throw over the old traditions ; in fact, in the greater part of Italy they never entirely abandoned them. Borne, both through its early churches and its still earlier Imperial monu ments, was still the quarry for ideas as much as it was a quarry for building materials. But about that time a less conservative tendency manifested itself; owing, in a great measure, to the increasing interest taken by the people in ecclesiastical matters, movement. MEDIEVAL ARCHITECTURE. 3 and to the growing power of the masonic guilds, which coincided with the gradual weakening of the hitherto overwhelming pre eminence of the monastic orders. A new era dawned. The younger generation cared little, and in certain parts of Europe knew nothing, about the models which had influenced the work of their forefathers. New forms, new ideas took the place of the old. It might almost be said that a new construction foUowed ; but this is not strictly true, since, although the methods of building piers, arches, and vaults changed somewhat, all the changes proceeded naturally in a regular unbroken sequence. The connecting links are clearly traceable between the large but simple arch of the Bomans and the smaller but more complex one of the mediaeval men, and between the early, heavy in tersecting vault and the later lighter vaults built with ribs. It is to the tracing of the gradual development in these and other parts of a building that the earlier chapters of this volume are devoted. Nearly all work in Europe from the beginning of the eleventh Divisions. century to the middle of the fifteenth in Italy, and to the end of the latter century elsewhere, might fairly be termed mediaeval, as it belongs to the most famous period of the Middle Ages, But when dealing with the examples in the different countries, in consequence mainly of the jump in progression mentioned above, it is as well to divide the work into two periods. The earlier, up to about 1150, will be termed Eomanesque, thus indicating the chief influence which moulded it, and the later by the title generally applied to it, Gothic.1 A halfway halting-place has its advantages, but it must be clearly understood that this forms no break. It is not a terminus ; it is merely a great junction to which many lines converge, and then, to some extent, branch off again. No greater mistake can possibly be made than to suppose that a hard-and-fast line separates Eomanesque and Gothic architecture. They are not two independent and separate styles ; they are parts of one great style, the greatest of all the arched styles of the world. The shape of the arch employed in different centuries has been counted as a barrier. But it is 1 The word conveys nothing as to the origin of the style, and is meaningless, but it has been in general use for long, and has become a term of endearment, although first used by the purists of the seventeenth and eighteenth centuries as a reproach. Addison describes the temple consecrated to the god of Dullness as "a monstrous Fabrick built after the Gothick manner, and covered with innumerable Devices in that barbarous kind of Sculpture." ment. 4 A HISTORY OF ARCHITECTURAL DEVELOPMENT. nothing of the sort. Eomanesque and Gothic belong to one and the same great wave of evolution, which advanced slowly at first, then, in the thirteenth century, progressed with great strides, and finally, having spent its force and passed its zenith, crumbled to pieces before the revival of learning in Europe and the con sequent resuscitation of classic art. These factors, coupled with the decay of the monastic orders, the spread of individual religious thought which culminated in the Beformation, and the restless desire for something new — a trait as marked in the sixteenth century as now — completed the ruin of mediaeval art, and brought about what is known as the Benaissance. Arrange- in the chapters immediately following this (and which with it form Part I. of the volume), which deal with the parts of buildings, their plans, proportions, construction, and detail, no division is made between Eomanesque and Gothic, because none exists. In the subsequent chapters (or Part II.), which describe the concrete examples in different countries, the halting-place is utilized ; the Eomanesque ones are taken first, the Gothic ones follow. Each country, as a rule, has two chapters devoted to it; one dealing with the earlier work, the other with the later. In some cases, however, more are given. The early churches in France, for instance, present such variety that further subdivision is necessary. Some are barrel-vaulted, some groined, others are domed ; some have naves and aisles, in others the aisleless plan is followed. The barrel-vaulted churches of Auvergneare in many respects different from the similarly covered churches of Burgundy. The early churches of Normandy and England possess traits of their own. In Gothic work the great cathedrals of Northern France are totally distinct from those of the South, and form a group by themselves, the later German cathedrals following in their wake. The English cathedrals are more varied than their neighbours across the water, and the line between the early and the later work — between Durham and Salisbury, for instance — is easily drawn. The peculiarities of Italian Gothic, if Gothic it can be called since Classic influence is always evident, divide it unmistakably from the more northern work. In the description and illustration of these different examples will be found the application of the principles described in the earlier chapters. In this lies their chief value. A student wants to know two things : how the different parts of a building developed, and the results of that development as expressed in actual examples. In Missing Page Missing Page MEDIEVAL ARCHITECTURE. 7 of the sexes was not insisted upon so strongly in the West, and it is necessary, therefore, to suggest some other reason for them in mediaeval examples. In monastic churches galleries were probably provided for the laity, as the monks at first monopolized the whole of the area below, and in cathedral churches they were possibly required to accommodate the crowds which on days of high festival overflowed the nave.1 Whatever the reason for them was, they were never universal even in Eomanesque days, and were omitted in France after the end of the twelfth century, although, in a modified fashion, they continued general in England until the middle of the fourteenth century. In the two largest Eomanesque churches at Caen, Normandy, one, S. Etienne, has them, the other, the Nun's Church, has none. In Gloucester Cathedral there are galleries over the aisles of the choir, but in the nave merely a low passage-way with insignificant arcading in front. Amongst other Eomanesque churches with spacious galleries are S. Ambrogio, Milan, S. Michele, Pavia, S. Eemi, Eeims, the Cathedrals of Ely, S. Albans, Peterboro', etc. ; and amongst later Gothic ones are Laon, Noyon, and Paris Cathedrals, West minster Abbey, etc. The Germans seem to have had a rooted objection to triforia in any form, and iu both their early and late work the spaces between the windows and the arcades are often left absolutely plain. The galleries in French churches are generally vaulted like the aisles below them ; in English churches in most cases they have merely timber roofs, an excep tion being round the choir at Gloucester, where the gallery is covered by a quadrant vault. Whether there is a spacious gallery at the triforium level, or merely a passage-way or an aisle roof, does not necessarily affect materially the design of the arcading in front. The effect produced, however, is different, because galleries always have windows on their outside walls, and are consequently light, whereas in the other case the background is dark. Each country possesses its own characteristics in triforium design, and in each these changed considerably from century to century. The developments by which in France the triforia became glazed, and in England disappeared altogether, are traced in subsequent chapters. In parish churches in England the middle storey is often Churches without 1 The galleries of N6tre Dame, Paris, are crowded now on Easter Day during triforia. High Mass. 8 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Churcheswithout clere stories. Churches withneitherclerestoriesnortriforia. omitted entirely, for the simple reason that the aisles are merely covered by lean-to roofs, which start low down and finish close under the clerestory windows (when there are any) only a few inches above the top of the nave arcades. In some abbey churches, and those not necessarily the smallest, triforia are also omitted, as in Pershore Abbey Church, but the passage-way is often retained below the clerestory, the difference being that the jambs of the windows, and sometimes other portions of them as well, are continued down to the string above the arcade without a middle storey being actually formed. Churches without aisles are also without triforia, because there are no side roofs to mask. Certain groups of churches abroad with spacious triforium galleries are without clerestories. In Northern Italy there are several, of wliich S. Ambrogio, Milan, is the most remarkable. This is a rib-vaulted church, but most of the examples of this type are barrel-vaulted. The majority are in Southern France, notably in Auvergne, and the upper windows are omitted because the vault allowed little room for them. Otherwise these churches in their sectional ordinance differ little from those with clere stories. The next type of church is totally different. The examples composing it have naves and aisles of equal or approximately equal height. The result is that clerestory windows are impos sible, and that the main light comes from the aisles. Many parish churches in England, especially those of the fourteenth century are planned on these lines, but the largest and most remarkable churches of this type are in Germany, and are known as Hallenkirchen. Under this heading may also be included a group of churches which have merely chapels at the sides in lieu of aisles, although these present special traits which are discussed later. CHAPTER II. ARCH DEVELOPMENT. The keynote of mediaeval art is arch construction. The lintel Arch con- occurs occasionally over small windows and doorways, especially struotlon- in domestic work, and in most continental churches the large entrance doorways are spanned by lintels, which, however, generally have arches over. Elsewhere the arch reigns alone. Without it the wide openings between columns and piers could not have been bridged, neither could the large floor spaces, necessary for congrega tional and ritual requirements, have been covered in a sound and satisfactory manner. Timber roofs might, it is true, have been employed, as in the basilican churches — they were to some extent, especially in England, Germany, and Italy — but the mediasval men, as a rule, like the Eomans, preferred the greater security from fire, and the more monumental effect wliich the vault affords. The dome did not travel westward. Except in some Eomanesque churches, chiefly in South- Western France, it remained in its native home, the East. The shape of the arch did not remain unaltered during the Pointed five centuries that mediaeval art flourished. Until near the middle of the twelfth century the semicircular form was general ; but it was not universal. So much has been written about the semicircular arch being the distinctive mark of early mediaeval or Eomanesque, and about the pointed arch being the exclusive property of Gothic, that it is as well to repeat, what has already been demonstrated in Volume I., that the pointed form was known to, and used by, the Egyptians, Assyrians, pre-Hellenic Greeks and Etruscans hundreds of years before the commencement of our era; and by the Copts of Egypt and their conquerors, the Saracens, some centuries before it appeared in the West. It was first employed in Europe in the barrel-vaulted buildings Pointed of the South of France, certainly before the end of the eleventh century, and possibly earlier still (Fig. 3). The old Eoman io A HISTORY OF ARCHITECTURAL DEVELOPMENT. )ate of atroduc-ion of lointed,rch. traditions were there very strong ; and one of these was the solid vault, with its extrados levelled to form two sloping sides, on top of which the outer covering, tiles, slates, or lead, could be laid direct, without the intervening protecting timber roof such as became customary later. In the majority of the eleventh century churches roofed in this manner, such as Conques (c. 1065), S. Sernin, Toulouse (c. 1096), and the group in Auvergne, the vaults are semi circular; but the builders soon perceived the advantages of the pointed form. Not only is it stronger, but it exercises less thrust, and the mass of solid material above the apex is considerably reduced. The nave of S. Nazaire, Carcassonne (c. 1096), is covered by a pointed barrel-vault, the aisles having semicircular vaults. SECTIONS THRO' BARREL VAVLTED ROOTS. SEAl-CIRCVLAR POIMTCD. Fig. 3. Elsewhere in Europe the date when the pointed arch was first introduced varies in different countries. In Northern France it first appeared between 1100 and 1120 ; in England it was not used constructionally before 1140, the approximate date of Fountains, Malmesbury, and Buildwas abbey churches. It occurs " accidentally," so to speak, in arcading formed by inter secting semicircular arches, both here and abroad, as early as the first half of the eleventh century, and hence arose the idea that the builders obtained the suggestion of the form from these arcades. But this theory has little to recommend it. Even if the pointed arch had not been used long before this interlacing arcading was introduced, the fact remains that the builders would be unlikely to abandon a form which had behind it centuries of tradition merely from a hint conveyed by an accident in decoration. Such might have occasioned a " freak " here and there, but the freak ARCH DEVELOPMENT. n would soon have died a natural death if the pointed form had not elsewhere proved itself to be a most valuable structural asset. In Sicily and in portions of Southern Italy its introduction was early, owing to Saracenic influence ; x in Germany it was late, as the conservative Teuton clung to his fine round-arched Eomanesque until the end of the twelfth century. In the greater part of Italy it did not obtain a footing until later still ; in fact, it never entirely superseded the Classic form. In the little Loggia del Bigallo, lAJTEPr ISEXTMG Fig. 4. Florence (c. 1360), and in many contemporary Italian buildings, all arches are semicircular. The steeper the pointed arch the less its lateral thrust ; Forms of and many of the early mediaeval examples are also the most acute, as though the workmen rejoiced in the new form, and were determined to make the most of its statical advantages. But this is not always the case. Some of the earliest are also amongst the flattest (having a rise only a trifle more than a semicircular arch affords), and suggest that the new form was used reluctantly, and with grave doubts regarding its appearance. It is easy to design a pointed arch so acute that it exercises no side thrust at all, only vertical pressure. Arches of such excessive steepness are not uncommon in blind arcading flanking windows in early thirteenth-century churches in England, but such extremes 1 It is impossible to say how early it appeared in Sicily ; but it was certainly in common use there before the Normans appeared in 1061, otherwise they would not have adopted it so universally in their churches in the island. 12 A HISTORY OF ARCHITECTURAL DEVELOPMENT. were rarely resorted to in arches over openings. An arch enclos ing a space approximating to or exactly forming an equilateral triangle was the general favourite until the middle of the four teenth century, and was by no means discarded entirely then. After that date the four-centred form in England, commonly called the Tudor arch, and the three-centred in France, are the most characteristic of the period. But both forms are weak, and are evidence of the decreasing vitality which marks the later work in PRI/SCIPAL, FORAXS OF~ARCHES. SEAM-CIRCVLAR.1000-1200 STILTED. FRE/SCH. \4CEATRED. 3 CENTRED. » 1350-1500 WEAK COAISTRVCTIon . Pig. 5. both countries. In England this was largely due to the " Black Death," the great plague of 1349, which swept away one half of the artisans of the country ; and in France, in great measure, to the hundred years' war, which lasted from 1346 until 1447 and seriously interfered with all building. The other semi-uncon- structional arches, the trefoil, cinquefoil, and ogee, are sometimes used alone over small windows and door's, but are more often enclosed by an outer arch. In the latter case, one appertaining ARCH DEVELOPMENT. '3 chiefly to window lights and to both pierced and sunk wall arcading, they are little more than shaped heads, and the arches are said to be " foliated," or cusped. The west doorway of Byland Abbey Church is trefoil-headed, and all the arch mouldings and the label are worked to this form, but above is a pointed dis charging arch, flush with the wall, which carries all the weight.1 FOLIATED ARCHES. Fig. 6. Stilted arches are the rare exception in England; on the Stilted continent they are the rule in all eleventh and twelfth- century aro es- work. They are not so common in Normandy as in Southern France, but in Auvergne, Aquitaine, Burgundy, etc., one seldom sees an arch that is not stilted, no matter what position it occupies or what its form is, semicircular or pointed. A good deal has been written about horseshoe arches in Horse- England which is misleading. It is doubtful if one single horse shoe arch was ever built in this country. There are many existing now in different parts, but aU were probably semicircular at first. Their present shape is due entirely to faulty construction, fires, settlements, or alterations, which have caused their haunches to bulge and their crowns to drop. The chancel arch of the Church shoearches. 1 Whether the pointed arch or the semicircular is the more beautiful is a matter of taste. The former is certainly the stronger ; but from the days of Imperial Rome down to the present time (excluding the work of the Gothic revival) it has only been in general use during three centuries, whereas architects, builders, and the public have preferred and employed the other throughout all the remaining ones. 14 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Construction of arches. Fig. 7. of S. Mary, Eastbourne, may be mentioned as an instance in which one or more of these causes have altered the original shape. Abroad there are many horseshoe arches which were built as such. After excluding the unmistakable examples in Cairo and other Eastern cities, the most numerous of European examples are on the west side of Southern Italy. There they occur over windows and doorways, as in the churches of Caserta Vecchia, Benevento, etc., and are undoubtedly due to Saracenic in fluence. In the south aisle of Coire Cathedral (Switzerland) are some extremely interesting pointed horseshoe arches, and in the triforium of Issoire Cathedral, and in other churches of Auvergne, are foliated arches curving inwards at the springing. Whence arose these forms in these parts is difficult to say. They may be owing to some lingering tradition left by the Moors when they passed through the country, or simply be due to the playfulness of the designer, who, perchance, had seen similar examples in the East (see Fig. 156).1 In Eoman arches of stone, big material is the rule, and the voussoirs in large blocks reach, in most cases, from the front to the back of the wall. The soffit is flush, or else panelled with moulded and enriched panels sunk below its plane. The mediaeval builder preferred smaller stones for two reasons : (1) they were easier to handle ; (2) an arch built with a number of small stones, which only bond with one another to a limited extent, is far more elastic than one built of a few large ones. The latter was the more important reason for preference. Elasticity was not necessary in Koman buildings where the mass of material was great, and the thrusts regular ; but in a mediaeval building, arched and vaulted in the manner which will be described later, it was of the first importance, owing to the multiplicity of thrusts and to the comparative slightness of all supports and abutments. The elasticity of mediasval buildings has proved their salvation ; but for this quality many a one would have fallen long ago. The majority of Eoman workmen were unskilled, fit only for 1 The original church at Coire is stated to have been destroyed by the Saracens. The cusped arches in S. Isidoro, Leon, Spain, are undoubtedly a reflex of Moorish Art. ARCH DEVELOPMENT. 15 rough work ; the mediaeval workman was an able craftsman who could be trusted to overcome unaided the difficulties resulting ROAYAM /METHODS, WITH ARC HI VOLTS. ROMA/MESQVE METHOD, WITH LABEL. methods. Fig. 8. from the use of small stones which were also generally of different sizes. The flat soffit of the Eomans was generally retained in Roman countries where Classic influence was strong until the twelfth ^^^ century was far advanced. It is also found in work of the first esque half of the eleventh century in England, Normandy, and other countries. The voussoirs, however, are of moderate size, and do not, as a rule, bond more than a few inches into the wall. The intervening space is filled in with rubble. This answered fairly well at first, as the wall above was either merely faced with ashlar, the core being of rubble, or else was rubble throughout; but was soon discarded in favour of a series of voussoirs on different planes, or in what is commonly called in different " orders." In this system of subordination the bottom ring of voussoirs, Subordi- or inner order, was built first, and formed the central part of the ^*h°° of soffit. It was not the full width of the wall above ; not more than half its width on an average. It carried the core of the wall, and, in addition, the outer rings, or orders. These partially rested on it, and partially projected in front of it on either side, making the arch the full thickness of the wall above. One advantage of this method of building was that it diminished the amount of centering required, as, when the first order had set, it acted as a centre and support for the rings above it. The "springer," or bottom stone of the arch, was, except in very large churches, in one piece of stone, which reached from side to side and from 16 A HISTORY OF ARCHITECTURAL DEVELOPMENT. front to back, with two or more orders often worked on it. This stone afforded a solid bed from which the smaller stones of the different rings could spring, and at the same time slightly diminished the span, and consequently the thrust of the arch. A large stone was possible here because there was no difficulty in hoisting it into position and setting it. In most English cathedrals the arches of the arcades are built in three or four rings, and the Eomanesque entrance doorway of Malmesbury Abbey Church has as many as nine. The builders in Italy and the South of France were slow to realize the advantages of subordination, and, in consequence, for some centuries they continued to make their arches with flush soffits, or, in technical language, of one order only. Even when they used two Mouldings. ARCH OF 1 ORDER WITH LABEL. ARCH OF Z ORDERS WITH LABEL. PLA/SS AT SPRI/SGI/iG. Fig. 9. orders, the lower one was by far the more important, the upper one projecting merely a few inches in front of it, as at S. Zeno, Verona, S. Ambrogio, Milan, etc. (see Fig. 15). In Central France, that stronghold of fully developed mediaeval art, the arches of the arcades of even the largest thirteenth-century cathedrals have seldom more than two orders; the reason being that larger blocks of stone were more easily procurable there than in England. An important aesthetic difference between Eoman arches and mediaeval is that when the former are moulded, which is by no means always the case, the mouldings worked on the voussoirs project in front of the wall and form an archivolt. In mediaeval arches mouldings are recessed. It is true that over the arch ARCH DEVELOPMENT. 17 is often a projecting hood moulding, or label ; but this, except in the smallest examples, is never worked on the voussoirs of the arch itself, but on separate stones in long lengths above them. Unmoulded arches are rare in England in churches built after the Norman Conquest, but examples occur on the north side of the nave of S. Albans Cathedral. In Italy and Southern France they were the rule during the eleventh and twelfth centuries, although the label over was moulded when the voussoirs were plain. Mouldings are as sure an index of the advance or retrogression of architectural art as the structural parts of buildings. Moreover, by them the date of any work can often be determined when other evidence is conflicting and unreliable.1 The scale and contours of mouldings underwent considerable change as the style developed. When the main structural lines 1050-1170. were big and strong, the mouldings were few and bold; as the former grew lighter the latter became smaller and also more numerous, until, in the fifteenth century, main lines and mouldings alike were thin and attenuated. Mouldings also foUowed in section, to a great extent, the shape of the arch in vogue at the time. Thus, when the arch was semicircular, mouldings were segments of circles. The most characteristic of early Eomanesque ones is the " torus," or three-quarter round — sometimes called the rounded " bowtell " — which, either with or without a small fillet on its face, was used alone or in combination with a cavetto, or hollow. In many English Eomanesque cathedrals the lowest order in the nave arcades often consists of a single semicircular moulding, which, as at Durham, produces a remarkably fine effect. 1 The sections of mediaeval mouldings on paper are very misleading; much more so than those of classic ones. To understand the effect- theyiproduce actual examples must be studied and measured. VOL. II. C 18 A HISTORY OF ARCHITECTURAL DEVELOPMENT. 1150-1170. 1150- 1250. When the pointed arch made its appearance, the pointed " bowtell " took the place of the rounded. For twenty or thirty years new ideas were fighting with old traditions. In the transitional work of this period one finds pointed arches which are moulded like most semicircular ones, and semicircular arches with the more deeply cut mouldings characteristic of somewhat later work. Of the latter type are the arches in S. Mary's Church, Shrewsbury. It was as though the workmen could not make up their minds to a complete change. If they altered the detail of an arch they clung to the old form; if they accepted the new form they quieted their conscience by working the old detail. About the middle of the twelfth century the sections of mouldings changed entirely, especially in England, owing largely 1250-1350. EMGLISH ARCri AV0VLDI/SG5, 1 150 1250. Fig. 11. to the general employment of soft, easily worked stone, such as Caen (from Normandy), Chilmark, Beer, Ancaster, Bath, etc. They became a series of bold rounds and deep hollows alternating, which produced marked contrasts of dark shadows and strong high lights. These deeply undercut mouldings are, however, by no means universal in English work of this period. In some parts of the country where the stone was hard and not easily worked, as in Yorkshire, simple champhers often take their place. Perhaps the most effective method of all is when, as on the west front of the Priory Church of SS. Peter and Paul, Brinkburn, Northumberland, one order is merely champhered, and the other richly moulded. The effect of the moulded order is by this means accentuated, and its richness emphasized. The mouldings of the next hundred years are much flatter, and the wave, ogee and double ogee are common forms. These are generally grouped, two or more together, each group being ARCH DEVELOPMENT. 19 separated from the next by a deep hollow where the stones of two orders meet. The ogee does not appear in English work until about the middle of the thirteenth century, although, owing to the influence of Classic tradition, it occurs much earlier abroad. In twelfth-century churches in the South of France and in Burgundy, most of the mouldings are of this section ; and the contours have a delicacy, especially in Burgundy, far greater than is ever met with in more northern examples. E/NGLISH ARCH AVOVLDIAGS 1350 - 1500 Fig. 12. When the flat four-centred arch supplanted, to some extent, 1350- the more acutely pointed arch, some of the mouldings in sympathy 1500, became wide and shallow. Fifteen-century mouldings, as a rule, are distinguishable from earlier ones by a certain angularity and hardness of outline which the others are free from. They are often also smaller in proportion. A small three-quarter round member, which forms an attached shaft and has its own capital and base, is generally met with in window and door jambs of the period. The difference in scale between these and the much larger similarly treated shafts of the late thirteenth and early fourteenth centuries is considerable. Until the middle of the thirteenth century all mouldings were Mould- worked only on the face and soffit of a squared stone. Even the V?|B on most elaborate combinations of deep hollows and bold rounds, planes. such as those on the nave arches of Salisbury Cathedral, come 20 A HISTORY OF ARCHITECTURAL DEVELOPMENT. under this category. In later work, however, the corners of the stone were champhered off before the mouldings were worked, in order to obtain the series of shallow mouldings wliich the architects of the fourteenth and fifteenth centuries preferred. Italian The same attention was not paid to mouldings in Italy as in mould- other countries. The workers there had not the love for the mgs. mason's craft which is one of the most distinctive traits m Northern art of the Middle Ages. When they wanted relief from too great a simplicity they preferred to rely on coloured decoration. In this they were only following the traditions which had existed in the country for centuries. The arches inside many a thirteenth-century church, such as S. Anastasia, Verona, and SS. Giovanni e Paolo, Venice, are unmoulded. Even when the contrary is the case, the mouldings are few and little difference is discernible between early and late examples. The absence of mouldings, however, is not due only to traditions of craftsman ship, but to the materials used and to combinations of materials. Marble is not so readily worked as stone ; and in arches built in alternate voussoirs of marble and brick, as many in Northern Italy are, mouldings are unnecessary, in fact, would destroy, to a great extent, the colour scheme. French In France it was quite different. There, as much as or even mould- more than in England, the mason's art was the first and chief consideration. On the whole, however, French mouldings are simpler and fewer than English ones, and they also underwent less change from century to century. The Frenchmen clung to the rounded "bowtell" long after it had been discarded in England. It was eminently suited to the somewhat hard, coarse-grained stone of which many of their churches are built. The deep hollows and bold rounds, so characteristic of English early thir teenth-century work, are rarely found in France, except in Normandy, where Caen stone was easily procurable. But even there the mouldings are generally bigger and fewer, as in S. Pierre, Lisieux, Coutances Cathedral, etc. In one example, how ever, the ruined church of the Abbaye d' Ardennes, near Caen, the details are practically identical with contemporary ones in England, and in Le Mans Cathedral they are very similar. After the hundred years' war French mouldings lost all their boldness, and became attenuated and weak. They are also in some cases very numerous, as in S. Maclou, Eouen. Inter- penetration, or the carrying of one moulding through another mgs ARCH DEVELOPMENT. instead of stopping it on top of the other, became general, although it was in Germany that this custom flourished most. Late French arch mouldings have little to recommend them, and contrast FRDNCH ARCH MOVLDIMGS 1050-1500. LE MA/NS LE MAMS 5.MACLOV, ROVE/N . Pig. 13. unfavourably with the squareness and simple severity of the twelfth and thirteenth-century ones. Whilst the mouldings of different countries possess their own peculiarities, contemporary ones in each country are strangely like one another. It is true that there are often localisms, dis tinguishing the work of one district from that of another, but these are, as often as not, the result of the material available. Thus, in CornwaU and also in Brittany, where the old churches are built of granite, the mouldings are naturally simpler and bigger than in those parts where a soft limestone was employed. But, local differences apart, a freemasonry undoubtedly existed. Workmen travelled from building to building, from town to town ; and wherever they went they carried with them, if not the actual templates they had used, a keen recollection of the mouldings they had worked. In no other way can the strong likeness per vading contemporary work be accounted for. This is more marked, perhaps, in England than elsewhere, as the country is small, and it was united; whereas France and Italy, besides being larger, were split up into different duchies and kingdoms. In England Similaritybetweencontemporarymould ings. Arch labels. 22 A HISTORY OF ARCHITECTURAL DEVELOPMENT. certainly, when a change in detail was introduced in one locality, time was short, as a rule, before it found its way into another. Labels, or hood mouldings, over arches, throw the wet to the sides, and prevent it from running down and injuring the mould ings and, in the case of traceried windows, the tracery below. There is, therefore, an excellent reason for labels outside a building, but there is none for them inside, unless they frame in sculpture, as in Lincoln Cathedral, or diaper work, as in Westminster Abbey. They are also useful when the walling is of rubble plastered over, as then they form a break between the plaster and the stones of the arch. But when the walling inside is ashlar and is left plain, they are far better omitted; and until mediaeval work E/NGLISti ARCH LABELS. tyvvvu^Lij ^L 1050 - 1250 WlTVVt/ 1250 - 1500. Pig. 14. became somewhat stereotyped, they often were. There are no labels over the arches inside Notre Dame, Paris, Amiens Cathedral, and the majority of large French churches, and the increased amount of plain unbroken wall space obtained by their omission is a great gain. In England, unfortunately, they are seldom absent; but there can be little doubt that the French plan is the more sensible and also aesthetically the better. The mouldings of labels, string-courses, and other minor details are always characteristic of their several periods. In Eomanesque times they were often little more than a fillet and champher; later, the forms changed frequently, keeping pace with the developments of other parts, and besides being moulded were frequently enriched by carved ornament. CHAPTER III. COLUMNS, PIERS, CAPITALS, AND BASES. Eomanesque columns in most continental work are more slender Columns. than contemporary ones in England and Normandy. The heavy bulk of the so-called " Norman " column is unknown outside the two countries mentioned. In Italy, Southern France, and Germany, many columns of the eleventh and twelfth centuries retain, to a considerable extent, Classic proportions, and not only taper, but have an entasis. These two refinements seem to have been little practised in early work elsewhere, although there are a few examples of the eleventh century in England and Northern France in which they appear. Columns in subsequent work throughout Christendom are straight-sided, and this in itself differentiates them from Classic ones. An important exception to the general rule is in the choir of the great abbey church of Vezelay, in Burgundy. Although this portion of the church was not built until about 1180, and is of fully developed Gothic, the columns round the apse diminish in diameter, and are entasised as well. English columns of the eleventh century and first half of the English twelfth are only faced with ashlar, their cores being of rubble. columns- Hence their excessive bulk, apparent strength, and real weakness ; the result often being disaster. They have no fixed or even approximate proportions, such as were universally followed in Classic work. The columns of Tewkesbury Abbey Church and Gloucester Cathedral, for instance, are about double the height of those at Hereford; and yet the diameters are practically the same (6'6 inches) in all three churches. Many English columns are ornamented by fiutings, which are vertical, spiral, zigzag, or else form lozenges. These were cut in the stones after the columns were built. All patterns occur in Durham Cathedral and Waltham Abbey Church, and the only two columns in the nave of Norwich Cathedral are carved in this way. The treatment was a relic of Classic times, when columns were generally fluted, 24 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Piers. although no doubt the masons in England knew nothing of these. They took the idea from the smaller shafts, ornamented in similar fashion, wliich surround cloister garths, especially in early con tinental churches. In Italy and Sicily the flutes or grooves in many of these are filled with an inlay of glass mosaic, a similar treatment apparently having been applied to the grooved shafts round the cloisters of Chester Cathedral.1 In Italy, cylindrical columns remained the usual support to arches all through the Middle Ages, and in France and Germany they never entirely disappeared. But the exigencies of construction called for their removal, and in all large churches they were finally entirely supplanted by that most characteristic feature of mediaeval archi tecture, the clustered pier. Piers, of course, were no novelty in the eleventh century. They had been used long before, in sixth-century churches in Syria, ARCH, PLAIN SOFFIT. PIER w> ARCH DEVELOPMENT. ARCH OF TWO ORDERS', PROM SAN AMBRO|GIO. OBLOflCr I I PIER T SHAPE PIER. /HAVE. SIDE "lAISLE "j — j SIDE. CRVCIFORrV piers. clvstereS pier: Pig. 15. and, in conjunction with columns, in S. Clemente and S. Maria in Cosmedin, Borne, and in S. Demetrius, Salonica (see Vol. I., Figs. 112, 114, 130). In other Italian churches they had been employed throughout, partly because the supply of antique marble columns, which for centuries after the faU of the Eoman empire had been so plentiful, had begun to fail, and partly because a simple pier can be built with unskilled labour, at much less cost and with cheaper material than a column. These early piers are all rectangular and unmoulded. The first alteration in their form was made in the ninth or tenth century, when aisles began to be divided longitudinally into bays by transverse arches. To carry these arches a " nib," or pilaster, was added to the back of each 1 Bound the tomb of Edward the. Confessor in Westminster Abbey are twisted shafts, fluted and inlaid, which, however, are later in date than the columns of Durham and shafts of Chester. COLUMNS, PIERS, CAPITALS, AND BASES. 25 pier, a corresponding pilaster projecting from the wall on the other side of the aisle. The pier thus became T-shape.1 The next step was to make it cruciform in plan, so that arches thrown across the nave as well as the aisles could have something substantial to spring from. In S. Miniato, Florence (c. 1013),2 the nave is divided longitudinally, into three divisions of three bays each, by piers quatrefoil in section, the front and back projections carrying transverse arches (see Fig. 117). For the changes which followed, the subordination of arches and vaulting considerations are mainly responsible. Neither a cylindrical column nor a cruci form pier is manifestly the best form to support an arch com posed of several members lying in different planes ; and both forms became still more unsuit able when, in addition to the arch, the ribs of a vault had to be carried as well. The unsuitability of a column to perform the first of these func tions in a satisfactory manner is manifest in, for instance, Gloucester Cathedral and Tewkesbury Abbey Church, where no attempt has been made to bring arch and column into harmony. In the choir of Peterboro' Cathedral, where some of the supports are columns and others twelve-sided piers, the capitals have been designed to agree with the orders of the arches above and the ribs of the aisle vaulting, but they fit somewhat awkwardly on to the supports below. The difficulties were the same when the pier was octagonal, as in Christ Church Cathedral, Oxford. To get over these difficulties and to reconcile all parts, the subordinated, or clustered pier was devised. It is so planned that PETERBORO' CATH: CAP: CHOIR A15LE. Fig. 16. Clusteredpiers. 1 T-shaped piers were used in Syria in the sixth century, in the church at Boueiha (see Vol. I., Fig. 134), for much the same reason as they were adopted in later Italian churches. * Some archaeologists place S. Miniato at the end and not at the beginning of the eleventh century. 26 A HISTORY OF ARCHITECTURAL DEVELOPMENT. each order of an arch, and, in early vaulted churches, each rib of a vault as well, has its corresponding member in the pier below. Nowhere can its development be better traced than in England and Normandy, as the architects of the Isle de France were by no means unanimous at first in adopting it, neither did they create so many varieties of it as we did. A clustered pier is formed by the addition of shafts to either a column or a rectangular pier. The simplest form, a somewhat incomplete one, is when one or more shafts are attached to the aisle side of a column to support the side vaulting, whilst a like shaft on the nave side is carried up the wall to the underside of the vaulting ribs, or, when the church is not vaulted, to below the tie-beam of the roof. In S. Albans Cathedral the piers on the north side of the nave are amongst the plainest of piers which are not simply rectangular. They have no attached shafts, but the angles are rebated to agree with the orders of the arch above. Lastly, comes the complete clustered pier, in which some of the attached shafts correspond with the orders of the arch above, whilst others support the vaulting ribs of nave and aisles. The result may be a nearly square form, or an oblong with its length from east to west. In Durham Cathedral large clustered piers alternate with cylindrical columns (see Fig. 179), the former carrying the main ribs of the nave vault. This alterna tion of large and small supports was a common one in Northern Italy and Germany in particular, and the reasons for it will be discussed later. English About the middle of the twelfth century all parts of a building 1150- became lighter, and the piers gradually lost their bulky pro mo, portions ; they also assumed a more clustered form. Typical examples, such as those in the naves of the abbey churches of Jedburgh, N.B., and Much Wenlock (c. 1180) (see Fig. 25), con sist of a series of attached shafts, in section alternately round and pointed, each of the rounded shafts having a fillet on its face. A circle drawn round them touches the extreme point of each shaft. Plain columns and octagonal piers continued to be built, especially in parish churches, but the most characteristic form between 1180 to 1250 was the column surrounded by detached shafts. This was employed almost universally in the South of England, and occasionally in other parts of the country, the shafts being generally of Purbeck marble from Dorset, as in Westminster Abbey and Salisbury Cathedral, etc. Elsewhere other marbles were used, notably in the Chapel of the Nine Altars in COLUMNS, PIERS, CAPITALS, AND BASES. 27 Durham Cathedral, where those in the responds are of Frosterley marble. They are always in long lengths, 10 to 12 feet long, and E/SGLISM COLV/n/SS AAD PIERS w 1060 - 1150 1060 - 1150 1150 - 1300 RE5POMD. <1250- 1350 1350 -1500 Fig. 17. where one piece joins another they are connected either by a moulded band of the same material, as in Westminster Abbey, or by a narrow ring of bronze, as at Salisbury. Examples of both 28 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Weight on shafts. methods occur in the presbytery of Worcester Cathedral. The moulded band sometimes merely grips the shafts and stops against the central column, sometimes it is carried round the column as well. The column can be of the same material as the surround ing shafts, but built up in drums independently of them, as at Westminster Abbey and Salisbury Cathedral, where columns, shafts, capitals, and bases are all Purbeck ; or it can be of stone, with shafts only of marble, as at Worcester. In the presbytery of the last-named the abacus of each capital and the top member of each base are also of marble, and the contrast between the two materials is very striking. The shafts generally touch the central column, but at Chichester Cathedral they stand 8 inches away from it. Their number varies considerably ; there are four round each column in Salisbury nave and Westminster choir, eight at Ely, Lincoln, Worcester, and Salisbury choir, ten at Lichfield, and they are sometimes so numerous as to hide completely the central column. This is not always circular ; it can be octagonal, as in the choir of Lincoln, or quatrefoil as at Salisbury, but the surrounding shafts are always centred on a circle. The detached shafts of the above piers had little structural value ; in fact, at Tintern Abbey they have all disappeared, and Ml —1 "th 1 __ '"*' If lift /"*— > ll CAPITAL OF n: COLVMM / SVRROWDED BY DETACH - r\ ED SHAFTS ^^^ jS£— ~_ f/9 ^ C.1220 idC a JjSp ~~&T Pig. 18. yet the piers stand almost as strong as ever. The builders of the thirteenth century were careful that as little weight as possible should fall on them. The upper members, at least, of the capital always followed round both shafts and pier, and, whenever size COLUMNS, PIERS, CAPITALS, AND BASES. 29 prevented their being worked on a single stone, the separate stones were thoroughly well bonded together, so that the capital acted as a corbel to some extent, and transmitted all pressure on to the central support. Notwithstanding this, many shafts have split or been thrust out of the perpendicular, and have had to be replaced by others, or refixed; the reason being that columns built in drums with many joints are likely to settle more than long shafts with few joints, and, in consequence, weight has fallen on the shafts which they were never intended to bear. The bands or rings allowed, it is true, a certain amount of " play," as the shafts could be somewhat loosely fitted between them, but the many restorations wliich have been necessary show that the allowance made was often insufficient. After 1250, detached shafts were abandoned, possibly because Piers, it was found that, notwithstanding the precautions taken, they ^||^~ were structuraUy faulty, and a return was made to a cluster of attached shafts. For small columns a quatrefoil plan, either with or without a fillet on each face was usual, and this was followed by a somewhat similar form, with shafts slightly ogee in outline and a deep hollow between each. The large piers consist of a series of shafts, eight in number at Lincoln and twelve at Exeter, which are placed diamond-wise, and not enclosed within a circle, as in earlier work. The moulded bands uniting the shafts dis appear ; there was no longer any reason for them, now that all shafts were attached and the whole pier built in drums. In the naves of Westminster Abbey and Worcester Cathedral, however, which were built a century or more after the choirs, they are retained, in order that the later piers should be in harmony with the earlier ones. In the work of the next hundred years a marked change is Piers, perceptible in all the details of churches, and nowhere is this 150^ more marked than in the piers. The builders had gained the confidence which successful experiments give, the construction was more skilful, and all the parts of churches much more slender. But although their skill was greater, the vigour which characterized the earlier work was lacking. The cry was for spaciousness; for a saving of material wherever it could be saved ; for small supports and wide spans. The rubble core was definitely abandoned, and the piers, and in many cases the walls also, became of ashlar throughout. Moreover, many piers were made oblong — the greater dimension being from north to south and the less from east to west — in order to interfere with the floor 30 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Responds. Conti nental piers. space as little as possible and yet be sufficiently wide to carry the wall above. Of this character are the piers in Cromer and Lavenham Churches and in many other parish churches through out England. The most general forms, however, are the octagonal, and a pier consisting of four attached three-quarter shafts with mouldings in between. The shafts alone have capitals, the mouldings continuing above the springing line and round the arch without a break. Many piers are panelled all over with small panels, the panelling generally continuing round the arch as well, as in Sherbourne Abbey Church and in the better known and most elaborate of all examples, Henry VII.'s Chapel, Westminster. It is a joinery treatment, and has little to recommend it. When extended to walls as well as to piers and arches, as in Henry the VII.'s Chapel and the Houses of Parliament, Westminster, it defeats its own object; as buildings are only rich or plain by comparison, and when there is no plain wall space to accentuate the richness of other parts, the elaboration of these is labour lost.1 In the central columns of an arcade, the lateral thrust of one arch is counteracted by that of another, but at each end additional abutment has to be provided. This, in mediaeval work, is partially afforded by external buttresses (see Chap. IV.). These, however, could seldom be made of sufficient projection, and hence the custom arose of building inside a church, at the ends of an arcade, a piece of wall known as the " respond." The respond, unlike the columns or piers, is generally the full width of the wall above, and differs from them in section. In parish churches it is often merely champhered in one or more planes, according to the number of orders in the arch above. Priority for English workers is not claimed in the above brief account of pier development, especially for the forms employed before 1200. The clustered pier found its way into England from Normandy, where it occurs, fully developed, in the Church of S. Etienne, Caen, commenced 1066, an earlier example than any we can boast. Elsewhere on the continent there are many examples of it in eleventh-century work. In the nave of S. Ambrogio, Milan, built probably in the latter half of the century, the large piers are clustered, the small ones cruciform. On the other hand, in the fine Eomanesque work in Germany the 1 The Westminster buildings possess one great advantage ; in one, the plain mass of Westminster Hall acts as a foil to the Houses of Parliament, and the choir of the abbey ohurch to the chapel in the other. COLUMNS, PIERS, CAPITALS, AND BASES. 3i piers are extremely simple, and whilst columns are general, clustered piers are rare. In S. Michael's, HUdesheim, a church of the early eleventh century, every third support is a plain rectangular pier, the intermediate ones being columns. In S. Maria in Capitolio, Cologne, the nave has rectangular piers, with a nib at the back of each to support the aisle arches, whilst in the choir and transepts are tapering columns with an entasis. Even in the large Ehenish churches of the eleventh and twelfth centuries, Speier, Worms, Mainz, the columns are plain rectangles, or T-shape, or cruciform. In France the builders found it difficult to abandon the French column, to which tradition had accustomed them. Even in Piers- churches otherwise so far advanced as Notre Dame, Faris (c. 1160), and Laon Cathedral, they are used in the main arcades. There was not great building activity in Central France before the ALBI. CHARTRE5. CMARTRC3. CHARTRES. ERE/NCIi PIERS AAD COLVM/MS,, 1050 - 1500. RrZIA\S. Fig. 19. middle of the twelfth century ; but in some of the few early examples, of which S. Germain des Pres, Paris, is one, there are cruciform piers, of the type common elsewhere in the country, consisting of a square with an attached column on each side (A). 32 A HISTORY OF ARCHITECTURAL DEVELOPMENT. This developed into the most characteristic of all French forms, which was employed all through the thirteenth century, and sometimes later. In it a column is surrounded by four attached shafts, which are sometimes semicircular, but generally three- quarter round (B). Of this character are the piers of the Cathedrals of Beims, Amiens, Tours, Bordeaux, etc. In Chartres Cathedral some of the piers are octagonal, whilst others are circular. The former are surrounded by four attached three- quarter round shafts, and the latter, by four attached octagonal shafts; an excellent instance of the good effect which can be produced by counter-changing (C and D). The French, outside Normandy, did not take kindly to detached shafts ; partly because they had no material so suitable as Purbeck marble, and partly, perhaps, because they regarded them as not altogether sound structurally. Still they used them occasionally, in some of the aisle columns of Notre Dame, Paris, and in Laon Cathedral for instance, whilst in Normandy, at Eouen, Bayeux, Coutances, etc., they are more plentiful. In the nave of the fine abbey church at Vezelay, Burgundy (c. 1126), the piers are a modification of the S. Germain des Pres type, which brings them one step nearer to the clustered pier of Normandy and England (E). Other piers in Burgundy, and in the South of France generally, have a special character of their own, owing to Classic traditions. Instead of attached shafts, they have fluted pilasters, as in the nave of Autun Cathedral (c. 1132) (F). In Auvergne are many interesting early piers which are somewhat different from those generally met with elsewhere. In the nave of Le Puy Cathedral, and in the Church of Polignac, close by, are cruciform piers with four de tached shafts in the corners, which stand clear of the arms of the cross (G). Still more effective and unusual are the piers at the crossing in the former church. In these a detached shaft is recessed on each western side with particularly happy results. ,. In Sens Cathedral (c. 1150) large clustered piers alternate with pairs of columns, one column being placed behind the other. This doubling of columns was a common device at the east end of French cathedrals with chevets, especially in Nor mandy, where it was important that the supports should be narrow and yet of sufficient depth to carry the wall above. In Bayeux and Coutances Cathedrals, double columns, which, how ever, are attached to one another at the back, are employed in COLUMNS, PIERS, CAPITALS, AND BASES. 33 this position, with small detached shafts partially recessed between them (H).1 There is not much difference between thirteenth and fourteenth century French piers, except that the latter are somewhat thinner and have often a greater number of attached shafts, as in the naves of Bourges Cathedral, S. Ouen, Eouen, Tours Cathedral, etc. In fifteenth-century work the supports, when not cylindrical columns or octagonal piers, have a number of filletted and moulded shafts worked on their face, as in S. Maclou, Eouen (J), which in most cases are continued round the arches without the intervention of capitals, as in S. Vulfran, Abbeville (c. 1490). In continental churches of the late fourteenth and fifteenth Canopied centuries, figures are frequently placed against piers on the nave ^g^ses on side. They stand on corbels, with canopies over them, about halfway between the floor and the springing of the arches of the arcade (see Fig. 84). Many of the corbels and canopies are delightful in design in themselves, but it is a question whether they and the figures are desirable additions. Their richness catches the eye unduly. In most churches the original figures have long since disappeared, and it is therefore difficult to say whether there is anything in the idea. When figures have been added in Classic times, as in some of the Belgian churches, there can be no doubt that the effect produced is a bad one ; but that may be entirely the fault of the figures. S. Ouen, Eouen, is one of the earliest churches in which these canopied pseudo- niches appear, and it is also one in which they are managed more effectively perhaps than in any other example. An important point in mediaeval building is the relative thick- Relative ness of arch and wall above to the supporting column or pier ^}ith 0* below. This depends, to some extent, on whether naves and wall. aisles are vaulted; but not entirely. In an un vaulted church the rule is that the wall is wider than the support below, and the two are brought into harmony by the capital, as described later. There are a few exceptions to this, but not many. In vaulted churches there is by no means such uniformity; and French methods are very different from English. In the majority of English examples, the plan is the same in vaulted as in unvaulted churches. Ln Salisbury Cathedral, for instance, the wall is wider than the central column and detached shafts combined (see 1 The columns of the chevet of Canterbury Cathedral are double, owing to William of Sens, who was responsible for the design. VOL. II. T> 34 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Fig. 17, A), and a similar arrangement was followed in those later examples, such as Exeter Cathedral, in which the piers consist of a cluster of attached shafts. In all typical large vaulted French churches, the opposite is found; the support, whether simple column, as in N. D., Paris, or column with attached shafts, as in Chartres Cathedral, is considerably wider than the arch and wall above. In Westminster Abbey there is an approach to a mixture of French and English methods (see Fig. 17, B). The supports are the same design as at Salisbury ; but the wall above is only slightly thicker than the central column alone, the detached shafts pro jecting in front of it. In Salisbury Cathedral the transverse arches of the aisle vault have a small bearing on top of the capital, whilst the diagonal ribs die against the wall. In Eeims and Chartres Cathedrals, the corresponding arches and ribs spring from the attached shaft which forms part of each pier, and from a portion of the pier itself (see Fig. 19, Beims). At Salisbury the vaulting shafts on the nave side start high above the column. At Chartres and Beims the shafts start from the capital of the column, and are supported partly on the cylindrical portion of the column, and partly on one of its attached shafts. The thinning of the wall in French churches has no bad effects in early work — in fact, there is much to be said for it — but in later examples it gives that wirey look which is so unsatisfactory in S. Ouen, Eouen, Auxerre Cathedral, etc. There is plenty of strength in the piers, because in these churches the vaulting shafts run down to the floor ; and the walls are strong enough because they are buttressed, so to speak, by these shafts, but the arches, in nearly all instances, look deplorably thin. Capitals. In plain rectangular piers, capitals may be dispensed with, as the pier is generally the same width as the arch and wall above ; and although a moulding is advisable to mark the springing line, this is purely an ornament. In the pre-Norman Church of S. Martin, at S. Albans, and in many early examples abroad, a simple abacus is all that parts pier and arch. In a few fourteenth- century churches, and in many of the following century, capitals are omitted entirely, because the plan of the pier agrees approxi mately, often exactly, with the section of the arch above. But a column supporting the portion of a wall at the springing of an arch, wliich is wider in all directions than the column, must have a spreading capital.1 1 In the naves of Gloucester Cathedral and Tewkesbury Abbey Church, the columns are so huge that little more than an abacus suffices. In some semi-Gothic, COLUMNS, PIERS, CAPITALS, AND BASES. 35 In many of the Byzantine churches of the sixth century, and Dosserets. in some of the early Basilican ones in which antique columns and capitals were largely re-used, the difficulty of starting an arch from the top of a column led to the introduction of a "dosseret," or shaped block, between the arch and the capital proper (see Vol. I., Fig. 107, p. 171). The dosseret of this form is confined almost entirely to Byzantine churches, and to early churches in Italy in which Byzantine workmen were employed, or in which the local masons copied Eastern methods. There is, however, a pseudo- dosseret, or upper abacus, which, in the eleventh century and early part of the twelfth, was largely used, not only in Italy, as in Pisa and Torcello Cathedrals, etc. (see Vol. I., Fig. 126, p. 192), but also in the South of France, especially in the Eomanesque churches of Burgundy and Auvergne. This upper abacus, a square-sided moulded slab, occurs most frequently above pseudo-Corinthian capitals. These have their own abaci, which, however, following the classic original, are curved in plan, and are not therefore the most suitable shape for an arch to spring from. In more northern Eomanesque work the dosseret is discarded Roman- entirely, and the arch starts direct from the top of the capital, as capitals EVOLVTIOA1 OF TYPICAL riORMAA CAPITALS Fig. 20. it always should do. The architects of S. Sophia, Constantinople, in the sixth century, had the good taste to recognize this, and to dispense with the unnecessary dosseret in the most important parts of their church. The capital they designed became the model for one half of the Eomanesque capitals of later date. In semi-Renaissance churches, such as S. Etienne du Mont, Paris, there are no capitals at all, and the mouldings of the arches die into the round of the columns below — which are the same width as the wall over — -as best they can. 36 A HISTORY OF ARCHITECTURAL DEVELOPMENT. The capi tal a corbel. Convexcapitals. it the portion below the abacus is convex, like the echmus moulding of the Greek Doric order. Other Eomanesque capitals derived their forms from ancient Ionic and Corinthian capitals, and, apart from differences in treatment, are easily distinguishable from those of the first-mentioned type by their concave outline. A capital above a column really partakes of the nature of a corbel, or rather a series of corbels branching out on all sides. A square block of stone or marble is taken, the upper part moulded to form an abacus, and the corners of the lower part rounded off so that it shall " sit " satisfactorily on the column below. When the capital is of any size, the abacus is worked on a separate stone. The result, in both cases, is a capital of convex outline. If the concave form is required, rather more stone has to be cut away in the lower part, but the method followed is the same in both instances. There can be no doubt that the convex form proclaims better than the concave the functions of a capital ; namely, to support the F.M.S. ISSOIRE CATH : CRYPT. 1904. 3. RE/AI , REI/-\S.VS96 Pig. 21. arch and transmit its thrusts to the column below. It is no wonder, therefore, that it was the favourite with Eomanesque builders, whose chief aim was to find the most suitable form for every part, and then to make it. The convex capital appertains to all countries, but each has several varieties. In Normandy, England, and Germany it is generally somewhat clumsy in form — COLUMNS, PIERS, CAPITALS, AND BASES. 37 the term often applied to it being the '' cushion " capital ; but in Italy and Southern France there are many charming examples of it of greater delicacy. A modification of the English variety consists of a number of cone-like forms divided from one another by sharp little ridges. This recalls strongly the egg-and-dart ornament of Classic work from which it was probably derived. Connecting links between the two occur frequently in Italian carvings of the ninth and tenth centuries in which deliberate attempts have been made to copy this best known of Classic ornaments. Most of the capitals with a concave curve below the abacus Concave are more or less correct imitations of the Eoman Corinthian. The capl a s- measure of correctness depends on the strength of Classic traditions in the locality. Thus, in Italy and Southern France the Eoman esque capitals are distinguishable only from antique ones by a certain coarseness of workmanship; whereas, farther north, the differences are so great that it is sometimes difficult to trace any resemblance. In work in Central France, even as late as the thirteenth century, the capitals are modified Corinthian, but other leaves are carved instead of the acanthus, and the form and mouldings of the abacus are very different from those in Eoman examples.1 Each shaft of a clustered pier requires its own capital. These Subordi- capitals are sometimes quite detached from one another ; at other capital ™ times they are joined together. In the latter case there is one capital composed of several members, which correspond with the divisions in the pier below and in the arch above, and the principle of subordination is preserved throughout. When columns and not piers support arches of more than one order, harmonious relation throughout is more difficult. At Tewkesbury and Gloucester none can be said to exist. In the choir columns of Peterboro' Cathedral an attempt has been made, and not unsuccess fully (see Fig. 16). The design, however, illustrates the difficulties, and shows how wise the mediaeval men were in employing, as a rule, the clustered pier with the subordinated arch. The plan of the abacus, or upper member of a capital, changed Plan of considerably during mediaeval times, and nowhere were the changes abaous- so numerous or are so clearly defined as in England. The abacus of a Eomanesque capital in all countries is square in plan, because 1 For further particulars of carved capitals, both Romanesque and Gothic, see " Carving," pp. 127-31. 38 A HISTORY OF ARCHITECTURAL DEVELOPMENT. the orders of the arch above are rectangular. In most continental work the square abacus continued general for many years after Gothic architecture was fully developed, the only change being that the corners were sometimes champhered off. In the typical French column surrounded by four attached shafts, the capital, in most cases, is considerably deeper round the column than round the shafts, as in the choir of Tours Cathedral. In England, when the mouldings of arches changed their character towards the end of the twelfth century, and became a series of bold rounds and deep hollows contained within a segment of a circle, the abacus and the mouldings below it changed also, and the whole became circular in form. Capitals circular in plan are not uncommon in Northern France, especially in Nor mandy, but they are not so characteristic a feature of French Gothic as they are of thirteenth-century English. It is a question if the change was really an advantage. Columns and arches are brought no doubt into more absolute harmony; the dif ferent parts of a church blend together more thoroughly ; but it is doubtful if this completeness and harmony are sufficient compensation Fig. 22. for of the greater Capitals,1150- 1500. the loss boldness and virility of earlier work. In English thirteenth- century architecture there are no sharp angles anywhere inside a church, and no strong contrasts. Column, shaft, arch mould ings, and capital, all are rounded, and the capital no longer breaks the vertical lines of pier and groining shaft as it did before. Moreover, the lower part of the capital loses its convex outline and becomes concave in all cases; its apparent strength and value of a support being thus considerably impaired. From 1150 onwards, the concave, or "bell "-shaped capital remained the characteristic form everywhere except in Italy and COLUMNS, PIERS, CAPITALS, AND BASES. 39 Spain.1 In France the builders wisely retained the proportions of the earlier capitals— with slight reduction— until the fifteenth century. Between early thirteenth and late fourteenth- century French capitals there is no great difference. The abacus in the latter is sometimes octagonal in plan, and the mouldings thereof slightly different from earlier ones ; the carving round the bell is stiffer and less bold, but otherwise there is little change. In England, on the other hand, the progression — if progression it E/NGL15M AOVLDEID CAPITALS & BASES. 1150-1250. 1250-1350. 1350 - 1500. Fig. 23. can be called, because it is very doubtful if the changes made were improvements — was continuous. About the middle of the thirteenth century the top member became octagonal, the lower members and bell remaining circular. A hundred years later all parts, even the bell itself, were made octagonal ; the slight angularity which resulted being more in keeping with the sharp arrises of the arch mouldings. As time went on, capitals became 1 These two countries never thoroughly assimilated Gothic methods, and never entirely threw over Classic ones, so it is no wonder that the developments elsewhere affected them but little. 40 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Much Wenlock. Door ways. smaller and smaller, until, in the fifteenth century, they lost all importance, and were often omitted altogether. Bases to columns and shafts followed in all respects the same evolution as capitals. Eomanesque bases are either copied exactly from classic ones, or else are modified versions of them. The upper moulded members are circular in plan and they rest on a square die, the two parts of the base being often connected at the Fig. 24. corners by carving. These bits of carving, known technically as " spurs," are more general in French work than in English. They are a sound, practical addition, as weU as being a beautiful one, as they emphasize the fact that the two parts of the base are worked out of the same piece of stone. When the top members of the capital became octagonal in plan, the lower members of the base followed suit ; and when the octagonal plan throughout became general for capitals, the whole of the base was made octagonal also. Fig. 25 is one bay of Much Wenlock Abbey Church (c. 1180), in which the three orders of the arch agree with the shaftings of the pier below. The drawing shows the relative positions of arch and pier mouldings, the projections of capital and base, and the corbel-like character of the capital. There is no need to trace the development in design in door ways and other minor arched openings, because this followed on similar lines to that described above for main arches and piers. Arches of doorways are semicircular or pointed, sharp or flat; shafts are attached or detached ; mouldings are large or small, numerous or few, according to the period. The capitals and bases of shafts in the jambs of openings are reproductions in miniature of the larger capitals and bases of the interior. The same development proceeded simultaneously throughout, and this can be traced as easily in the smallest fitting as in the most important structural feature (Fig. 26). , , VAVLTt^G SHAFTS — -T-» , WE/NLOCK ABBEY Cti: SHROPSHIRE. ARCH ¦"" ri * ""mOVLDI/SG PLAM OF PIER 1/iLOOKI/SC DOW/S VZLOOK1/SG VP 1 O 2 ? 6 i» fa" 0 1 2. SCALES OTIilil I I I I I I AMD I i I i I I IFEET FOR ELEVATIO/S FOR PIER & ARCH Fig. 25 (G. Langshaw). 42 A HISTORY OF ARCHITECTURAL DEVELOPMENT. DOORWAY5 Pig. 26. CHAPTER IV. OUTSIDE OF CHURCHES, WALLS, BUTTRESSES, PLINTHS, ETC. Stone was the favourite material of the mediaeval builders, Wall although marble, brick, and, in certain parts of England, flint were all used for walling. In Northern Italy a very fine effect is often produced by courses of marble, alternating with four or five courses of brick which together are about equal in height to one course of marble. In Central and Southern Italy walls are often faced entirely with marble, which in some cases is merely an applied veneer, but in others is structural. In Southern France there are many fine churches of the eleventh and later centuries built entirely of brick; and brick was the material generaUy employed throughout the Middle Ages in Northern Germany, HoUand, and Belgium, owing to the scarcity of stone. In England, the art of brick-making, which had flourished during the Eoman occupation, had died out, and was not reintroduced until the four teenth century. Most of our brick churches are a century or two later, and they are chiefly in the Eastern Counties, Norfolk, Suffolk, and Essex, where a close intercourse with the Low Countries was maintained.1 Churches with walls built and faced with flint, are most common in Sussex; two excellent examples being at Old and New Shoreham. Flint was also largely used in Norfolk and Suffolk, sometimes in its natural rounded form, and sometimes, especially in the fifteenth century, split smooth, and squared. In the latter case it is generally framed in by stone, and is merely a panel, as in the Chapel of Houghton le Dale, Norfolk. In the eleventh century the craft of the mason was more Stone advanced in some countries than in others. Italy naturally came first, as it had behind it centuries of building tradition. England was perhaps the country most behindhand of any, although in 1 The tower and other eleventh-century parts of S. Albans Cathedral are built chiefly of old Roman brick taken from the ruins of Yerulam close by. walls. TME PILGRIMS CMAPEL, MOVGMTO/N - LE- DALE . /NORFOI K ERECTED C 1550 doorway 12 6 O I 1 I I I I L ARCM SCALE or FEET for DETAILS. i W1/ND0W BVTTRE55E5. C. 1250-1400 Fig. 29 OUTSIDE OF CHURCHES, WALLS, BUTTRESSES, ETC. 49 When a buttress is in different stages, these are separated from stages of one another either by sloping weatherings, or set-offs, or else by steep-pitched gables which die against the face of the stage above. There is no rule to regulate the slope of the set-off. This differs considerably in different examples. It might have been thought that the top slopes would be steeper than the lower ones, because, being more above the eye, they would become more foreshortened. They are made so occasionally, but by no means universally. In some buttresses the top slopes are the flattest. At the top the buttress finishes much the same way as each stage finishes. When it is crowned by a gable, this may be below the parapet, or level with the top of it, or be carried considerably above it. The advantage of the last arrangement is that a considerable mass of masonry is poised above the haunches of the arch and vault (although not necessarily immediately over them), and its weight, exercising a vertical pressure, helps considerably to neutralize the side thrusts (see Fig. 28 and Fig. 29). This is the reason for the pinnacles which play so important Pinnacles. a part in Gothic architecture. They are not merely ornaments ; they are, in some cases, absolute necessities, and are standing proof of how the mediaeval architect converted a structural requirement into an ornamental feature. It is true that they are sometimes introduced in fifteenth-century work when the need for them does not exist (for instance, in churches with flat timber roofs which exercise little or no lateral thrust) ; also that they are frequently omitted in twelfth-century buildings. But in the latter case this happened either before their advantages were properly appreciated, or else when the mass of the buttress was so great as to render them unnecessary. Thirteenth-century buttresses often have shafts at the angles, Treat- which are detached like the shafts round columns of the same ^,e3Lofangles ¦ period. Sometimes the corners are widely champhered, so much so that the front becomes semi-octagonal. In many churches also — Salisbury Cathedral is one — the stages of the buttresses diminish in width as well as in projection, in which case weatherings occur on all sides. In the middle of the thirteenth century the custom was intro- Comer duced of making one buttress do the work of two at an outside Presses corner (Fig. 29). This was placed diagonally. One buttress was really sufficient, as, although additional strength at a corner is always advisable, the only thrust worth considering likely to VOL. II. E 50 A HISTORY OF ARCHITECTURAL DEVELOPMENT. come at this point (openings are generally too far away from corners for their arched heads to need other abutment than the wall) would be from the diagonal rib of a vault, and that is best counteracted by a buttress on the line of its thrust. Diagonal buttresses, however, never became universal ; and two buttresses, at right angles to the wall and to each other, were used as often as a single one up to the very last. Niches. Niches for figures are sometimes worked on the face of buttresses in Eomanesque churches, especially in Southern France ; but they were comparatively rare until the great revival of figure sculpture about the middle of the twelfth century. Then the sculptor demanded place for his figures; and, not content with the opportunities afforded him in the jambs and heads of doorways, claimed the buttresses also. Beautiful examples of niches, with cusped canopies over them which afford protection to the figures below, occur in the east end of Ely Cathedral, and on the west front of Wells. In Wells Cathedral the niches are on all sides of the buttresses, and, with their figures, produce a richness of effect unsurpassed elsewhere. But- T^g differences between thirteenth and fourteenth-century but tresses, . J 1250- tresses are slight, being merely in detail. About the middle of the 150°- latter century, the custom became general of covering both walls and buttresses with panelling. The result is that a buttress no longer stands out so boldly as before, as its outline is lost to a great extent in the general scheme of rich elaboration. Partly perhaps as a corrective to this, its upper stages are often different in plan from the lower ones. They are octagonal, hexagonal, or else a square placed lozenge-wise ; the different faces being crowned by canopies, above which rises a crocketted pinnacle common to all. In the fifteenth century, especially at the angles of towers, buttresses are often octagonal in plan from top to bottom; and when all their sides are panelled, as well as the wall, they may be said to be almost non-existent, so far as appear ance goes. Batter. The face of a buttress often slopes inwards a trifle (2 inches or so in about 10 feet). This inward inclination, or "batter," helps the set-offs, and assists in producing a pyramidal, and there fore strong, effect. This refinement was undoubtedly practised by the Eomanesque builders in Italy, although it is doubtful if it were employed in more northern countries until the Gothic style was fairly developed. Like the entasis of Greek columns, OUTSIDE OF CHURCHES, WALLS, BUTTRESSES, ETC. 51 it is not perceptible to the eye in most cases, although in others it is more marked. It is not confined to buttresses, but occurs also in towers. In the tower of Stanton Harcourt Church, Oxford shire, it is so unmistakable as to be almost disagreeable. Diminu tion in width is difficult to arrange satisfactorily in octagonal buttresses (as each set-back widens the wall space between them), and a batter is almost an impossibility. The result is that towers so buttressed often present a top-heavy appearance.1 Flying buttresses made their appearance above the aisle roofs Plying which at first concealed them about the middle of the twelfth century, and rapidly became a marked feature of the outside; especially in France, where the great height of the large churches necessitated many buttresses one above another.2 They give an appearance of vigour and active force to a building which nothing else can afford, but when employed extravagantly, as was fre quently done abroad, they create a feeling of unrest which is far from satisfactory. In most Eomanesque churches in Italy there are not even the Pilaster wide, shaUow buttresses common to England and Normandy, but j^' the walls, especially in Lombardic work, are divided into panels by narrow, vertical pilaster-strips, and by horizontal string-courses. These have no structural value unless the walling is of either rubble or brick ; then they act as bonding courses, and strengthen it (see Fig. 113). Pilasters and string-courses are the descendants of the Eoman columns and entablatures, the connecting links being easily traceable in late Eoman buildings, such as the theatre at Orange, France, and in some early churches in Italy. Until the fourteenth century, arcading was the favourite way wall of decorating the walls of mediaeval churches, outside as well as arcading. inside. It was a Classic tradition which filtered through early Christian work from the buildings of Ancient Borne (see S. Apolli- nare-in-Classe, Vol. I., Fig. 117), and it did not disappear until large traceried windows swallowed up all the wall space and left no surface to decorate. In general design and in detail, arcading followed the same evolution as the structural parts of buildings; the arches being semicircular, pointed, foliated, or ogeed, and the shafts being 1 This defect is particularly marked in the Victoria iTower of the Houses of Parliament when the whole tower is visible. 2 For their development and functions, see Chap. VI., on ribbed vaults, pp. 95-100. 52 A HISTORY OF ARCHITECTURAL DEVELOPMENT. attached or detached, moulded or plain, according to the period when the church was built. Throughout the eleventh and twelfth centuries intersecting arches are common (see Fig. 4). In the North these are always semicircular, but in Southern Italy and Sicily, as at Amalfi, Monreale, near Palermo, etc., they are pointed. Between 1280 and 1350 walls are seldom arcaded, for the reason just given ; but after the latter date the desire for over- elaboration which possessed the people, produced the flat panelling, already referred to, which covers walls, buttresses, piers, and even arches. Hori- Plinths, string-courses, oversailing-courses, and parapets form zontal the horizontal lines outside. The main horizontal line of a church, lines. . of course, is its roof (except in fifteenth-century examples, when it is often hidden behind the top parapet) ; and the great, quiet mass of the high-pitched roof in many a Gothic cathedral and church exercises a steadying influence on, and forms an admirable back ground to, the silhouettes of pinnacle and flying buttress below. Horizontal lines could not be dispensed with altogether in mediaeval work, but in the main they are kept subordinate to vertical ones. They are stronger in French Gothic than in English, especially in the western fronts of the great Cathedrals of Laon, Paris, Beims, etc. (see Figs. 196 and 197). Plinths. The mediaeval builders showed their wisdom in starting all but their smallest buildings from a moulded plinth. The Greeks felt the need for something below their columns and walls to give their temples the effect of rising from the ground, and not of sinking into it, and devised the stylobate. The Bomans raised their columns on bases and pedestals, and their buildings on a high podium. It was, however, left to the mediaeval men to develop the moulded plinth, which, in a series of stages, some moulded, some plain, affords a visible footing from which the wall can start. String- String-courses in both Eomanesque and Gothic are always placed immediately below windows on the outside, and often on the inside as well. In these positions their chief office is to separate the pierced portion of the wall from the unpierced ; as their small projection outside affords but little protection to the wall below. Externally they either stop against the buttresses, or else break round them. In Eomanesque churches the abacus which marks the springing line of the window arch (it is often the abacus of the capital of the shaft in the jamb) and from which courses. etc. OUTSIDE OF CHURCHES, WALLS, BUTTRESSES, ETC. 53 the label of the arch starts, is often continued horizontally as a string; and in early Gothic it is the label itself which is returned at the springing for the same purpose. Later, when windows reached from buttress to buttress, no room was left for a string in this position, and that under the windows is the only one.1 In small churches roofs are often carried down to the eaves, Parapets, and their timbers rest on a projecting course which is moulded according to the fashion of the day, and sometimes carved. In large churches they generally stop behind parapets. Considerable thought was given to the designing of these. The parapet projects in nearly all cases in front of the wall below, and in work of the eleventh and twelfth centuries is generally carried on corbels, which, more often than not, are carved as grotesque heads. These corbels frequently support small trefoil or pointed arches (see Fig. 29), which form a rich band under the plain masonry of the parapet above. About the middle of the thirteenth century, the corbel course and arches were discarded in favour of an " over- sailing" course, which, besides being moulded, was frequently carved. The large square flowers in fifteenth-century examples are particularly striking. The parapet at first was quite plain, and so it continued until about the end of the twelfth century. Then panels were either sunk on its face or pierced through it. The designs of these changed from time to time. Stiff and geometric to begin with, the lines later became flowing, as in Malmesbury Abbey Church, and then stiffened again towards the fifteenth century, following much the same development as the traceried heads of windows (see Chap. V., pp. 60-64). In the fifteenth-century, battlemeuting, which had been introduced some time previously, became general, and was either plain, or panelled, or pierced. The reason for the projection of the parapet was to provide more space behind for the gutter and flashings, and also to act, to some extent, as a protection to the wall below. The parapet certainly affords a more dignified finish than the eaves-gutter, and therefore possesses aesthetic advantages as well as mere practical ones. But it has its disadvantages. Defective leadwork in gutterings has caused the timbers to rot in many a roof, and has let in damp to many a wall. 1 The mouldings of strings are much the same as labels over openings (see Pig. 14). 54 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Gur- goyles. The mediaeval VITRE. Lead ^ j qVRC^OYUEA \ builders employed long stone spouts for carry ing off the rain from the gutters, and many of these gurgoyles are carved most grotesquely. They are effective as ornaments ; but are other wise objectionable, inasmuch as they discharge large volumes of water on to the ground at given points. They have in consequence been superseded in modern churches by the often un sightly but more practical rain-water pipe. The gurgoyles are not always in stone ; some of the most effective are in lead. The latter are rare on churches, and are generally found on Fig. 30. mediaeval, or early Eenaissance, domestic buildings CHAPTER V. DEVELOPMENT OF WINDOWS. Windows in Byzantine and early basilican churches are, as a rule, numerous and large ; but in Eomanesque churches, especially in the early ones, they are few in number and generally small.1 The change occurred in Italy in the ninth century, when the Eastern custom, until then prevalent there, of filling windows with pierced slabs was discontinued (see p. 181, and Vol. I. p. 222). These slabs obstructed a considerable amount of light ; when they were done away with one reason for large openings disappeared. Some writers ascribe the change to the temper of the people, who, about that time, discovered that a dim religious light was necessary for their devotions. But apart from this are the practical con siderations that much light was no longer needed — as the old methods of decorating waUs with mosaics had fallen into general disuse, and fresco pamting had hardly taken its place — and that the fewer and smaller the windows the better, since the builders required far stronger abutments for their vaults than had been imperative in the days of timber roofs. Soon it was found that thrusts of vaults could be transmitted by flying buttresses, that walls could be reduced in bulk, and windows, in consequence, became larger than ever. Eomanesque windows are generally placed singly, but in Windows, domestic work, in church belfries, and in the east ends of churches especially, two or more are often grouped together. When in groups the lights are divided from one another by either walling or shafts. In the former case they appear from the outside as though they were separate windows, connected only by a hood moulding, and not always by that; whereas from the inside, 1 Some Romanesque windows in large churches are of considerable size, especially in England and Northern France, where more light was required than in the south. In many English cathedrals such windows have, at a subsequent period, been divided in two by a central mullion, and tracery has been inserted in their heads, as at Peterboro' Cathedral. 1000- 1150. 56 A HISTORY OF ARCHITECTURAL DEVELOPMENT. owing to the wide splaying of the jambs, they unite to form one window. The outer jambs are sometimes plain or merely champhered, but more often they are enriched by shafts, from which spring arches concentric with the window-head. In walls arcaded on the outside, the practice was general to pierce some of the bays of the arcading for light, and this continued the custom so long as walls remained arcaded. The division by shafts is the more usual one in Italy, Germany, Switzerland, and Southern France. The shafts are often placed in the centre of the wall, in which case their capitals have con- Windows, 1150- 1250. ROMA/NE5QVE 2-L1GMT WI/SDOW Vz EXTERNAL ELEVI Vz l/NTERMAL ELEV. SCALE or 12" O SECTIO/S. 10 l FEET. Pig. 31. siderable projection in front and at the back, but very little at the sides, like many Byzantine ones of earlier and contemporary date (see Vol. I., Figs. 145 and 149). In English work before 1066 the central position for the shaft was generally adopted; but after that date (in Northern France as well as in England) the shafts are nearer the outside face of the wall, so as to leave more room for splayed jambs on the inside. In Eomanesque windows the lights are seldom more than thrice their width in height, and sometimes are considerably less.1 ¦ In S. Mary Buildwas Abbey Church, the clerestory windows are about twice their width high, although the east windows are nine times. DEVELOPMENT OF WINDOWS. 57 When the pointed arch supplanted the semicircular, the lights elongated considerably, and to some extent narrowed also. In Kirkstall Abbey Church west front (c. 1160), in Tynemouth and Hexham Priory Churches, and elsewhere, are early instances of Vz rLXTERVNAUT tel/NTERrtAL ELEV. , -rJlik ELEV. 5-L1GMT Wl/NDOW. C-1200. EXT. JAMB MLDC. _„„._,. ra O 1 3 5 C A LE5 of laxl__l_l_L j_aFEET, or i i? i i i m al/SCMES. rOR ELEV§ ETC. FOR DETAILS. Fig. 32. the tall proportions which afterwards became so characteristic a feature of English work. In Hexham Priory Church the lights of the transept are 2-6 feet wide by 24*6 feet high; in Brink- burn Priory Church, Northumberland, the lights of the west window are each 1 foot 10 inches wide by 19 feet high ; in the 58 A HISTORY OF ARCHITECTURAL DEVELOPMENT. west end of Bomsey Abbey Church they are 2 feet wide and as much as 30 feet high from sill to apex. The famous "five sisters " window in the north transept of York Cathedral has the proportion of about 9 to 1. In France much wider lights are the rule, and the builders early curtailed their height in order to develop the tracery of window-heads. The tall English pro portions are consequently seldom met with outside Normandy. This is one instance out of many which will appear in subsequent WI/NDOWv REIMS. JO-f o- 10- ZO-* FEET. WI/SDOW CMARTRES. Fig. 33. pages of differences in development on the two sides of the Channel.1 The lights of thirteenth-century lancet windows are grouped 1 In Soissons Cathedral (c. 1210) the single lancet lights of the aisles are each 7 feet wide, and each bay of the clerestory has a pair of lancets, only a little narrower than the aisle windows, with a foliated circle pierced through the tympanum above. The great width of the lights in early French Gothic and their comparative stumpmess are especially marked in the stately desecrated Church of S. Frambourg, Senlis (c. 1177), one of the finest of the smaller examples of early Gothic in France. The church is aisleless, about 30 feet wide by 130 feet long, exclusive of the semicircular apse, and is vaulted throughout. DEVELOPMENT OF WINDOWS. 59 in much the same way as Eomanesque ones, although a greater number than before are in many instances united under one hood moulding, five and even seven being not unusual. In the jambs outside and on the splays inside are often many detached shafts, and there is an increased richness in mouldings and ornament. Sometimes the lights are so close together that they are separated merely by wide mullions of ashlar, as in Oakham Church, Surrey (c. 1230), and literally form one window. Mullions, however, are more distinctive of windows of later date with traceried heads. As stated above, the Frenchmen were engaged in the develop- Plate ment of tracery whilst we were mainly concerned with perfecting traoery- the proportions of the long lancet light. Most of the early French HEAD OF E>Af\ Tr\ACERY WINDOW. 5Howi/ic joirmnc Or TRACERY Fig. 34. tracery is of the kind known as " plate," which is composed of circles, quatrefoils,' and other geometrical figures, pierced through slabs of stone, which fill the window-heads. In Chartres Cathedral, N. D., Paris, and in other large French churches of the end of the twelfth century, there are many such windows. " Bar " tracery, as it is sometimes called, in which the design is worked in different pieces of stone cut to the required shape and often dowelled together, was a somewhat later creation. The latter does not seem to have been employed in England before 1250 — the date of the Chapter House of Westminster Abbey — or at all events not to any extent ; but in France it was general for some years before. The windows of La Sainte Chapelle, Paris, finished 1248, have bar-tracery heads throughout. 60 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Origin of tracery. Early circular windows. Early Frenchtraceried windows. But it does not follow necessarily that tracery was introduced into England from abroad. The germ of it had existed in this country for nearly a couple of centuries before the French first brought it to a logical conclusion. In Eomanesque work it is no uncommon occurrence to find two lights side by side and over them a circular window ; and when the pointed arch appeared the arrrangement became more and more common, and the lights and the circle were brought closer together. The next step was to enclose them under one hood moulding. When this is done, especially when both lights and circle over are foliated or cusped, the result is a plate-tracery window. An interesting attempt at tracery occurs in the north aisle of Stone Church, Kent (c. 1240). Here there are three windows side by side. Outside they are all alike, each consisting of two lights with a quatrefoil above enclosed by a hood moulding. Inside the design is the same in two, except that in one case the quatrefoil has many mouldings round it, whilst in the other it is merely champhered. On the inside face of the third, or easternmost one, however, is an open screen, quite detached from the window itself, which consists of a cylindrical shaft with richly moulded trefoil heads and quatrefoil over. It is partially plate tracery ; but the mouldings intersect, and if the spandrils had been pierced, the tracery could have been formed entirely by bars of stone. In addition to these tentative attempts at tracery in windows with two or more lights, the circular windows, which occur singly in so many English Eomanesque churches, are often filled with shafts which radiate from the centre and divide each window into many lights, the heads of which are foliated. They are known as wheel windows. In Barfreston and Patrixbourne Churches, both in Kent, are fine early examples, and in the gables of the west front of Peterboro' Cathedral and in the south transept of York Cathedral are others, similar in character, but larger and later in date. One characteristic trait of early French traceried windows, which does not occur in English examples, is that the arches of the window-heads, especially those in the clerestories, are much stilted, or, in other words, the tracery comes down far below their springing line. The whole of the window-head proper is filled by a large circle, sometimes cusped, at other times plain, which is the full width of the lights below (see Fig. 33, Beims). Such windows, generally of two lights, occur most frequently round eastern apses, DEVELOPMENT OF WINDOWS. 61 5TO/SE CHVRCM, KE/NT I/STERAiAL ELLEV.^AISLF: WINDOW. 5ECTIO/S . I i I i I SCALE OF FEET. _u Fig. 35 (A. H. Kersey). 62 A HISTORY OF ARCHITECTURAL DEVELOPMENT. 1250- 1350. because the spaces available for windows there are much narrower than in the bays of the choir west of the apse and in the nave.1 With the exception just mentioned, there is little difference between English and French tracery from 1250 to about 1300. The designs always consist of geometrical figures, circles, quatre- foils, trefoils, spherical triangles, etc., a tendency towards an elongation of the figures being noticeable at the end of the thirteenth century. Early in the following century, however, in England the designs lost all stiffness and the tracery lines became flowing. There can be no doubt that whatever credit is due PATRIXBOVR/NE CHVRCH, KE/NT. 12 b" O I I I I I Fig. 36. for this departure belongs to the English masons. The French examples of curvilinear tracery, as it is sometimes called, are, with very few exceptions, of much later date. Amongst the most beautiful windows in England of this type may be mentioned the east window of Carlisle Cathedral (c. 1330-1380), which still retains a trace of geometric formality, the west window of York Cathedral (c. 1330), and the windows of Boston Church, Lincoln shire, Cottingham, Beverley, Selby, and other large churches in Yorkshire. 1 Stilted window-heads are not confined to apses alone. In many French churches all the window-heads are stilted. DEVELOPMENT OF WINDOWS. 63 DEVELOPMENT OP WI/NDOW TRACERY. HITCHAM, C.I300. BVCKS. /SCTLEY.HAAITJ. C | 11 II I \ COTTIMGHAM, YORK5. C. 1330. C. 1450. CLAYPOLE, C. 1380. LIAHCOL/S. Fig. 37. 64 A HISTORY OF ARCHITECTURAL DEVELOPMENT. 1350-1500. Fifty years later England and the Continent definitely parted company. The English straightened the bars of their tracery, until the entire heads of windows became filled with a series of long, straight-sided, pierced panels ; whereas the French adopted and carried to excess the flowing lines, producing flame-shaped openings, hence the term " flamboyant" generally applied to French traceried windows of the latter part of the fourteenth and the following century. One instance of flamboyant design crossing the Channel is in Brede Church, near Hastings, where not only the tracery but the mouldings as well are curiously French. The transition from flowing lines to straight seems to have come suddenly in England, and it coincides with the devastating Black Death of 1348-1350. Other changes in design, from lancet lights BftEDE CHVRCH. EAST WlfiDOW OF SCVTH AISLE OF CHATIG JAM a 6 ARCH Fig. 38. Transoms. to traceried windows, from geometrical forms to curvilinear, came gradually. There are many examples which are neither one thing nor exactly another ; but there are few windows in England combining both flowing and vertical lines. The west windows of Monmouth Church and the little chapel of Houghton-le-dale, Norfolk (c. 1350) (see Fig. 27), and the east window of Claypole Church, Lincolnshire, are instances. One reason for the change was that perpendicular tracery requires less stone, less work, and less skill in setting out than either of the earlier forms. Workmen and clients were both in a hurry. They wanted to make up for the loss of time which the Black Death and the general distress which followed had caused, and, besides building quickly, wished to build economically. In nearly all large fifteenth-century windows the lights are DEVELOPMENT OF WINDOWS. 65 divided horizontally by transoms, of which there are sometimes two or more to each light, as in the west window of Winchester Cathedral, the windows of Bath Abbey Church, etc. Transoms first began to be inserted to any extent about the middle of the previous century, but they are not unknown in earlier work, especiaUy in domestic buildings. The windows of the Banqueting Hall in the Bishop's Palace, Wells (c. 1230), for instance, are divided by them. They were of some advantage to the stained glass designer, inasmuch as each of his figures could be framed in top and bottom, as well as at the sides, and this is sometimes stated as the reason for their introduction. But it is doubtful if his wishes would have had much effect on the mason if the latter had not desired cross-stays to his tall, thin mullions.1 It may be urged that the increased popularity of transoms was due to a desire to bring the window into unison with the panelled wall ; but the converse is more probable. The panelled wall was more likely the result of the panelled window. Much of the beauty of traceried windows is due to the Subordi- emphasis given to certain parts of each. Some mullions and of'^™aow tracery bars are wider than others, and have a greater number mould- of mouldings worked on them. The lights and the heads are mgs- consequently grouped. There is no hard-and-fast rule as to which mullions shall be emphasized, but in four- and six-light windows the centre one is generally thicker than the others ; or in the latter there may be two thick mullions dividing the window into three divisions of two lights each. In a five-light window the two central mullions generaUy have extra mouldings, and in a seven-light either the central ones or the second from either side.2 In windows of many lights, mullions of three different sizes are not uncommon, the largest being of three " orders," the next size of two, and the smallest only of one. The same scheme runs through window-heads. The main and secondary divisions are emphasized by similar means, and great delicacy is often given to a design by the slightness of the bars which surround 1 Each light is by no means always filled by one figure. In the great east window of York Cathedral (filled with glass in 1405-1408) there are several figures in each ; and in many fifteenth-century windows without transoms, the lights of which are much longer in proportion than in any window with them, the glass stainer has shown how he could employ one figure only, and yet fill the whole space by the aid of elaborate canopy work. 2 This is not a universal rule. Many windows of several lights have all their mullions the same thickness. VOL. II. F 66 A HISTORY OF ARCHITECTURAL DEVELOPMENT. the smaller openings. In most large windows, and in some small ones, the central light (or lights) is wider that the side ones, partly because of the stronger mullions which frame it, and partly because, even when all mullions are of the same thickness, a central light has a tendency to look narrower than the side lights if all are of the same width.1 Mullions. There is an enormous difference in proportion and detail between mullions and bars in early and late traceried windows. Early mullions are very wide, but have small projection from the face of the glass ; the later ones have greater depth, but are much thinner. There is not really very much difference in superficial area between the two — the early ones are a trifle heavier — and the appearance of strength is much the same, especially when DEVELOPMENT OF AWLLIO/S5 C.I330 - l+SO Fig. 39. Cuspings. the window is viewed sideways. As regards the contours of the mouldings at different periods, when the design is geometrical the principal mouldings of mullions, jambs, and tracery are segments of circles, generally three-quarter rounds, although sometimes the curve is broken by a fillet. When the lines of the tracery are flowing, the mouldings are mainly ogees, and the ogee continued general after the lines straightened. Thus, much in the same way as arch mouldings altered as the arch form changed, so the mouldings of tracery changed to agree with its lines. In most plate-tracery windows, and in a few early ones of the other class, the cusps spring from the flat soffit alone, not from the splayed or moulded part of the tracery at all. Such cusps are 1 For the same reason the central openings in a Greek peristyle are often wider than the side ones (see Vol. I. p. 91). DEVELOPMENT OF WINDOWS. 67 either worked on the solid, or else are in separate pieces of stone which are tenoned into the head of a light or into a circle, or what ever the opening may be. Many of these have dropped out, and have not been reinstated in "restoration." This "soffit," cusping, as it is called, disappeared as bar tracery developed, but it occurs in the heads of the triforium openings of the Angel Choir, Lincoln (c. 1260), and Westminster Abbey ; and as late as the fourteenth century in the highly enriched windows of Leominster Church (c. 1310). In most windows from 1250 onwards the cusping either starts from a fillet or else cuts into the moulding below. CVSPS AND- CVSP ENDINGS C.I30O FILLET CVSPIMG CI200 CHAMPHER C.I400 CVSPIMG — FlG. 40. The back of the cusp is champhered so as to form an " eye," which is sometimes merely sunk, sometimes pierced and glazed. The cusp terminations in the different centuries vary consider ably. In the thirteenth century they finish square or with a blob ; in the fourteenth they come to a sharp point, and in the fifteenth to a broad point. An examination of the windows of the different periods shows how well suited each form is to the shape of the cusped opening in which it occurs. In the windows of the eleventh, twelfth, and early thirteenth Window centuries the arch on the inside face of the wall over each light is aL0^ ?-on~ generaUy concentric with the window-head, but rises high above it because of the wide splaying of the jambs (see Figs. 31-2). When many lights are grouped together to form one window, Circularwindows. 68 A HISTORY OF ARCHITECTURAL DEVELOPMENT. being divided from one another merely by mullions, another treatment becomes necessary. The inside arch is made much flatter than the outside one, and, notwithstanding that it is far wider, it is in consequence very little, if any, higher ; sometimes it is actually lower. Late thirteenth, fourteenth, and fifteenth- century window openings have, in most cases, three arches over them, which are struck from different centres. On the outside is the window arch, which includes the jamb mouldings, if there are any, and the tracery, its duty being to carry the outer face of the wall ; on the inside is another arch, called the " scoinson," which carries the inner face ; and between the two, an intermediate, or rear, arch which comes under the core. This construction is Fig. 41. i another instance of the desire of mediaeval builders, mentioned before, to give elasticity to their work, and so, by not connecting all parts together too rigidly, prevent the possibility of cracks if settlements occurred. The lowering of the inner arch was ric*ht for another reason. Tops of windows are always above the eye ; in most cases considerably above it ; and the eye consequently sees under an inner arch. The top of the tracery may be above the apex of this arch and yet the whole of it be visible from below. Besides, two arches of different curvature form pleasant contrasts. The early wheel windows, mentioned on a previous page, are the forerunners of the large " rose " windows which form such effective features in the great cathedrals abroad. Mediaeval DEVELOPMENT OF WINDOWS. 69 builders in England never took kindly to them. There is a traceried circular window of fair size in the south transept of Lincoln Cathedral, and a somewhat larger one in the Chapel of the Nine Altars at Durham. In the transepts of Westminster Abbey are similar windows, the design of which has been altered more than once. Apart from these there are few of any size in England. On the continent it is different. Hardly an important cathedral is without its circular window, either in the west front or in one or both of the transepts. The design of these varies according to date, and the whole gamut of French tracery can be studied from circular windows alone; from the plate-tracery examples in the west fronts of Laon and Chartres, through the geometric ones of Paris and Beims, down to the ones filled with flowing lines, of which the finest is probably that in the south transept of S. Ouen, Eouen. Even when a French western or transeptal window has vertical lights, an enormous traceried circle often fills not only the head but halfway down the sides as well, as in the south transept of Beauvais Cathedral. This design is the natural development from the earlier windows in which the arch is stilted to allow of a circle the fall width of the lights below it, as shown in Fig. 33. CHAPTER VI. VAULTING. Intro- The most characteristic trait of mediaeval architecture is vaulting, and yet it is by no means universal in either Eomanesque or Gothic churches. In Italy, the traditions of the wood-ceiled basilican church account for its absence during the eleventh and twelfth centuries, and even in later ones. Elsewhere in Europe, in England, Normandy, and Germany, it was not so much tradition as the difficulty of vaulting wide spans which caused many of the large churches to have timber roofs over their naves. The aisles could easily be vaulted, and generally were; but the workmen were not always sufficiently skilled at first to attack the more difficult problem of vaulting the naves, choirs, and transepts. The vault was no invention of the Middle Ages. Of the two forms of vault, the barrel, or cradle, vault and the intersect ing vault, the former was well known to the Egyptians and Assyrians, and both were extensively used by the Eomans. But although the forms remained much the same, the methods of construction finally adopted by mediaeval builders — especiaUy as regards the intersecting vault — are in many respects different from those of earlier work. The steps by which they reached the solution of the difficulties which beset them, and the further modifications wliich they introduced, form the subject-matter of most of this chapter. Barrel Some mediaeval barrel vaults, like Eoman ones, are solid ; and vaults. . the outer covering, generally tiles, is bedded on top of the vault itself. In the majority of cases, however, they are only a foot or so thick, and are covered by a protecting timber roof, in the same way as intersecting vaults are covered and protected. One practical disadvantage of a barrel vault over an aisled church, is that, being continuous, it presses equally on voids and solids, on arches as well as on piers. In the words of Sir G. G. Scott, it VAULTING. 71 entails "an illogical arrangement of divided substructure and continuous superstructure." This, so far as it affects appearances merely, is partly overcome in many twelfth-century churches, especiaUy in Burgundy and Auvergne, by placing transverse arches above the piers, as in Autun Cathedral (see Fig. 161). These, however, are of httle use structurally, although they do strengthen the vault sUghtly at these points. The vault is not carried on them, as in the Eoman building known as the Baths of Diana, Nimes (see Vol. I., Fig. 75, p. 122); they are too far apart.1 Another objection sometimes urged against a barrel vault is that, unless a church is of great height, clerestories are impossible, as aU upper windows at the sides have to be kept below its springing. As a matter of fact, this is really a blessing in disguise, because the principal Ught to a church should come from the west end, and obtrusive side light is a nuisance. Barrel vaults were never popular in England in either Eomanesque or later times. Early examples are over the nave and gaUeries of S. John's Chapel in the Tower of London (c. 1080), the vaults being semicircular ones. Later vaults occur over Eoslyn Chapel, Scotland, and the little S. Catherine's Chapel, Abbotsbury, Dorset. In the latter building the vault is a pointed one, and the whole is panneUed in stone, with exceUent results. The intersecting vault permits of windows at the sides, as Advan- high as its apex, but that is not the reason why it was preferred, infersect- Its great advantage is that it concentrates the thrusts and weight ing vault. of a vault over the points best capable of receiving them, namely, the piers ; a fact that was known to and thoroughly appreciated by the Bomans. The earUest Eomanesque intersecting vaults, Uke Eoman ones, Develop- are continuous, as in the nave aisles of the Abbaye-aux-Dames, ^Jj'.of Caen, and the crypt of Eochester Cathedral. The first modifica- transverse tion made by the eleventh-century builders was to divide the arohea- vault into bays by transverse arches and longitudinal arches (or wall ribs, when the vault is enclosed on two of its sides by waUs). These arches spring from pilasters or attached columns projecting in front of waU or pier, and as their span is less than that of the vault enclosed by them they show below 1 With steel and reinforced concrete there is no reason why the barrel vault, with transverse ribs to mark the bays, should not come again into favour, and be built on sounder principles. Its form is to some far more beautiful than the intersecting vault. 72 A HISTORY OF ARCHITECTURAL DEVELOPMENT. it. The infilling of each bay (or severy, to use the more archi tectural term), now rests on these arches, thus removing still more the weight of the vault off the voids below, and concentrat ing it above the piers.1 CO/STIN VOVS VAVLT5. A.A.-TRA^SVERSE ARCHES. B.B.-LO^GITVDIAJAL « C.C.-WALL RIBS. VAVLT \n BAYS. Fig. 42. Diagonal ribs. The next development carried the system one step further. This was the introduction of diagonal ribs below the line3 of intersection, or " groin " lines of the vault. The vault now becomes a ribbed vault, in contradistinction to a groined or inter secting vault, and each severy consists of two distinct parts, (1) the constructional frame, composed of transverse arches, longi tudinal arches or ribs,2 and diagonal ribs ; and (2) the infilling, or web, which rests on the frame. The latter, as it has only its own weight to carry, can be made merely a few inches thick, and practi cally no part of it bears on the walls. The principle of a combina tion of thrusts concentrated at given points is now complete, and this concentration facilitated their neutralization, as will be seen later when dealing with flying buttresses. The structural value of ribs has been so emphasized that their aesthetic value is sometimes forgotten. It is on them the eye rests ; and not on the infilUng, 1 The Byzantines used transverse arches occasionally, as, for instance, in the vaults over the narthex, outer narthex, and aisles of S. Sophia, Constantinople; but there is this difference, the arches do not spring from projections, and would not show at all if it were not that the vaults are domical. 2 In many early vaults in England and Prance wall ribs are omitted, as the great thiokness of the walls below rendered them unnecessary. VAULTING. 73 the curves of which are, to a great extent, lost in most instances, owing to distance and insufficient lighting. The Englishman possibly felt their aesthetic value more than the Frenchman, whose attention was mainly concentrated on structural considerations, and this, to some extent, accounts for the differences between EngUsh and French vaulting mentioned in the following pages. Diagonal ribs first appeared in churches towards the end of the Early eleventh century, but they did not become general for some fifty ^J^ years. Most Eomanesque vaults of the first quarter of the twelfth century are without them, and in Burgundy, owing to the con servative instincts of the Monastic orders which were especially FUSBED VAVLTS. A-TRA/H5VER5E ARCH. B-WALL RIB. C- DIAGONAL RIB. < ARCME5 C WITH C.SEAM -ELLIP TICAL, B. STILTED & A.SEAU-CIRCVLAR.. WITH CSE/^M-CIRCV- LAR A/ID B.5T1LTED. A-^ "^ ^?"€ rC^^P #T '/A ^^ WITH CSEMI-CIRCV- LAR A/SD A.& B. STIL TED, A5 AT VEZELAY. PiG. 43. powerful there, they are often omitted in later work. There can be no doubt that they simplified considerably the building of vaults of wide span, and the Eomanesque builders therefore deserve great praise for realizing their advantages. But did they, strictly speaking, originate diagonal ribs ? Did they not really merely reintroduce them in a modified form? In Eoman intersecting vaults there are diagonal ribs of brickwork (see Vol. I., Fig. 74, p. 120). These, however, do not project below the surface of the vault. Moreover, as successive layers of concrete were added, they lost their independent functions, and became imbedded in, and part and parcel of, the vault. Still they are there, and the Eomans used them for much the same reason as they were used 74 A HISTORY OF ARCHITECTURAL DEVELOPMENT. later; namely, as permanent centers to carry the infilUng.1 It is quite possible that by the eleventh century the old methods had been entirely forgotten, but, even admitting that as proved, the builders of the Middle Ages cannot be credited with an entirely new invention. The date of a vault is always difficult to fix, partly because it is the last part of a church to be buUt. When a church was commenced may be known exactly, but there is very seldom documentary evidence to show that work proceeded without a break; and where documentary or other evidence does exist, it generally proves the contrary. Besides, builders were inexperi enced at the beginning, and many early vaults undoubtedly collapsed and had to be rebuilt. Fires also necessitated rebuUd- ing to an almost incredible extent, and in a fire a vault suffers more than almost any other part of a church. Wliich country, therefore, deserves the credit of being the first to use visible ribs below the groins cannot be decided with certainty. StiU less can an authoritative statement be made regarding any particular church. All countries were engaged trying to find a satisfactory solution of the same difficulty. In France, ribbed vaults appeared simultaneously in different parts at the beginning of the twelfth century : at Bordeaux and Poitiers in the south, at Quimperle (Brittany) in the north, round Beauvais, and in the valleys of the Seine, Marne, Oise, etc.2 At Morienval (Oise), a narrow am bulatory round the eastern apse is covered by ribbed vaults, and the date of these is said to be c. 1120.3 In the district round Beauvais and Soissons many churches of about this date have aisles vaulted with diagonal ribs. The nave of S. Etienne, Beauvais, is also vaulted in this manner, but the vaults are 1 In Roman intersecting vaults, such as those of the Basilica of Constantino, etc., the weight of the vault was concentrated over the piers quite as much as in mediaeval vaults. 2 M. Oorroyer's theory (see his " Gothic Architecture," translated) is that diagonal ribs were first used in Aquitaine under vaults built in rings like a dome, but this receives little support from his own countrymen, and from others. 3 The east end of the church at Morienval has been altered so often that it is difficult to give a date to any part. The chanoel was evidently originally barrel vaulted — springers of the vault remain on both sides — and ended with a semi-dome over the apse. The present semi-dome starts higher than the original one, is pointed, and has two ribs under it, although built dome fashion with concentric courses. The ambulatory beyond is an addition, and although its correct date may be 1120, the work is so rough that it may be earlier. The transverse arches (semi circular) across the ambulatory are only about 2 feet 6 inches wide. VAULTING. 75 certainly later than the original date of the church (1120-1123), and were probably rebuilt after a fire in 1180.1 Some of the aisle vaults, which are ribbed, date from before the fire. The claim of priority for Italy rests chiefly on S. Ambrogio, s. Ambro- Milan, the wide nave of which has ribbed vaults. The east end of sio>Mllan' this church is ninth century. The nave, according to Cattaneo, was built " in the second half of the eleventh century." H this date is correct, one of the widest of Eomanesque vaults — its width is 44 feet — and also one of the most effective (see Fig. 112), was built twenty years or more before the similar smaUer vaults of France. It is difficult to beUeve this, although it may be true. Cattaneo continues, "In 1129 the second belfry was erected, and in 1196 they repaired the damage done to the edifice by the faU of an arch in the principal nave." This sentence mentions two dates when building operations were in progress. The first rebuilding of the nave may have merely preceded the buUding of the campanile tower by a few years, which would give c. 1120 for the nave vault. But "the faU of an arch," if that were a transverse arch, means the coUapse of two bays ; and there are only three vaulted bays to the nave of S. Ambrogio. Cattaneo's own statement therefore shows that part of the nave vault must have been rebuilt after 1196 ; perhaps the whole of it was. On the other hand, the angle shafts which carry the diagonals are so important that they must be part and parcel of the original design in the first rebuilding (see Figs. 15 and 112). They would never have been made so large if they had been intended merely to carry groin lines. One may, therefore, fairly assume that diagonal ribs were intended from the first, and the real question is, was the nave buUt forty or fifty years before the tower — the date of which is known to be 1129 — or at about the same time ? This vault was not so difficult to build as many later ones in England and France, because, although its span is considerable, it springs low down, only a few feet above the crown of the arches of the main arcades. Angle shafts, on a smaUer scale than those in S. Ambrogio, are Angle not conclusive evidence that the vaults over had diagonal ribs. jfba^|and There are angle shafts to the vault under the gallery of the north vaulting. 1 Mr. Moore, in his " Gothic Architecture," states that the present vaults were copied from the original ones, which also had ribs. I was at Beauvais this year (July, 1907), but was unable to discover any signs which would prove this, as the church is under restoration, now nearly completed, and the greater part of the stonework inside, especially the capitals of the piers which might have supplied a clue, appears to be quite new. 76 A HISTORY OF ARCHITECTURAL DEVELOPMENT. transept, Winchester, but no diagonal ribs (see Fig. 44). Angle shafts were introduced before diagonal ribs were thought of, because the groin lines of the vault otherwise had to start from the WI/1CMESTER V/HDER GALLERY OF flORTH TRANSEPT Pig. 44 (J. Bilson). inner angle of the pier carrying the transverse arch, which did not provide satisfactory springing. Durham For England a strong claim can be made for priority ; but even Cathedral. tjia(. cann0^ \,Q entirely substantiated. In Durham Cathedral, the small vaults over the aisles of choir and transept are ribbed vaults, and these Mr. Bilson places as early as 1093-1096.1 The period given by him for the nave vaults is between 1128 and 1133 (see Fig. 176). These vaults will be referred to later (see p. 251). It is natural that the question of priority regarding diagonal ribs should arouse considerable controversy, because they went a long way towards solving the problem of vaulting large spaces, and thus greatly assisted the development of mediaeval art. Early Diagonal ribs were a great advance, but they caused a difficulty cutties which had not existed before. In a square vault without ribs, the arches of which are semicircular, the groins are semi-elliptical and therefore weak in form. This weakness does not matter in vaults of this description, as the groin lines are unconstructional and, so to speak, accidental. But when diagonal ribs are added, 1 See " The Beginnings of Gothic Architecture," by Mr. John Bilson, in The B.I.B.A, Journal, vol. vi., Nos. 9 and 10. Part II. should be read for the arguments regarding these vaults and other early ribbed ones at Winchester and Peterboro', and Part I. for an interesting summary of the pros and cons for continental priority. VAULTING. 77 these cannot well be built on the same lines, as they have to carry a great part of the weight of the infilUng and must therefore be strong. The Eomanesque builders tried making the ribs segmental, as in Durham Cathedral, or else semicircular (see Fig. 43). The latter form was strong, but when the other arches of the vault were semicircular also the result was practically a dome, and should have been built as such.1 The domical form was avoided in some cases by stilting considerably the transverse and longitudinal arches so that their crowns should be on about the same level as the apex of the diagonal ribs. A better expedient still was followed at Ouistreham Church, near Caen, where the arches are semi- elliptical; and this was also sometimes done in early English examples. Most early ribbed vaults are over aisles ; their spans are consequently smaU, and the builders could " cook " the curves without producing weakness or ugly lines. Exact setting out of these vaults was also rendered difficult by the fact that, in England certainly, although many side vaults are approximately square, few are exactly so, and some are unmistakably oblong. These rough and ready devices were all very well over aisles, Diffi- but they were unsuitable for wide spans. Nave vaults required g^ygj3!, more scientific treatment. Searching about for a remedy, the pointed workmen bethought themselves of the pointed arch, which had, ar0 es' some years previously, lessened the difficulty of the barrel vault. The solution of the difficulty was found. There was no longer any occasion for stilting arches or for weak diagonal ribs ; there was no longer any need for a domical form, although for many years, it must be admitted, especiaUy in Southern France, the latter pre- vaUed in vaults built entirely with pointed arches, and arches, especially wall ribs, continued to be stilted when there was often little reason why they should be. Fig. 45 shows how transverse and longitudinal arches and diagonal ribs can, by the aid of the pointed arch, aU be of the same height, no matter what the shape of the vault may be, or how wide its span. In most French vaults, in which the system is complete, the diagonal ribs are made semicircular, and the radius of these dictates the height and consequently the forms of the other arches. In England they are generally slightly pointed. The Normans were the pioneers in vaulting in Northern Early Europe, but they clung to the semicircular arch, and their fame ribb®d 1 M. Corroyer says that in the church of S. Avit Senieur, in Southern France, . it is a dome, the stones being laid in rings, and the ribs acting merely as stiffeners. ?rcneq 78 A HISTORY OF ARCHITECTURAL DEVELOPMENT. rests more on their eleventh-century work than on their twelfth. In Germany and Italy also, traditions were too strong to aUow of early change, and the "round arched Gothic," as it is sometimes caUed, of the former country is amongst the finest in Europe.1 France and England were the first to adopt generally the pointed form. The vaults over the aisles of Malmesbury Abbey Church (c. 1140-1150) are amongst the earliest ribbed vaults in this country in which the transverse arches, and the arches of the nave arcade alongside them, are pointed. The diagonal ribs are semicircular. In Durham nave the transverse arches are also pointed and, if Mr. Bilson's dates are correct (1128-1133), these arches are earUer than the Malmesbury ones. These vaults have no wall ribs. In OBLO/NG VAVLT WITH POI/STED ARCHES OF EQVAL HEIGHT. DITTO. VAVLT WITH DITTO. ARCHES, BVT WALL, RIBS STILTED. A=TRA/NSVERSE ARCH B = WALL RIBS C = DIAGO/SAL RIB5. D RIDGE RIBS. Pig. 45. France the pointed arch was employed in vaults almost as early as diagonal ribs. Many small churches round Amiens, Beauvais, Soissons, etc., have pointed ribbed vaults, which are attributed to 1120-1130. The first large church with vaults of this description, the date of which is known, is S. Denis, near Paris. The narthex was finished 1140, the choir in 1144. In S. Maurice, Angers (the cathedral), the magnificent vault of about fifty-four feet span is pointed throughout, and cannot be later than 1150-1153. The dates of the above examples show that the pointed arch was used 1 It must not be forgotten that at first the pointed arch was only used for arches of arcades and for vaults. Windows continued to be semicircular headed until near the end of the twelfth century. ISM I Mm Hi Photo: Author. Fig. 46. — S. Georges de Boschehville : One Bay of Choir Vault. There are no Diagonal Bibs ; Lines are only Painted. Photo: Author. Fig. 47. — Angers Cathedral : Nave. [To face 21. 76 VAULTING. 79 for vaulting almost simultaneously in France and England ; although the Frenchmen had the advantage of a few years. One important difference between early and late ribbed vaults Trans- is the great width of the tranverse arches in the former. This is arches especiaUy marked in churches of the South of France, such as very wide. S. Badegonde, and the Cathedral, Poitiers, S. Pierre, Saumur, S. Trinite, Angers ; in S. Ambrogio, Milan, and in most German churches of the tweUth century. It is also noticeable, though to a less extent, in the somewhat later great churches of Northern France, such as N. D., Paris, Laon Cathedral, etc. The transverse arches in Eomanesque vaults in this country are also heavy, but then so are aU the other ribs. In the churches of Poitiers and Saumur, etc., the diagonal ribs are exceptionally sUght. The wide transverse arch possesses one great advantage : it unmistakably divides the vault into bays corresponding to the bays below, and prevents the somewhat monotonous effect which a vault with all its ribs about equal in size presents. In the churches of Southern France it was a necessity because of the domical character of the vaults. It was needed for strength, and it was required for appearance (see Fig. 164). In England even the earliest vaults are only sUghtly domical in form, the later ones not at all, and, conse quently, wide transverse arches, except in quite early work, are practicaUy unknown. The pointed arch not only solved the difficulty occasioned by the Vaulting use of diagonal ribs, but it also faciUtated the vaulting of oblong °paoe°ng spaces. Such spaces, it is true, had been vaulted before diagonal ribs and pointed arches were thought of. Some of the vaults in Cara- caUa's Baths, Bome, for instance, are not square. The best instances in Eomanesque times of oblong vaults without ribs and with semicircular arches occur over the nave of Vezelay Abbey Church and the choir of S. Georges de Boscherville, Normandy (c. 1120) (see Fig. 46). Here the wall ribs are much stilted and the transverse arches slightly stilted, so that their crowns are nearly straight. The result is not altogether satisfactory, as the weight of the vault comes too much on the transverse arches and too little on the walls. It may be asked, why should not all vaults be square ? In the majority of churches the nave is about double the width of each aisle, and, consequently, if a bay of one be square the corresponding bay of the other must be oblong. Before the pointed arch became general the oblong form was avoided, as in S. Ambrogio, Milan, and in many early French and 80 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Sex- partitevaults. German churches vaulted with semicircular transverse arches, by making one bay of the nave vault correspond to two bays of the aisle on either side ; all bays being thus approximately square. But the objection to this plan is that each bay of the central vault is enormous, and the span of its diagonal ribs, in particular, immense (see Fig. 51). As a corrective to this the French designed what is known as the sexpartite vault. All early ribbed vaults are quadripartite, that is to say, each bay is divided into four compartments by the diagonal ribs. In the sexpartite vault an intermediate trans verse arch is introduced, which cuts the diagonal ribs at their SEXPARTITE VAVLTIMG. / ABBAYE AVX DAA\E5,CAE/1 F.fVS. 1906. Fig. 48. intersection and thus supports them. This, in the early essays, was simply an arch built under the web, which was formed independently of it. Examples are in the nave of the Abbaye- aux-Dames, Caen, and the church of Bernieres-sur-mer, both in Normandy. The next step was to let these intermediate arches take their share in supporting the web. The result was the sexpartite vault, i.e. a vault divided into six compartments. It was a great favourite in France for some thirty or forty years, but it was definitely abandoned early in the thirteenth century. iEsthetically it is unsatisfactory, as it leads to many twisted curves in the web, which, at the back of the intermediate arches, is often no wider than the arches themselves for a considerable height above the springing. In some French examples the bad Photo: Author. Fig. 49. — S. Pierre, Lisieux: Quadripartite Vault of Choir. 'Photo: Author. Fig. 50. — Laon Cathedral : Sexpartite Vault of Nave. [To face p. 80. VAULTING. 8 1 effect is minimized by starting the intermediate arches some feet above the other arches of the vault. Nearly all these sexpartite vaults are much scooped ; the intermediate arches have therefore to rise considerably higher than the main transverse arches, which is an additional reason for stilting them. In the church of La Trinite, Angers, the intermediate arches are stilted as much as eight feet. The following French cathedrals, amongst others, ¦ have sexpartite vaults : Laon, Paris, Bourges, etc. Noyon was '£: ',?;" r ... I ?, '(-- -.-- •:<&- 5QVARE VAVLTI/NCOVAD- Rl PARTITE. C.1050-1150 SOVARE VAVLTIAIG, SEXPARTITE. C.llOO. // i *\ ii u // •i i ,\ "i I!!! V 'a\ il'i .. i* -<- v c "\~ . "'1 - ' ' ¦5.- ^ %Vi>'' n-,1 ^ J- , £ 1 ¦#¦:¦- 1 -- - I'll _ _IMi J\.-J llll ^ ' ,5"c „ 'A- ,;W^.' 'V;,.' !!'! *** ,**' * I ¦ ¦ B J =3 * SOVARE & OBLONG VAVLTI^IG C. 1300. t" H=--]i Ml V OBLO/SG VAVLTl/^G. C. 1200 - 1300 Fig. 51. designed for sexpartite vaults, but the present ones are quadri partite. In England there are few examples of the sexpartite form. The choir of Canterbury Cathedral (the work of WiUiam of Sens, a Burgundian) and the south transept of Lincoln Cathedral, however, are vaulted in this manner.1 1 In early sexpartite vaulted churches the columns are alternately large and small — which is logical — but in some of the later examples — Paris, Laon, etc.— the columns are, with some exceptions, of the same size. At Noyon columns alternate with larger piers. VOL. II. G 82 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Oblong The difficulty of vaulting wide square bays with quadripartite vaults, and the unsuitabiUty of the sexpartite form forced the builders in most countries to make their nave bays oblong, the length being from north to south, and to make each bay of the nave the width of an aisle bay each side. The aisle bays took their chance. They are sometimes approximately square, at other times oblong, their length being from north to south or from east to west as was necessary. Owing to the pointed arch, the shape of the plan of a vault was immaterial, provided the span of the diagonal ribs was not too great. The ItaUans alone clung to the square for their central bays until the advent of the Eenaissance. They dispensed with the intermediate pier of the Eomanesque builders, and in the aisles, instead of two bays to each bay of the nave, had one long oblong bay. The result is that in DVRHAM ,/^IAVE CHOIR, LI/SCOL/^I Fig. 52. Italian churches of the fourteenth and fifteenth centuries the supports are always far less numerous than in more northern examples.1 The French having returned to the quadripartite vault and perfected it — the credit for this is undoubtedly theirs — stopped satisfied. Even in their fifteenth- century churches the vaults are of this simple character, with very few exceptions, the only ribs being the transverse, diagonal, and waU ribs, with occasionally a ridge rib. But the English were restless. They had shown in the early vault over Durham nave — which differs from other vaults in being in seven compartments, there being no transverse arch above the intermediate piers, where one might naturally have been expected — that they liked something different from other 1 The builders were doubtless influenced by the old Eoman plan of few supports, but they forgot that the Roman supports are exceptionally large. Hence is owing, to a great extent, the unsatisfactory character of Italian Gothic. VAULTING. 83 people, and this feeling is shown again in the choir vault of Lincoln Cathedral, where the diagonal ribs do not run from corner to corner (see Fig. 52), and in subsequent developments. Before considering these, however, it is best to deal briefly The web. with the methods of building the infilling or web. In early vaults this was often Uttle more than rubble, plastered on the underside, but in later work it was coursed stone. Whichever method is adopted, its surface is sUghtly concave, so that it forms a A\ETHODS OF LAYING WEB. COVRSES /NOT PARALLEL WITH RIDGE COVRSES PARALLEL WITH RIDGE. Pig. 53. flat segmental arch from rib to rib. Viollet-le-Duc, in his " Diction- naire Baisonne," describes two methods for coursed stone, one which he dubs French, and the other English. In France, he says, the stones were always built parallel to the ridge, and in England and in the semi -English work of Southern France, they were placed either at right angles to the diagonal ribs, or else at right angles to an imaginary line drawn between the diagonals and the trans verse arch, or the diagonals and the wall rib. But this is not quite correct. The so-called French method is, it is true, general in Northern France ; but it is also not uncommon in England. The two methods occur, side by side, in the vault over the cloisters 84 A HISTORY OF ARCHITECTURAL DEVELOPMENT. on the north side of Westminster Abbey. In Westminster Abbey there may be strong suspicion of French influence, but there can be none in the Lady Chapel of Chichester, or at Fountains Abbey ; and yet in the vault of the former, and in that over the huge cellar of the latter, the stones are laid, to all intents and purposes, parallel with the ridge.1 The truth is that in nearly all examples, French as well as English, the web is by no means bmlt through out in exact accordance with any hard-and-fast rule. The stones at the bottom often start anyhow; and up to the point where they could be safely built merely by hand and without centering of any kind, the courses vary considerably in height and often in direction. Above that point more care was required, and more regular coursing commences. The web is generaUy of the lightest material available, clunch {i.e. hard chalk) being not uncommon, although courses of harder, heavier stone are sometimes introduced at intervals, as in Westminster Abbey. It varies in thickness from 4 to 8 inches, although in some early vaults a kind of concrete is laid on its extrados, as in Chichester Cathedral and Westminster Abbey nave vaults. In later work this is omitted, probably because it was found that the added strength was dearly bought by the additional weight.2 Sections In Eomanesque vaults the web is generally laid on the extrados of the ribs, wliich are wide, shallow, and heavy, in agree ment with the other parts of a church. Early in the thirteenth century, however, the ribs were made much deeper, and were rebated to receive the web. They are also far thinner, but their strength is as great as, if not greater than, the earlier ones, as their depth is increased by an amount equal to the thickness of the web. The joints were often strengthened by metal or slate dowels, especially in late work. Many of the later ribs are not 1 In one of the bays of the choir aisle of Tewkesbury Abbey, one side is built French fashion, and the other in so-called English fashion, but one or other may have been restored. 2 In Chichester Cathedral the concrete is a foot thick, the chalk below 6 inches. In Westminster Abbey the vaults over the eastern part of the church and over about two-thirds of the nave have about 9 inches of concrete above them. The remaining bays to the west, which are a oentury or two later in date, have none. The Boman tradition of layer upon layer of concrete may possibly have suggested this covering ; but it is more probable that the old traditions were lost, and that the early medieval builders added the layers because they were afraid of the strength of their thin shells. of ribs. VAULTING. 85 only elaborately moulded, but are also enriched with carved ornament. The task of building the web was by no means an easy one The build- in some of the wide, lofty churches of Central and Northern ^egb°£the France. The EngUshmen never mastered the science of vaulting so thoroughly as the French, although in beauty English examples more than hold their own, and present greater variety than con tinental ones. The mediaeval builders were confronted with much the same difficulty as the Eomans centuries before. Although wood was plentiful in Europe in the Middle Ages, the cost of scaffolding and ordinary centering for a vault, such as that over Amiens Cathedral, would have been a serious item if special devices had not been adopted. According to Viollet-le-Duc, in the Isle-de-France at least, the workmen employed hand-centers, consisting of two planks of wood, curved on top, to give the desired concave surface. A slot was cut along each, and the HA/ND CE/ITER FOR WEB. Fig. 54 (Viollet-le-Due). two were loosely joined together by wedges. At the ends were angle-irons, which rested on the ribs. Where the space to be fiUed was narrow, i.e. at a few feet only above the springing, the planks lay side by side ; as the space widened they were lengthened out until they were almost end to end. Two men working together could easily manage such a center. The objec tion, however, to this theory is that, although possible when the length of a course of infilling does not exceed, say, 10 feet, the top courses of many of the great French vaults are more than double that. This would entail a center too heavy to be managed by hand.1 Moreover, it is the top courses which most require 1 The upper courses were very likely laid on single boards, propped up temporarily from a light scaffolding below. Quick-setting mortar would be used, and workmen after setting a course in one compartment could pass on to a course in another, giving the first time to set. It must be remembered that the web is often very thin. Viollet-le-Duc says that in the vaults of Notre Dame, Paris, it is only 10 centimetres, less than 4 inches, thick, 86 A HISTORY OF ARCHITECTURAL DEVELOPMENT. support when building ; the joints of the lower ones are more horizontal, and the stones, in consequence, rest more on the ones below. The adhesive power of the mortar alone almost keeps the latter in their places. Extra It is more than probable that it was the desire to avoid long ribs- courses which led the English architects to adopt the intermediate ribs (see p. 88, fiercer on) termed superfluous by some French and American writers, and so reduce by more than half the width of the compartments of the web. Thus, in Exeter Cathedral and in the nave of Westminster Abbey, the top courses of the web are not 8 feet long. That intermediate ribs were introduced to facUitate the building of the web is, to some extent, proved by the many churches in which the nave vaults have them, while the aisles have them not. Westminster Abbey and Lichfield Cathedral are cases in point. Extra ribs began to appear in EngUsh vaults early in the thirteenth century. The first was the ridge rib, sometimes called the lierne ; and this was rapidly followed by intermediate ribs, or tiercerons, wliich spring from the same spot as the diagonals, and divide the four big compartments of each vault into several smaUer ones. VioUet-le-Duc attributes the introduction of the ridge rib to the desire of the EngUsh to cover the ugly joint at the summit of a vault which results when the courses of stone on each side are not parallel. But it has abeady been pointed out that in many EngUsh vaults the courses are parallel ; and it should also be remembered that even when such is the case, some coaking is required to make the two sides meet satisfactorily, because the surface of the web is not straight, but concave. Moreover, in so many vaults the stones were not intended to show, but were plastered and painted, that another reason may be suggested. Reason for When vaults are constructed with semicircular or semi- n ge n s. e]|jpticaj arcnes there is no ridge line ; with the pointed arch there is. The suggestion is consequently not unnatural, that the need for emphasizing this soon became evident to the artistic mind of the English builder, and he therefore introduced the ridge rib, irrespective of how the stones of the web were laid. Early examples of it in England are in Lincoln and Chester Cathedrals, Westminster Abbey, etc. In Southern France it was common in the vaulted churches of the end of the twelfth century, such as S. Pierre, Saumur (see Fig. 164), S. Eadegonde, Poitiers, Poitiers Cathedral, etc. The three great bays of the nave of Angers S^SS»1 B V\%v flu l\ \3 i^^B ^K ^| IvV [1 ^1 lire ^k iMSa] ^^¦\il gBR U A ^»l i%* - '* 1 BBSr1 ^j^H ¦ SI SPHk^^B; t J/M ¦ B'| |B^j^|bl>. ^t %& ^B liLmW/f^A ¦ fj WW WHW^^ I1 f "i Photo : E. K. Prideaux. Fig. 55. — Tierceron Vault of S. John's Chapel, Exeter Cathedbal. Photo: E. K. Prideaux. Fig. 56. — Exeteb Cathedbal: Vault of Nave. [To face p. 86. VAULTING. 87 Cathedral (c. 1150) are without ridge ribs, but the later choir vaults have them. In the above-mentioned foreign examples the ridge ribs (like the diagonals) are extremely slender, and were probably made so because otherwise they would have emphasized disagreeably the dome form of the vaults. In England this had / -D \ < VAVLT5 WITH RIDGE RIBS. > < VAVLTS WITH TIERCERON. > K riORWICM. WINCHESTER. VAVLTS WITH LIER/1E5. > o- BOSSES. A-TRA/1SVER5E ARCH. B-WALL RIB. C- DIAGONAL RIES. D- RIDGE RIB. E-T1ERCERO/1.E- LIERA1E. W DIAGRA/AS OP RIB DEVELOPMENT. Fig. 57. not to be guarded against, as the ridges are straight in most Gothic examples, the exceptions being mainly those in which they slope upwards from the central boss to the sides, in order to provide greater height for windows. The architects of Northern France used ridge ribs sparingly, and to the last regarded them as not absolutely necessary. 88 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Tier- cerons. Liernes. Fan- vault. In oblong bays of vaulting, one tierceron is generally introduced in each central compartment, and two or more in each side one. The French, who delighted in overcoming structural difficulties by sheer tours deforce, and to whom compromise was distasteful, rejected these ribs as useless and unconstructional. But they are not so really. Their function has aUeady been indicated; and if the English method is to be regarded as a compromise, a man must be indeed bigoted who will not admit that it is a very beautiful one. Whenever tiercerons are used, ridge ribs are abso lutely essential, because, as Sir Gilbert Scott has pointed out, " without them the point at which the intermediate ribs meet at their apex would want abutment." 1 The Englishman's next introduction was a purely ornamental one, namely, small ribs to connect together the principal and inter mediate ribs, and so form with them starlike and other patterns.2 The spaces between the ribs are often now so smaU that each can be fiUed by two or three pieces of stone, as in the nave of Norwich. The original idea of a vault as consisting of a framework and a filling-in between of small stones, was before long lost in a web of liernes which covered the whole surface of the vault, no one rib being any stronger than its neighbour, as in the choirs of Gloucester and Tewkesbury. The next development, in perfectly natural and logical sequence, threw it over altogether, and resulted in an ashlar stone vault, in which rib and panel are worked on one and the same stone. One result of the change was that the vault once more became of one substance. It had not the homogeniety of the Eoman concrete vault, but it exercised far less thrust than the strongly-ribbed vaults of the thirteenth and fourteenth centuries, and consequently required less abutment. Thus was born the van- vault, so called because its ribs take the form of a half-opened fan. It is an essentially English design, and an appropriate and beautiful ending to centuries of vault develop ment. The earliest example of it is found in the cloisters of Gloucester Cathedral (c. 1370), the main arches being pointed. In nearly all other examples of fan vaulting, Henry VII.'s Chapel, Westminster, S. George's Chapel, Windsor, Eton College 1 In his lectures on " The Rise and Development of Mediaeval Architecture," vol. ii. p. 208. Ridge ribs are sometimes omitted in English vaults between the tiercerons and the wall ribs, as in Linooln Cathedral. 2' These small ribs are generally also called " liernes." The term is applied to any ribs which do not start from the springing. PLA;N TROTA BELOW LOOK1/NG VP WARDS. HALF TRANS VERSE SECTION. SCALE OF FEET. IO 5 O 5 IO I I I. I I I I I I I I 1 L EA/N VAVLT IMG from HE/NRYVII CMAPEL, WEST/A1/N5TER. rR£V\ WILLIS, vavlT3 or Tfic MlCTOl^C ACEs'i, Fig. 58. 90 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Relationof wall rib to window. Bosses. Chapel and King's College Chapel, Cambridge, they are four- centred. The builders, deUghted with their new form, indulged in many charming fancies. They carried down the ribs so as to form pendants, as in the side chapels of Henry VII. 's Chapel ; they started strong transverse arches, letting the lower part show and concealing the upper part, as in the central portion; they cusped the heads of the panels and carved the Uttle spandrils above, until the vaults excelled in richness any Eastern one. The pendant was by no means a freak ; it covered with its supporting bars what otherwise — in the case of four-centred vaults — would have been a flat and uncomfortably weak-looking centre. Sir Christopher Wren, in some of his pseudo-Gothic churches, such as S. Mary Aldermary, London, formed a saucer-dome over this space, but it is hardly an improvement on the earUer form. One advantage of the four-centred arch in vaulting is that the wall rib can be made concentric with the head of the window below it more easily than when more pointed forms are used. In earlier vaults there was often a difficulty in bringing these two into harmony. In some cases it is solved by the aid of a scoinson arch (see p. 68) over the window on the inside face of the wall, and this is the most natural method when the window is not the full width of the bay. In late French cathedrals, however, and in a few EngUsh ones (York nave, for instance), the entire space below the vault down to about the springing line of the waU rib is fiUed by the window-head, and the rib then forms its natural outer member. l In the nave of Lichfield Cathedral the difficulty is overcome by making the clerestory windows spherical triangles, but the experiment is not altogether satisfactory. Much the same thing occurs in the triforia of Westminster Abbey. Bosses, or key stones, are usual at the intersections of all ribs, except in Eomanesque vaults, in which they are rare. Short lengths of all the meeting ribs — 4 to 6 inches long — are worked on the boss, so as to form skewbacks, and the web butts against, or more often rests on, the solid central mass, so that thorough abutment is provided at the apex for both ribs and web. Some of the finest of Gothic carvings are on bosses. In the simple quadripartite vault, bosses only come at the intersection of the diagonals, as in the naves of Chichester and Salisbury Cathedrals, 1 In French work this was all the easier because the springing line of the wall rib is always far above that of the main ribs, which is not so general in England. It is so at York, however. VAULTING. 9' where they appear as jewels in an immense setting. As soon as ridge ribs were introduced they became necessary at the apex of transverse arches; and with the advent of tiercerons, at the points where these cut the ridges (Fig. 55). With liernes connect ing together the main ribs, they multiplied enormously, and also varied considerably in size, there being great difference in this respect between the bosses at main intersections and those at subordinate ones. In the latter positions they could not have been dispensed with, because the liernes often cut so awkwardly into the sides of larger ribs, that but for the bosses the junctions would have been exceedingly ugly. The following Ust states the character of the vaulting, etc., Summary in some of the English cathedrals, etc. : — examples. Timber roof.— (Ceiled) Naves of Ely, Peterboro', S. Albans; (open) Eochester. Barrel vault. — S. John's Chapel, Tower of London, S. Cathe rine's Chapel, Abbotsbury. Quadrant barrel vault. — Gloucester choir triforium. Intersecting vault. — (Plain) Eochester, crypt ; (with transverse arches) Durham Castle, crypt ; Gloucester, aisle of choir ; S. ¦ Albans, aisle of nave ; S. John's Chapel, London, aisle. Bibbed vaults — (1) Simple four-tite. Durham, nave and aisles ; Chichester ; Fountains' Abbey, cloisters ; Salisbury, nave, choir, cloisters, etc. ; S. Saviour's, Southwark ; WeUs, nave ; S. Cross, Hampshire, chancel (vault is domical) ; Boxgrove Priory. (2) With ridge ribs. Chester, choir ; Gloucester, nave ; West minster Abbey, choir ; Worcester, choir. (3) Six-tite vaults. Canterbury, choir ; Lincoln, south transept (with ridge rib). (4) Tierceron vaults. Ely, presbytery; Exeter, nave and choir ; Hereford ; Lichfield ; Lincoln, presbytery and nave ; Westminster Abbey, nave ; Worcester, nave. (5) Tierceron and Uerne vaults. Chester, nave; Ely, choir and Lady chapel ; Winchester, nave ; York, choir ; Worcester, cloisters ; Norwich, nave ; S. Mary Eedcliffe, Bristol. (6) Liernes. Tewkesbury, choir and crossing ; Gloucester, choir; Wells, choir. Christ Church, Oxford, con structed much like central vault Henry VII.'s Chapel, 92 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Setting out of ribs. Westminster, and Sherborne Abbey Church, form con necting links between Uerne and fan vaults.1 Fan vaults. — (With pendants) Henry VII.'s Chapel, central portion and side chapels ; (without pendants) King's College Chapel, Cambridge ; Bath Abbey Church ; Gloucester, cloisters ; Winchester, Bishop Waynflete's Chantry Chapel. The next points to consider are the setting out of the ribs at the springing of a vault; the counterpoise necessary to resist their thrusts ; and the disposition of the shafts from which the ribs spring. In Eomanesque work the different ribs are independent of one another, and each has its separate skewback. In Gothic RIBS DETACHED AT SPRIMCI/SC. Fig. 59. work this was impossible, because, in English work at least, the ribs are more numerous, and all the parts of a church below the vaults so much smaUer. Nor did the builders desire it. They realized that concentration of all thrusts was the first essential ; and that so long as each rib acted independently, the satisfactory neutralization of their thrusts was extremely difficult. They therefore united all the ribs together, for a certain number of courses above the springing, up to a point at which their separation became imperative. These courses are all built with 1 The Sherborne vaults are practically fan shape, but the Hemes are not curved as in fan vaults, they are straight. VAULTING. 93 horizontal beds, the top surface of the top course being splayed to the necessary angle to give each rib a skewback to spring from. The number of courses varies in different examples according to the shape of the vault, and its pitch. In Westminster Abbey there are seven courses ; in Lincoln Cathedral choir as many as ten; in other examples only three or four. In a square vault the ribs separate from one another much sooner that in an oblong; and in a steep or much stUted vault they remain in contact longer than in one of flatter pitch. The number also depends, to a considerable extent, on the point from which the diagonals start. In many vaults the centre line of each diagonal meets the centre line of the transverse arch on the wall plane (as shown on Fig. 60) ; sometimes the lines meet considerably behind it, and occasionally even in front of it. There is no hard- and-fast rule as to this. The meeting point depends entirely on whether it is desired to compress the ribs at the springing into the narrowest possible compass, or to spread them out. These bottom courses of the vault are coUectively called the Tas-de- " tas-de-charge." The principle it embodied was not new. In c arge" Egyptian, Byzantine, and Sasanian barrel vaults the bottom courses are often laid with horizontal beds up to an angle of about 30°. In Eoman intersecting vaults the diagonal ribs, men tioned in Vol. I. (Fig. 74), spring from a tas-de-charge of concrete ; and in some Eomanesque ones without ribs, the groin lines start from one of stone, as in portions of Winchester Cathedral (see Fig. 44). But to the Gothic men belongs the credit of applying it to ribbed vaults.1 The merits of the tas-de-charge are : (1) that it unites all thrusts and simplifies the question of counter poise ; (2) that it reduces to a minimum the danger of the slipping or crushing of any one rib at its springing, and the consequent fracture of the vault ; (3) that it lessens both the span and the rise of the active portion of the vault, viz. that built with separate voussoirs, and therefore the amount of centering required; (4) that it brings the thrusts well within the wall, the tas-de-charge itself acting as an abutment, and taking a good deal of the thrust. In aU early French Gothic vaults the wall ribs do not start French from the same springing Une as the other ribs, but from a point ^,lisIl 1 To determine the point at which the tas-de-charge ends and independent waU n°s' voussoirs commence, draw lines from the points at which the ribs intersect on plan, at right angles to the direction of each. Where these lines cut the extrados of the ribs is (approximately only when the ribs are of different depths) the top of the tas-de-charge. 94 A HISTORY OF ARCHITECTURAL DEVELOPMENT. far above it; a shaft, detached in most cases, resting on top of the capital of the main vaulting shaft, and carrying the rib. The English never employed the upper shaft, and in many of their VAULTING. 95 early vaults the wall ribs spring from the same point as the others. In nearly all later ones, however, the rib is much stilted, as in the nave of York Cathedral, and the result, therefore, is the same as in French work. There can be no doubt that the principle of separating the wall ribs from the others is a sound one. Their thrusts, unlike those of other ribs, are contained within the buUding ; none passes to the outside, and has to be neutralized. The thrust of one rib is met by the counter thrust of the next, and when these thrusts meet above the point where the other thrusts are focussed, they exercise a downward pressure which has a steadying influence. Moreover, the absence of wall ribs from the tas-de-charge undoubtedly facilitates the concentration of the thrusts of the others, as the area of thrust is lessened. StaticaUy, therefore, it is right, but aesthetically it is of doubtful advantage ; as it occasions considerable cutting back of the web, in order that this shaU not come in front of the wall rib. There is no difficulty in providing sufficient resistance to the Counter- thrust of a barrel vault, especially in a bunding of a single span. ^°J.srei The wall being continuous below the vault, all that is needed vaults. is that it shall be sufficiently thick. In a building divided into nave and aisles the continuous thrust of the central vault is counteracted by continuous aisle vaults, either semicircular, as in S. John's Chapel, Tower of London, and the church at Car cassonne, etc., or of quadrant shape, as in the narthex of S. PhiUbert, Tournus, and many other churches in Southern France (see Figs. 150 and 151). When the central vault is divided into bays by ribs, corresponding ribs are often buUt under the side vaults, and on the outside of the aisle walls are flat buttresses, which, however, are not much more than complimentary, owing to their slight projection. The task of counteracting the thrusts of intersecting vaults, Counter- especially ribbed ones, is not so easy, and it is no wonder that {^rsect the early mediaeval buUders did not solve the problem straight ing vaults. away. All the thrusts being concentrated at given points, the wall in between these points is of little assistance. This was recognized so thoroughly after the twelfth century, that, in France especiaUy, the wall became window. For the same reason a series of supports was substituted for continuous abutment. These stretched across from the back of the nave wall to the inside face of the aisle wall, under the aisle roof. They are not solid walls, for that would have weighted unnecessarily the 96 A HISTORY OF ARCHITECTURAL DEVELOPMENT. transverse arches over the aisle below, and also have interfered with free passage in the triforium, but are, as a rule, quadrant arches. They are, in fact, flying buttresses, although hidden.1 The hidden buttress possesses one advantage over the exposed one, that it is less liable to decay. It was sufficient so long as the churches were comparatively low, or when the triforium above LAO/^I CATH. Fig. 61. the aisles was a spacious gallery, and the roof over consequently as high as the springing of the nave vault, as at Laon, in Northern France ; but it ceased to be effective when the upper gallery was omitted, and the soaring ambition of the French architects led to greatly increased height for all parts of a church, the central nave especially. 1 Exactly the same abutment was provided by the Romans in their great vaulted buildings, such as the Basilica of Constantine. VAULTING. 97 The builders did not trust entirely to these hidden buttresses to counteract the vault pressure. As a matter of fact they relied more on the thickness of their walls, and on the height that these were carried above the springing of the vault to the underside of the tie-beam of the timber roof above. They knew that any weight above a point on which there is a side thrust neutralizes, to some extent, that thrust, and directs it downwards. They also, as an additional precaution, weighted the extrados of the vault above the tas-de-charge, so as to keep the thrusts still more within the walls. With the modest dimensions of our cathedrals these hidden Flying buttresses, coupled with the other precautions mentioned, sufficed ; t^s*seg- at all events at first, as at Worcester and portions of Lincoln, SaUsbury, etc. But the faUure and sometimes complete collapse of early vaults soon showed that more direct resistance was required. In many cathedrals, both in England and abroad, extra buttresses have been added which rise above the aisle roofs. In aU large churches abroad, commenced after the latter part of the twelfth century, visible flying buttresses form part of the original design. Flying buttresses transmit thrusts rather than resist them. But- Starting from the upper part of the nave wall — as a rule, from transmit sUghtly projecting buttresses — they conduct the thrusts to but- thrusts. tresses built out from the aisle wall. In N6tre Dame, Paris, which has double aisles, the principal flying buttresses are in single spans. As a rule, however, in double-aisled churches, the portion of the wall over each pier between the aisles is carried up to form an intermediate buttress, so that each flying buttress con sists of two arches. The first arch transmits the thrust of the vault from the nave wall to the intermediate buttress, and the second carries it over to the outer buttress. This has to be sufficiently strong to resist the thrust of the flying buttress itseU, and such transmitted vault thrusts as have not been neutralized on the way. There was evidently no hard-and-fast rule to determine the Relative relative position of the flying buttress to the vault. This differs gy^0"^! considerably in the various examples. The builders evidently did tresses to not make careful calculations to ascertain the amount of pressure v ' exercised, the exact direction it took, or the precise point from which it started. They judged from experience, from failures, and from successes. It is a question whether such calculations, if made now, would be of any use, because, given one single VOL. II. H 98 A HISTORY OF ARCHITECTURAL DEVELOPMENT. instance of bad workmanship, or of defective material, and the calculations are upset, unless an enormous margin of safety is allowed (see Figs. 61 and 62). In the majority of examples the top of the intrados of the arch of the flying buttress — or of the most important one when there are several — i.e. where it butts against the nave or choir wall, is, approximately, on a line with the top of the tas-de-charge. Its position varies, however; sometimes it is above it, at other times below it, the latter being most general in early, unsatis factory examples. The pitch of the vault (the thrust of a steep vault is more vertical than that of a flatter one), the height of the tas-de-charge (this varies, as abeady stated), and the weight of the wall above the springing line of the vault, aU these must be taken into account in determining the exact position of the flying buttress. Again, the depth of the buttress from soffit to top of coping, and the slope of the coping, make considerable difference. In early buttresses the bottom curve is often a quadrant, and the top slope of flat pitch. The mass of masonry against the wall which offers resistance to the vault thrust is consequently smaU. Experience showed that this form was not satisfactory, and the tops of later buttresses are consequently of steeper pitch, and the underside curves are equivalent to half a pointed arch. The early form was most unsatisfactory when the flying buttress was low down; for then, if the upper part of the wall ceased to be perpendicular, owing to the thrust of the vault or other causes, the top of the flying buttress dropped and began to press hard against the wall, with the result that it soon collapsed, being subjected to a thrust outward above and to a thrust inward below. In many later buttresses, and in some earUer ones, there is often a con siderable space between the top of the voussoirs composing the soffit curve and the bottom of the coping. This was sometimes solid, and was plain or had sunk panels on its face, sometimes, especially in late work, it was pierced. Such a buttress provides a considerable margin when any uncertainty exists as to the exact direction of the vault thrusts. Moreover, it acts over a large surface, and consequently affords greater support to the wall. Flying Flying buttresses are not used merely to resist or transmit a"gtifi.ses vault thrusts. In the lofty cathedrals of Northern France, such eners. as Beims, Amiens, etc., where two flying buttresses come one above the other, the upper one has no relation whatsoever to the vault. Its function is simply to aid in supporting the high wall VAULTING. 99 of the clerestory and to take the thrust of the timber roof above the vault, which, owing to its construction, is often considerable. At Beauvais, where there are three tiers of buttresses, the lowest is far below the vault, the middle one slightly above the tas-de- charge, and the top one close under the top parapet. The first and CATHEDRAL CK055 5EXTIO/1 SCALE OF FEET. 150 -| Fig. 62. last have nothing to do with the vault, they are merely stiffeners to the wall. It must not be supposed that the builders, when they adopted flying buttresses, dispensed with the other methods of counteract ing the thrusts mentioned on a previous page. On the contrary, they developed them. The side walls of the nave were still carried up above the vault, in many cases far higher than before. In Eeims Cathedral there are several feet between the top of the vault and the bottom of the tie-beam, and this means the same ioo A HISTORY OF ARCHITECTURAL DEVELOPMENT. MALME5BVRY ABBEY CHVRCH. PLY I r\ G BVTTRESS A/ID PI/M/MACL.E. amount of extra wall at the sides and a large increase in the weight pressing on the vault thrusts.1 Pinnacles. In addition, the builders multi plied pinnacles, the statical advantages of which have already been men tioned (see p. 49). They placed them above nave walls ; over intermediate buttresses — when the church had double aisles — and above the outside aisle buttresses. Many early flying buttresses are without them, but it was soon reaUzed that they effected a considerable saving in material. An aisle buttress weighted heavily on top can be haU the size of an unweighted one. Their aesthetic value is also immense. In Malmesbury Abbey Church the pinnacles are not only particularly well placed to counteract the thrusts, but are also extremely graceful. They are set back from the outside face of the aisle wall, start from above the haunch of the flying buttress, and are strengthened by gables in front which add beauty as well as stabUity. Pinnacles did not play so important a part in EngUsh work as in French, where the greater height of the churches, the apsidal east ends, and the frequency of double aisles led to multiplication of flying buttresses and consequent multiplication of pinnacles. In the opinion of some, the number abroad is sometimes too great, and the effect over-done; but there can be little doubt about the Fig. 63. 1 This additional height was probably chiefly for effect. In England, where great height was not desired, there is no great space between the vault and tie-beam ; sometimes the tie-beam rests on the vault. VAULTING. 101 success of the abside of Le Mans Cathedral, where flying but tresses and pinnacles combine in unusual profusion, owing to the forking of the buttresses, between the projecting apses (Fig. 97). The principle of thrust and counterpoise may be briefly summed Sum- up as foUows:— mary- 1. The tas-de-charge unites all thrusts and consolidates them at given points. 2. The tas-de-charge itself offers resistance to thrusts and also decreases the amount of thrust, inasmuch as the span and height of the " live " portion of the vault are lessened. 3. A high waU above the point of pressure tends to divert a lateral thrust into a vertical. 4. Pinnacles act in the same way. 5. Flying buttresses transmit thrusts to abutments. 6. Abutments weighted by pinnacles can be smaller than when unweighted. Only two points remain for consideration : one, the arrangement of the vaulting shaft ; the other, the different points from which the vaults spring. In early mediaeval work the vaulting shaft starts from the Vaulting floor and, forming part of the pier, passes through the triforium s a ' to the springing of the vault. This is the arrangement at Durham, Chichester, the Abbaye-aux-hommes, Caen, and in most vaulted Eomanesque churches of Italy, France, and Germany. As the desire grew for greater lightness and smaller piers, this English method was abandoned in England in the thirteenth century and never again became general here, although in parts of Lincoln and Lichfield Cathedrals, and in the fourteenth and fifteenth-century work in the naves of Winchester, Canterbury, and York the old plan is followed (see Fig. 219).1 At Exeter, Ely, Chester, the choir of Lincoln, etc., is seen the typical EngUsh arrangement. In these cathedrals each vaulting shaft springs from a carved or moulded corbel, which nestles between the arches of the nave arcade, immediately above the capitals of the piers. It is a beautiful and at the same time a perfectly sound structural device. Sometimes the shaft starts higher up, from the string course under the triforium, as at Carlisle, with merely a very small corbel below. In Salisbury and Wells Cathedrals it starts higher still, from above the triforium, and possesses little importance. 1 In Winchester Cathedral the old Romanesque piers were remodelled, which, to some extent, accounts for the shafts starting from the floor. method. methods. 102 A HISTORY OF ARCHITECTURAL DEVELOPMENT. The arrangement in these last two churches emphasizes the fact that verticality was not the sole aim of English workers. By the omission, practically, of the shaft they lost the appearance of height which vertical lines give, but the gain in breadth of effect was ample compensation. French In France the design depends somewhat on whether the supports are simple columns, as in Notre Dame, Paris, and Laon Cathedral, or whether they are columns with attached shafts, as at Beims and Chartres. In the former case the shafts, either three or five in number, start from the top of the capital of each column. To aUow space for these the wall has to be considerably thinner than the diameter of the column below, as already men tioned (see pp. 33 and 34). The shafts in these two examples could not have started from the floor without disturbing the cylindrical form of the column which the architects were anxious to retain. In Sees Cathedral a detached shaft is carried down in front of the column, the effect produced being particularly good. . In the choir of Canterbury Cathedral, the work of WUliam of Sens, the same arrangement is foUowed as at Paris. In cathedrals with shafted columns, the shaft on the nave side is sometimes carried up its full size, as in Amiens Cathedral, with merely the abacus of the capital carried round it; a subordinate shaft on each side of it starting from the top of the capital to support the diagonal ribs. In Beims Cathedral this shaft is stopped by the capital, and the vaulting shaft above it is more slender. In later work, however, no attempt is made to lessen the feeling of verticaUty. In S. Ouen, Eouen, and in the naves of Auxerre and Tours Cathedrals, three or more grouped shafts form part of each pier and, starting from the floor, run without a break to the springing of the vault, thus adding considerably to the apparent height of the church in each case. Springing There is no hard-and-fast rule to denote the point at which the vaulting shaft shall end and the vault begin. In the Cathedrals of Salisbury, Exeter, Ely, Lichfield, Bourges, Paris, Laon, etc., the vault springs from the string between the clerestory and triforium. At Durham and Lincoln it springs from below it, and from above it at Chichester, Chester, Wells, etc. At Amiens and Beauvais Cathedrals the springing line is nearly halfway between the string and the top of the clerestory windows, which accounts, to some extent, for the great height of these two examples. As a general rule, late French churches follow the example of the two last-mentioned cathedrals. a CHAPTER VII. TOWERS AND SPIRES. Church towers may be placed singly — (a) over crossings, (b) at Position west ends, central with a nave, (c) at either side, sometimes near of towers- the west end, sometimes near the east, in some cases starting from the nave walls and spanning the aisles, in others attached to the aisle walls only, and (d) detached entirely from the church to which they belong. They may be in pairs — (a) at west ends, (6) at ends of transepts, (c) over transepts, (d) flanking either eastern or lateral apses, (e) at the extreme east end, as at S. Nicolo, Bari, although this position is very rare (Fig. 123). A central tower over the crossing in a cruciform church gives Central the roofs of nave, choir, and transepts something to stop against. towera- One in a church without transepts allows the roofs over the western and eastern ends to start from and rise to different levels without an ugly break resulting. There is, therefore, a sound aesthetic reason for a central tower, except when a church is so lofty as to render it an undesirable addition, as in the majority of large French churches. In England central towers are the rule in cathedrals, and there are also some striking examples in Nor mandy. The Germans seldom bunt them ; they preferred octagonal cupolas over crossings, Uke those in Northern Italy. In their Eomanesque work they excelled in the grouping of these cupolas with circular turrets, of which there are generally two at least, and often as many as four, in a large church (see Fig. 134). A possible objection to a central tower is that it requires piers to support it considerably larger than are needed elsewhere. But in a cruciform church with transepts as high as the nave and choir, even when there is no central tower, the piers at the crossing stiU have to be stronger than other piers in order to resist the thrusts of the arches of the arcades. In the eastern transepts of SaUsbury Cathedral the arcades are continued across, and over each arch is an inverted arch, thus obviating the necessity for stronger io4 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Westerntowers. Towers over transepts. piers. When the nave and choir are sufficiently wide, and are also long, larger piers at a crossing are sometimes an advantage, as the deeper arches they carry form pleasant breaks in a vault, and these, coupled with the projection of the piers, improve the per spective and tend to heighten the mystery of the interior. Central towers are most effective from inside, when they are also lanterns, i.e. so arranged that the roof or vault over each is sufficiently high to allow of windows in their sides above the adjoining roofs, which windows throw a flood of light into the middle of a church. Eomanesque central towers, so far as can be gathered from existing remains, were generally lanterns, although many, like Tewkesbury Abbey, now have later vaults under their windows at about the level of the main vaults. The central towers of Lincoln, Canter bury, and York Cathedrals, and Pershore Abbey Church, are lanterns, and amongst continental examples, four of the finest are at Coutances and Evreux Cathedrals, S. Maclou, Eouen, all in Normandy, and Braisne Church, between Soissons and Eeims. The most effective and suitable position for a pair of towers is at the west end. They frame in the principal doorway and the central part of the front, and, in addition, they mask the aisles at the sides. The end elevation of an aisled church is seldom satisfac tory when its lines follow truthfully the sectional outUne behind it. To conceal this was one of the chief aims of the mediaeval buUders, and in choosing a pair of towers for this purpose they showed sound judgment. For the same reason, the transepts of many French churches, which have an aisle on each side, were flanked by twin towers until, for circumstances which are suggested later (see p. 277), they were finally abandoned. Towers over transepts are far from common, and yet this position has many advantages. The towers can rise from soUd walls, and have not to be balanced aloft on piers and arches ; they can be built independently of the rest of the church, and can be allowed to settle as they please ; their full height shows ; they cause no cramping of the width at the crossing; and they add considerable stability instead of being, as is often the case with a central tower, a source of weakness, at times even a positive danger. Moreover, provided they are of fair height and not too far apart — they should start close against the aisles — they provide more striking grouping than even twin towers at a west end. Towers come over the transepts in Angouleme Cathedral, in a few other churches in Southern France, and in Treguier Cathedral, Photo : Author. Fig. 66. — Coutances Cathedral: Lantern. Photo: Author. Fig. 67. — Evreux Cathedral : Lantern. [To face p. 104. Photo: Author. Fig. 68.— Beaisne Chubch: Lantebn. Photo: Author. Fig. 69. — S. Maolou, Rouen : Lantebn. [To face p. 10 4 TOWERS AND SPIRES. 105 Brittany, where there is a central tower in addition. Exeter Cathedral alone in England has towers in this position. The two are practically all that remains of the Eomanesque church. In early days a pair of towers so placed was probably not uncommon, but beyond the examples mentioned there are but few others exist ing now. In later work they are stUl rarer, the largest and most striking examples being those of Vienna Cathedral. There are not many churches in wliich the eastern apses are Towers in flanked by towers, except in Germany where, however, they are poaitions. mostly more turrets than towers. S. Abbondio, Como, is an Italian example, the two towers rising above the aisle roofs at the junction TOWERS TREGVIER CATM: Fig. 70. of the nave and choir. In the church of Morienval, France, the towers are somewhat similarly placed at the east end of aisles which are cut off from the chancel by walls. In Tournai Cathedral towers crowned by spires flank the apsidal transepts with striking effect (see Fig. 140). No statement can be made regarding the most usual positions of detached towers. Some are close up to a church, others many feet away from it, as at S. Zeno, Verona. In France few large churches have single towers at the west end, Western S. Eicquier, near AbbevUle, being one of the chief exceptions. In ^ we* 3 ™ England a single tower at the west end is an old tradition. The number of Anglo-Saxon examples still remaining is proof that before towers. 1 06 A HISTORY OF ARCHITECTURAL DEVELOPMENT. the Conquest this position was general, which is all the more interesting as single western towers are rare in early work abroad. The Italian basilican church had a narthex at the west end, never a tower. Byzantine churches have no towers at aU, and although Eavenna boasts many early ones belonging to its basiUcan churches they are not central western ones. In Normandy, in the eleventh century, towers over crossings were the rule, and western ones, except in pairs, the exception. In the centuries immediately foUowing S. Augustine's mission to England, a tower attached to a church was almost a necessity in this country ; not so much for appearance as for refuge. In most viUages the church occupied the most central and highest point, and it was therefore only natural that attached to it should be the stronghold. A tower provided refuge for both people and priests, and although there was no particular reason for choosing a western position, this was at least as convenient as any other and had aesthetic advantages as well. Fortified towers attached to churches, and churches forti fied throughout, are neither rare abroad. Albi Cathedral is an example of the first ; the church at Boyat, Auvergne, with its machicolations all round, of the second. In England, in the debatable land on the Scottish border, there is at least one church, Burgh-by-Sands, Cumberland, which has a fortified tower, the only approach to which is from the church by a narrow opening, pro jected by a strong iron door. Whatever was the reason why the English so early adopted single western towers, they never tired of them. Ely is the only cathedral with one tower at the west end, but there are thousands of parish churches, of aU dates and in all parts of the country, with this feature. In fact, it remained the rule long after Gothic had passed away. English Before the Conquest some central towers had been built over crossings in England, as at Dover Castle, and after it very few large churches were designed without them. Amongst Eoman esque examples, the finest are those of S. Albans Cathedral, built with Eoman bricks from the ruins of Verulam close by ; Tewkes bury Abbey, exceedingly massive and bold ; and Norwich Cathedral, more slender, and crowned by a later spire. The smaU height of our Gothic cathedrals, i.e. small as compared with continental examples, rendered the building of a central tower comparatively easy. The result is that this feature, sometimes crowned by a spire, and sometimes not, is one of the most characteristic features of English church architecture. The pair of towers at the west TOWERS AND SPIRES. 107 end were not abandoned, but some early cathedrals, like Salisbury, and some later ones, like Winchester, are without them. Amongst cathedrals with both central and western towers are Durham, Canterbury, York, Lincoln, Wells, and Peterborough ; the western towers of the last-named, however, are small. Lichfield Cathedral is unique in England in having three towers crowned by spires. At Chichester only the central tower has a spire, not either of the end ones. Cathedrals with a central tower and spire, but no west towers, are SaUsbury and Norwich, and to these at one time could be added Old S. Paul's, London. Gloucester, Hereford, Worcester, and some others, have now central towers only. Some EngUsh towers at present without spires were intended originally to have them, and in a few cases did have them, but they have disappeared. The highest tower is the central one of Lincoln, which rises 262 feet above the ground, over 50 feet more than the western towers, which are also unusually high. The tower of Gloucester Cathedral, early fifteenth century, comes next in height, 225 feet. In beauty of design it is only surpassed in contemporary work by the central tower of Canterbury. Italy is the land of towers, and yet the Italians never appre- Italian ciated their fuU value as important integral parts of a design. They are generaUy detached entirely, Uke Giotto's CampanUe at Florence, and even when attached often appear like after thoughts. The one touching the side wall of the Cathedral at Trani has a roadway running through it. Fear of earthquakes accounts largely for the number of detached towers in Italy, but the fact that in the Eastern church towers were omitted altogether possibly also had some weight in determining the Italians not to make them part and parcel of their designs. A central tower over the crossing is almost unknown in the country, although there are frequently low cupola turrets at that point ; and twin towers at the west end are exceedingly rare. The earliest church towers in Italy, and therefore in Europe, are those of Eavenna. A few are square, but most are circular, one or two of the smaUest of the latter starting from square bases. The oldest is the one near the cathedral, which dates from the fifth century. The towers of S. ApoUinare in Classe and S. Apollinare Nuovo belong to the middle of the following one. That these towers were belfries seems certain. Cattaneo says that bells were undoubtedly used in the sixth century, but judging from the number of their windows, especially in the upper storeys, they may also have been towers. Photo : Alinari. Pig. 72. — Toweb op SS. Giovanni e Paolo, Rome. [ To face p. . 108. TOWERS AND SPIRES. 109 S.SATIRO,MILA/i . TOWER bands or string-courses. The tower of S. Zeno, Verona, is of the Venetian type, but it has two open loggie, instead of a single loggia, and it is built in alternating courses of brick and marble. The conical spire which crowns it is of brick.1 Lombardic towers differ chiefly from Venetian in that each storey Lom- has windows and is marked by a horizontal string-course, of the \^™s usual Lombardic type, which stops against pilasters at the angles. The tower of S. Satiro, Milan (c. 879), is probably the oldest square church tower in Northern Italy. It has four storeys, and the windows increase in size from narrow slits immediately above the ground, to two very wide openings on each side at the belfry level. The tower of S. Ambrogio, Milan, built 240 years later, is similar in design, but richer (see Fig. 113). Eoman towers are without pilasters, Roman but solidity is given to the angles by owers- keeping the windows well in the centre. They are divided by cornices which gird them, and take the place of the string-courses which in other towers generaUy stop against angle pilasters. The finest are those of SS. Giovanni e Paolo and S. Maria in Cosmedin, both of the twelfth century. In both towers the number of lights is the same in each of the four or five upper storeys. The tower of SS. Giovanni e Paolo is of brick, but the window shafts and capitals and the little corbels in the cornices are of white marble. As in other towers in Borne, roundels of green marble and porphyry are inserted in different places with excellent effect. A tower, unique in many respects, is the circular one at Pisa, Pisa. which leans 13 feet out of the perpendicular. The generally accepted theory regarding its non-verticality is that this is the 1 Many Italian spires are built of glazed brick of different colours ; the spire above the square tower of S. Giovanni Evangelista, Ravenna, being in alternating bands of green and white. Fig. 73. no A HISTORY OF ARCHITECTURAL DEVELOPMENT. CIRCVLAR TOWER, PISA. result of bad foundations, or of earthquakes. This view, however, is combated by Mr. Goodyear,1 who holds that it was mtentional from the first. The ground storey has a blind arcade, like the cathedral, and above it are six tiers of shafted and arcaded galleries with a belfry of smaller diameter on top. The upper storeys and belfry are nearer the perpendicular than the lower ones, showing either that an attempt was made when the tower was half built to correct the inclination, or else, as Mr. Goodyear claims, to give an entasis to the outline. It is difficult to endorse his theory, as it seems incredible that any architect could wilfully perpetrate so mad a design, or that any people should be so insane as to allow him to carry it out. That the tower has stood so long and so steadily is owing to the great thickness of the wall, which at the base is about 14 feet thick, and yet the internal diameter of the tower is only 25 feet. The galleries project about 3 feet, leaving 8 or 9 feet for the thickness of the waU above the ground storey. The Lombardic type of tower was the one at first followed in other coun tries, but more pronounced buttressing soon led to modifications of it. Cam panile towers pure and simple are rare in both England and France. A small but graceful example is at Yatton Keynell, North Wilts. The sides batter as in Italian examples. Most Fig. 74. Eomanesque towers have at the angles the wide buttresses of slight projection typical of the period, and otherwise are quite plain except at the top, where there are often one or more bands of arcading. The two small ones on the west side of the transepts of Canterbury Cathedral are particularly well-proportioned. There is no need to trace the variations introduced from century 1 In the Architectural Record for January, 1898. Photo: Author. Fig. 75. — Jumieges Abbey Chubch. Photo: .Vcnrdeili. Pig. 76. — Coutances Cathedeal : Centbal Lantebn Toweb. [To face p. 111. TOWERS AND SPIRES. in to century in tower design in England, as they are similar to those already described when dealing with other parts of churches. The finest late towers we have are those in the west, in Somerset, Gloucester, etc. Distinction is given to many of these by the pierced stone panels which fill the windows and take the place of the wood louvres which are customary elsewhere.1 These panels bear a strong resemblance to the slabs in the windows of Byzantine churches, and are the only instances in which the custom prevailed in Gothic work. Examples are in the towers of S. Mary Magdalene, Taunton, and Huish Episcopi, Somerset; Tiverton, Devon ; and ColeshUl, Berks. A characteristic French form of tower is that in which (the Charac- lower part being square) the upper part is octagonal. In front of French the canted sides are turrets, which preserve the square form to form. some extent. At Jumieges Abbey Church, Normandy, and S. Germain, Auxerre, are Eomanesque examples. The majority of towers of this type, however, belong to the thirteenth century, whUst a few are later, like the west towers of Beims Cathedral. The object undoubtedly was to faciUtate the starting of the spire, and prevent the somewhat abrupt change which takes place when a tower is continued square to the top and an octagonal spire springs directly from it. This plan is followed in the south-west tower and spire of Chartres Cathedral, and in the famous towers at Laon, where niches, which start from the buttresses below, project at right angles to the canted sides (see Fig. 196). The finest towers of this type are the three at Coutances. The angle pinnacles of the central tower are octagonal, and reach nearly to the parapet ; those of the two western towers are square, and are carried far above it. Although many French towers are now without spires, it seems probable that wood ones at least were intended in most cases. The spire of the south-west tower of Chartres is stone, and so are the spires of the two western towers of Coutances. At Reims a spire of the same material has been started over one of the western towers, but was never finished. An excellent way of finishing a tower, when funds do not Saddle- admit of its being carried high, is in the Church of S. Nicolas, Caen. j?aok The east and west ends of it have gables, with a roof between. These saddle-back towers are numerous round Soissons, France, and are far from uncommon in some parts of England, notably 1 In Prance the louvres of tower windows are larger than in England, and are frequently covered with small slates. 112 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Herefordshire. The tower of Braisne Church, near Soissons, has a gable on each of its four sides, and this was the general custom in German towers crowned by spires. There will always be differences of opinion as to whether the campanile towers of Italy, or the buttressed towers of more northern countries, are the finer. The former are certainly the more graceful, the latter the more robust. In a less clear atmo sphere than that of Italy, the delicacy of such a tower as the S. Zeno one would be lost ; and in that atmosphere the shadows cast by buttresses like those of the west towers of Canterbury Cathedral would be too strong. A tower of unbroken lines is TIA\E>ER SPIRES. A. SHEWING TIMBERS . B. " SHI/SGLES . C .TYPICAL GER^VA/N SPIRE Pig. 77. Spires. undoubtedly best for Italy ; one with bold projections has so far been regarded as most suitable for our cUmate. But it .does not follow that such will always be the opinion ; and as Classic feeling gains ground in England, we may ere long see the abandonment of features which in the North of Europe in mediaeval days were considered essential. The majority of Eomanesque towers were merely crowned by steep-pitched timber roofs, which graduaUy were elongated into spires. These were mostly square in plan, like those over the towers of Tournai Cathedral (see Fig. 140). Nearly all later spires, whether in wood or stone, are octagonal. The usual way TOWERS AND SPIRES. "3 in which an octagonal timber spire sits down on a square tower is very simple. A trimmer is inserted between each pair of hip rafters near the base, and from its junctions small trimming rafters are carried to the corner, with ordinary rafters in between. The triangle thus formed covers what would otherwise be an exposed angle. The same form is occasionally worked in stone, but it is most appropriate to wood.1 In most stone spires of the thirteenth century in England, each angle is bridged by a " broach," 5TO^E SPIRES. *=# Fig. 78. the top of which dies into one of the canted sides, whilst the bottom completes the square. The canted sides of a spire are carried on squinch arches thrown across from side to side of the tower to form the necessary octagon. Until near the end of the thirteenth century all spires in England started slightly in front of the face of the wall from projecting oversailing courses. In the 1 Wood spires are covered by lead, slates, tiles, or oak " shingles," according to the locality. The old shingled spires of Sussex, which have toned down to a soft grey, are particularly pleasing. VOL. II. I U4 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Construc tion of stonespires. Angle of spires. German spires. fourteenth century the custom became general of finishing a tower by a parapet, with the result that the spire started behind it. The buttresses at the angles, which hitherto had usuaUy stopped under the oversailing course, were now continued above the parapet, generally in the form of pinnacles, and in some cases flying buttresses were thrown across from the pinnacles to the canted sides of the spire. A broach was sometimes retained, but more often a flat slab covered each angle, leaving the passage-way clear. The sides of a stone spire are not more than from 6 inches to 9 inches thick, according to size. Near the top they are often thinned down to 4 or 5 inches. The actual top for a few feet is soUd, and finishes with a finial. All joints are horizontal, a spire being really built up in a series of corbels, and consequently exer cising no side thrust, only vertical pressure. All spires ought to have a sUght entasis to prevent the sides appearing concave, and many EngUsh ones have this refinement, although it is not uni versal. In the stumpy spires of Italy, the entasis is often so strong as to be disagreeable, and in early examples in France it is more marked than in this country. The later the spire the steeper its slope. In EngUsh spires of the thirteenth century, their angle is often not more than 73° to 75°; the timber ones being generally less steep than the stone, The spires of SaUsbury, Norwich, and Lichfield are about 80°, the Coutances spires 82°, the spire of Patrington Church, Yorkshire, is a trifle more (Fig. 251), whilst the spire of Louth Church, Lincolnshire, is as much as 85°. In England richly ornamented spires are rare. The Patrington Church spire has some arcading round its base, but as a rule the simple lines are aUowed to show from parapet to finial, banded in places and crocketed at the angles in fourteenth- century examples, but otherwise plain. The windows, or lucarnes, vary in design according to date, but are seldom more than single lights or pairs of lights, crowned by a gable. On the continent the spires are often more elaborate. The French generally covered their early ones with scalloping, etc., cut in the stone, and in the fifteenth century sometimes nearly concealed the whole of a spire under a network of Uttle buttresses, carved panels, niches^ and crocketed canopies. The two western spires of Chartres Cathedral are examples of the two periods, the south-west spire being early, and the north-west one late. The Germans paid especial attention to spires, and although TOWERS AND SPIRES. 115 they buUt few stone ones, the designs of their timber ones show considerable variety. Many are similar to the English tiled or shingle-covered spire. In the most typical form, however, the tower below is gabled on all four sides, and the spire starts from the apexes of these gables, and is continued down between them to their springing. The same is done over octagonal towers, the spires then being also octagonal. Sometimes the spire is octagonal above a square tower and sixteen-sided above an octagon. The question as to whether a tower alone or a tower and spire Tower or combined is the more effective depends on circumstances. In spire- mountainous districts spires are a mistake, unless they stand in a wide vaUey. At the foot or side of a hill their height is dwarfed by the mass rising above them, and on a summit they appear thin and their outline becomes blurred by distance. In the Lake district in England most of the towers are without spires and are broad and stumpy. Spires are most telling on flat or undu lating country, and for this reason are especiaUy numerous in the shires, in Leicestershire, Northamptonshire, Butland shire, etc. The home counties, Kent, Sussex, and Surrey, have also several. Towers abound in England in the north and west, although portions of the latter are flat. Boston " stump " is also a landmark for mUes round in the Fen district, and is probably more striking than if it had a spire. The fine effect both towers alone and towers with spires produce in towns, especially from a distance, may be well seen at Oxford, Eouen, etc. In mediaeval days, when cities were surrounded by waUs, they were the only features that showed above them. S. Bernard was wrong when he forebad towers in Cistercian churches, and the Benedictines were right in always adding them to theirs. For a church without a tower can be hidden completely by trees in the country, by houses in a town ; and what is out of sight is too often out of mind. CHAPTER VIII. MURAL DECORATION — SCULPTURE, CARVING, STAINED GLASS. The interiors of many mediaeval churches, as seen to-day — as clean and colourless as a hospital ward — give but a faint idea of what their appearance was originaUy. When built they were a blaze of colour from floor to roof. The windows were filled with stained glass, just sufficient clear glass being left to admit Ught and to frame the dazzling figures of saints, apostles, and martyrs ; the walls were painted, and the mouldings of arches picked out with colour and often gUding; the floors were covered with encaustic tUes, yellow, red, and brown. Puritanism, sanitary considerations, and restoration have changed aU that. In place of coloured wall, we too often have now newly scraped stone, clean new plaster, or, worse stUl, rubble walUng with the joints picked out, sometimes in black mortar. The brilUant glass has, in many cases, disappeared altogether, making our churches, especially fifteenth-century ones, more like conservatories than places of worship. The difference is as great as between a paint ing and an engraving ; the composition remains the same, the colour is absent. Continental methods have been even more drastic than those of this country. To give one example only. The east end of Laon Cathedral, a few years back, contained most interesting remains of early thirteenth-century and later decorations. The first was applied direct on to the stone, and this, in the fifteenth century, had been covered by a thin layer of stucco, which in its turn had been painted. The work afforded valuable examples of the methods employed in both centuries and of the disregard which later workmen showed for the efforts of their predecessors. The Httle that remained was only a smaU portion of what formerly existed there, but even that was doomed. It MURAL DECORATION, ETC. 117 had to go, said a workman, " pour le faire propre." 1 Cleanliness may be next to godliness, but if not tempered by judgment and respect for the past, it tends to destroy both art and history. What is wanted in old churches is the cleanliness of the painter, not that of the charwoman or "restorer." Colour can be introduced in a church either in the form of Methods appUed decoration, such as marble veneer, mosaics or paint, or by °}oneoora" the employment of different materials structuraUy. In the north of Italy brick and marble or brick and stone are interchanged with the happiest results. Farther south brick is abandoned, and marbles of different colours alternate, the effect being most satis factory when one colour predominates, and not, as at Siena, when bands of black and white marble are about equal in height. The old Eoman custom of covering walls with a thin veneer of marble, which was practised so successfully by the Byzantines, was also frequently followed in central and southern Italy. S. Miniato, Florence (Figs. 117 and 118), shows a particularly good example of this treatment, the slabs being smaU and the patterns varied. In France and Germany materials of different colours were seldom used. In England also stone facings inside are generally of one colour, although in parts of the country where red and yeUow stone exist side by side, often in the same quarry, arches and piers are some times of one and the walling of the other. English work, how ever, especiaUy that of the thirteenth century, is often marked by strong contrasts, owing to the use of detached shafts of marble in combination with white stone. In the south of England these shafts are generally from the quarries in the isle of Purbeck, Dorset, and in the north from Frosterley, near Durham, and else where. In the nave of Salisbury Cathedral and throughout the whole of Westminster Abbey the columns, surrounding shafts, and the capitals and bases are all marble. At Worcester only the shafts, the abacus of each capital, and top members of each base are of that material, the remainder being stone. The shafts in jambs of triforia, clerestory and aisle windows, doorways and porches are also frequently of marble. The contrasts now are in many cases too marked, especially at Salisbury, where the jump from the dark Purbeck of the piers to the white stone of the arches is too abrupt, and the general effect disagreeably cold. That this 1 My first visit to Laon was in 1896, when the scraping tool was getting perilously near the east end. I was there again last year (1907), after the above was written, and was glad to find that something had been spared. In La Trinite' Angers, so far as can be seen, not a vestige of old stone or colour remains. 118 A HISTORY OF ARCHITECTURAL DEVELOPMENT. frigidity is part and parcel of the original design is extremely doubtful. The white stone of arches and wall was most likely painted, probably in rich reds and occasional yellows, and the effect in mediaeval days was entirely different. Some of the most striking schemes of decoration in materials of different colours are in the volcanic district of the Puy-de-Dome, France, and in Eomanesque churches elsewhere, but these are mainly confined to exteriors, and are referred to later in the chapter deal ing with the churches of Auvergne. A similar treatment occurs in Chichester Cathedral in the tympana over the openings of the triforia, which in some cases are filled with lozenge-shaped stones of different colours. Mosaics. Whether mosaic work originated in the East, and thence was carried to Borne, or in Italy, is a disputed point; but it seems likely that although the Eomans were early workers in marble mosaic for floors, the art of glass mosaic commenced elsewhere. The latter certainly reached perfection under the Byzantine Greeks of the sixth century, and much subsequent work in Italy, down to the end of the twelfth century, was executed by Greek workmen. Amongst noted late examples of it are the blue Madonna in the apse of Torcello Cathedral, the head and shoulders of Christ in the apse of Cefalu Cathedral, the same design in a similar position in Monreale Cathedral, both in SicUy, and the wonderful band of figures round the nave, under the clerestory windows, in the last-named church. In S. Mark's, Venice, are also remarkable mosaics of the eleventh and twelfth centuries ; but these and other contemporary examples in Greece and Constantinople have already been dealt with in the first volume. In the thirteenth century the art died, except in the East, where it lingered iu places a short time longer. One reason why it was abandoned is that Eomanesque and Gothic churches do not lend themselves so well to applied decoration as basiUcan churches. The triumphal arch of the latter, with its fine expanse of wall over it, crying aloud for colour, disappeared in the desire to carry the vault from end to end without a break ; and the open triforium galleries at the sides occupied the space over the arcades formerly available for mosaics. A revival has taken place of late years, as the fine work in Westminster Cathedral shows, and there is no reason why this most durable form of mural decoration should not again become general.1 1 Although so largely used by the Byzantine Greeks on domes, mosaics were Photo: Author. "j Pig. 79. — Chichesteb Cathedbal: Tbifoeium. [To face p. 118. MURAL DECORATION, ETC. 119 The remains of early fresco paintings are not numerous, Fresco id *i 1 ii t i n p" but there can be little doubt that the art was practised in Italy through the first centuries of church-building. In the under ground church of S. Clemente, Eome, are fresco paintings of all periods, some as early as the fourth century, others as late as the beginning of the twelfth, when the present upper church was built. Examples in Eomanesque churches are few. The Eomanesque workman was a builder, he revelled in his bricks and mortar and stone; as a decorator he was inferior; but the absence now of paintings in his churches does not necessarily mean that none ever existed there. Many have doubtless been destroyed. These probably were not great works of art. Until towards the end of the thirteenth century the fresco painters produced little worthy to compare with the figures of the mosaic worker. Then came a Eenaissance, and in the work of Cimabue, Giotto, etc., in the Franciscan church at Assisi, in S. Croee, Florence, etc., the art reached its highest level. Before oil became the general medium for mural decoration Dis- early in the fourteenth century, work not executed in fresco was ^^J done in tempera, fixed sometimes by a final coat of oil. Tempera paintings. differs mainly from fresco in being applied to a dry surface and not to a damp one. Sometimes the work was done direct on stone, but more often a thin coat of stucco was laid first, as this afforded a more even and satisfactory surface to work on and prevented absorption. The process is an exceedingly ancient one, extending back to the days of Egypt, Greece, and Eome, and was practised all through the Middle Ages in all countries. The palette was an extremely limited one until the end of the thir teenth century, and consisted of little more than reds and yellows. There are many fragments of pictures painted in this medium in England, but all are more or less damaged. Many have been covered by later paintings, or by plaster, to form a key for which the earUer surface, plaster or stone, has been hacked. Others have been hidden by whitewash, applied either in the days when whatever was beautiful was held to be idolatrous, or during periods of great sickness. After the plague visitations, which were so frequent in the Middle Ages, disinfection of churches was as necessary as disinfection of houses, and much never, so far as I know, applied to the compartments of ribbed vaults. The art was moribund before they were invented, and the idea of applied decoration (save paint) was altogether repugnant to the mediaeval mason. 120 A HISTORY OF ARCHITECTURAL DEVELOPMENT. of the whitewashing generally attributed to the Puritans was very likely a precautionary method taken much earlier to prevent the spread of disease. Whatever the reason, the damage done was irreparable. Whitewashing mosaics only preserves them, but it spells ruin to both fresco and distemper. The paintings on the western sides of the nave piers of S. Albans Cathedral are fair examples of early work. So far as can be judged from somewhat scanty existing remams, such pictures only came at intervals, or in bands, the walling in between, above, and below them being coloured a plain tint, generally red in the thirteenth century, or else whitened, with double Unes of red or yellow forming rectangular spaces, which corresponded approximately to the joints in the walling behind. There was, however, no attempt at imitation jointing, and in the centre of each oblong was frequently painted a four- or five-leaved flower. POLEBROOK CIIVRCrl , rtORTt1A/M5. PAinTincs O/l SCREE/- DARK GREErt O/l LIGHT CHOCOLATE BACHGROW1B PATTER/1 R.EO GFIOV/IO LI/1E3 V/HITE & CHOCOLATE Pig. 80. Fifteenth- century muralpaintings. With the introduction of oil as the usual medium, greens and blues were added and predominated, and mural decoration became far more elaborate. Pictures on the walls do not appear to have been so general as in the thirteenth century, but many in this medium have survived on wood fittings, on triptychs, panels of pulpits, screens, etc. Walls were frequently covered with large diaper patterns, sometimes consisting of single sprigs of flowers — variety being obtained by turning the stalks and flowers different ways — -and sometimes of elaborate conventional designs in imitation of Florentine hangings. The mouldings of arches especially were all painted, the fillets being frequently Photo : Author. Pig. 81. — S. Vulpean, Abbeville. Photo : Author. Fig. 82. — S. Jacques, Lisieux : Painting, 1552. [To face p. 121. o MURAL DECORATION, ETC. 121 gilded. Not very much of this work remains on walls in England, but there are many examples of it on screens, stalls, pulpits, tombs, timber roofs, etc., especially in the Eastern Counties, Norfolk, Suffolk, etc., and in the west. The mediaeval builder had no respect for the natural colour or grain of oak. That feeling is a modern fancy entirely. Oak was the common, the only wood for all fittings, and for aU structural parts of churches requiring to be in timber. These were painted and gilded, so that in many cases not a square inch of the natural material showed, and the grain was obUterated entirely. In S. Albans Cathedral the shrine of S. Alban, although of Purbeck marble, is painted blue and green, the tracery and mullions being gilded. The painting on the mouldings of many a fifteenth-century oak screen affords an excellent guide to what the painting on stone mouldings was like, and enough remains on the latter in different churches, at home and abroad, to prove that both were treated alike. In stone vaults the carvings of bosses were generally gilded, and between the leaves or figures, as the case might be, was colour to emphasize them further. The ribs were painted, sometimes merely for a short distance from the bosses, and the infilling, when plastered, was treated in the same way, although few remains exist of mediaeval work on webs in northern countries. The fine decoration in S. Anastasia, Verona, is Gothic in feeling, but was executed after the Eenaissance had begun ; and the paint ings on the vaults of S. Vulfran, Abbeville, and S. Jacques, Lisieux, are sixteenth century, the latter being dated 1552. In no country can such extensive remains of old colour be Colour in seen as in Germany, and in none are modern redecorations and Germany- restorations — not always successful — so plentiful. The churches of Cologne, for instance, S. Martin, S. Gereon, etc., are extensively painted inside. One of the most satisfactory redecorations is in the porch of the Liebfrauenkirche, Niirnberg, which is a mass of colour, the figures and other portions being gUded solid. Such decorative schemes are not confined to churches. Eestaurants, hotels, private houses, etc., are often painted much in the same way as buildings were in the time of Diirer. As in printing, the Germans have preserved mediaeval traditions in decoration, and black-letter inscriptions on waUs, especiaUy in the smaller towns, are almost as common now as they ever were. The colour schemes of the mediaeval painters are a succession Colour of contrasts. A broad effect is obtained, not by the use of one soliemes- 122 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Quality of colour. Colour ex ternally. single colour, but by the juxtaposition of many colours, no one being allowed to prevaU over the others. One mouldmg may be green, the next red or blue, or white with smaU sprigs of flowers painted on it — the last is reserved for hollows — and the dividing fillets either gilded or painted a strong yellow. A complete accord must have existed between mason and painter, otherwise such schemes would have been impossible, as colour completely changes the appearance of mouldings. By its aid some can be emphasized, others almost obUterated. Considerable controversy at times has arisen over the quahty of the colours used. Some have held that it was crude in the extreme, and the present fine tones they attribute to time alone. As proof, they point to examples from which whitewash or plaster has recently been removed. In' thirteenth-century work the colours possibly were crude, although the linseed oil that was often the final coat must have toned them down somewhat. The late G. F. Bodley always maintained that, in the fifteenth century at least, although crude colours were applied first, a semi-trans parent glazing colour was added as a finish. His contention was that when whitewash has been removed from old work, the glazing coat has unfortunately been removed also, and that the colour re maining is very different in tone from the work when the painters left it. The patchy appearance of much old painting may be advanced in support of this, and even if absolute proof is not forthcoming, his suggestion is such a reasonable one that it may well be accepted. Fragments alone remain to enable one to judge of the extent to which colour was employed externally on churches, but these are sufficient to show that in most cases both it and gUding were appUed to the figures, canopy work, and walling behind in porches and doorways. Colour has to be sought for in churches for the entire renovation of which funds were luckily not available. The less renowned the church, the greater the likelihood of finding remains of decoration. The north and south porches of Chartres Cathedral are famous throughout the civUized world, and it is useless to search for colour in them. And yet there can be no doubt whatsoever that they were painted and gilded originally, but both paint and gold have been cleaned off with a thorough ness worthy of a better cause. It is only in the crevices that anything can be discovered at all, and even in these the patches are so minute that their nature is not clear. Over the west door Biiinni i rri :; «,, ¦Mill mi I f i r£r annum i ¦ i kjswj nii!ii< : i nn .... ^ ' 11 urn m i riV ' l!!im t(l,«5 illlllllt 1:1 IIIIIIMLI I E! iMIilll I I U illlHIII I B! iiiiii m ft 1 1 I I II I I I I: ! Er*..*lllllllll t t-.itx m 1 1 i t fi : i it imiiii! ii 1 1 ; I llllillll! I 1 1 1 u 1 1 1 i ; | a Photo: Author. Fig. 84.— S. Ouen, Rouen: Niches on Piebs. [ To face p. 123. MURAL DECORATION, ETC. 123 of Angers Cathedral a considerable amount of colour, red, blue, and green, remains ; and the bands, etc., on the garments of the figures, the wings of the angels, and the animals, of which there are several, are gilded solid. In the triple western porch of Sees Cathedral, Normandy, the same three colours were used ex tensively, mainly on the mouldings and on the walling behind the figures. A good deal of this decoration still exists, because the restoration has not yet reached the west front. All through the Middle Ages, and especially in the eleventh Sculpture and twelfth centuries, the sculptured pictures were the books of the people. They read in stone the stories the monks, and a few others alone, could read on parchments. The wicked descending to heU, the righteous being carried up to heaven in chariots simUar in form to those which they saw every day in the street, were pictures all could understand. The angels on the one side, the devUs on the other, and the Saviour enthroned between, as over the doorway of Autun Cathedral, expressed in concrete form their beUef. The Lamb suspended by a cord, the Figure on the Cross, conveyed far more to them than the words of the preacher ; and the monks of the eleventh and twelfth centuries were wise in their generation in choosing such subjects. Sometimes the carvings iUustrate a biblical incident, as over the north-west doorway of Eouen Cathedral, where the daughter of Herodias is shown dancing — or rather tumbUng, as an old Bible expresses it — before Herod. Elsewhere, the skill of the sculptors is shown in repre sentations of domestic animals, oxen and sheep, as at Laon Cathedral, which made a homely appeal to an agricultural com munity, or in fantastic forms, as at Paris, and Notre Dame, Dijon. Early in the thirteenth century symbolic sculpture and bibUcal pictures became rarer, and the finest work was devoted to the glorification of apostles, saints, and kings. Apart from the intrinsic beauty of mediaeval sculpture, two Mediaval reasons account for its especial effectiveness. The first is that fr^ea*6 it was executed by men thoroughly in harmony with the work surrounding it, and in close touch with those who carried that out, and secondly, that the sculpture in most cases is framed in. OccasionaUy, in fourteenth-century work, a figure stands on a corbel against a waU outside, or in front of a pier inside, as originally at S. Ouen, Eouen, without any, or with but slight, surroundings; but figures so placed are exceptions. As a rule they stand in niches with canopies or arches over them, or else 124 A HISTORY OF ARCHITECTURAL DEVELOPMENT. are carved on the tympana of entrance doorways, in which position they are still framed by an arch. In Paris, Chartres, Amiens, and other great French cathedrals, they occur in bands of arcading along the west fronts. In jambs of doorways they often partially take the place of shafts, and in the orders of the arches over they, with their canopies, form concentric bands of great richness. In the fronts and sides of the buttresses on the west front of Wells Cathedral they are in profusion, standing in niches, or seated or crouching inside quatrefoils or trefoils. Inside, save in S. Ouen, Eouen, and in some later churches, they rarely touch the structure. They are reserved mainly for reredoses, as in Winchester Cathedral, and Christ Church, Hampshire; for screens round choirs, as at Chartres Cathedral ; for tombs, pulpits, fonts, and other fittings. There are exceptions. On the inside face of the west wall of Eeims Cathedral is some fine sculpture and still finer carving in panels. The angels in the spandrils over the triforium arches in the eastern arm of Lincoln Cathedral have earned it its title of the Angel Choir. Even finer than these, although much smaller, are the figures over the seats in the Lady Chapel of Ely Cathedral. In pose and in drapery treatment they are as perfect as Greek sculpture. England has Uttle sculpture that it can pit agamst the sculpture of France, with the exception of that at WeUs. The west front of that cathedral is unusually wide, owing to the positions of the towers (see p. 305), and the whole front forms an immense screen, specially designed to contain as many figures as possible, and to show them off to the best possible advantage. No French cathedral west front can boast so fine an array. On the other hand, the greater size and magnificence of French doorways and porches gave the sculptors abroad a far finer opportunity than their brethren as a rule enjoyed here, which they turned to excellent account. Schools of Viollet-le-Duc says there were five schools of sculpture in sculpture. J r France in Eomanesque times — those of Toulouse, Limoges, Provence, the Ehenish, and Clunisian. Of these by far the most fascinating is the last, or Burgundian school, of which such magnificent examples exist under the porch of Autun Cathedral and over the doorway leading from the narthex into the nave of Vezelay Abbey Church. s One of the chief peculiarities in these two examples is the smallness of the heads of the figures. In all other Eomanesque work the heads are unnaturally large, and like raPKs Fig. 85. — Figubes feom Westeen Dooeway, Chaetbes Cathedeal. [To face p. 124. MURAL DECORATION, ETC. 125 proportions extended, to some extent, all through Gothic times. The folds of drapery on Clunisian figures bear a close resemblance to archaic Greek sculpture of the seventh and sixth centuries B.C. ; and the resemblance is interesting as showing how early workers at different periods expressed themselves in similar fashion, although many centuries intervened. Viollet-le-Duc's theory is that the special characteristics of this school of sculpture are taken from Byzantine Greek paintings, which, he claims, the work greatly resembles, especiaUy in the treatment of drapery. On the other hand, it is not impossible that craftsmen from the East were employed on it. To what extent Byzantine Greeks penetrated into France and Germany in the tenth and eleventh centuries is a point that has never been decided, perhaps never will be ; but it is one that weU deserves thorough investigation. The Abbots of Cluny certainly had close relations with the East, and not only sent missions to Antioch and other parts of Syria, but also im ported works of art from there — sUk hangings, carved ivories, paintings, etc. It is quite possible that the monks persuaded workmen to return with them. This Clunisian School spread as far as Chartres and Le Mans. The figures flanking the west doorway of the former and the south porch of the latter are almost identical in treatment with those at Vermenton Church, Burgundy (c. 1130). The figures themselves stand on high bases and form the shafts at the sides, their heads projecting forward, whilst behind each head rises a short portion of a cyUndrical shaft, which is crowned by a capital. This was the customary design in the first half of the twelfth century. By the end of the century much of the archaic feeling noticeable in all early sculpture had disappeared, and the faces and forms had become more reaUstic and the Unes of drapery more flowing. The sculpture on the north and south porches of Chartres is representative of the work that foUowed. That on the south is earUer and finer than that on the north, and its restraint and architectural suitabiUty are greater. Both porches are mines of wealth; but whilst giving them due praise it is possible to have a sneaking preference for the earlier and more conventional rendering found in the figures of the west doorway of the same cathedral. The work of the pioneers of a movement is often more fascinating than that of those who have carried that movement to what one is pleased to call perfection. It is more individuaUstic. It may have more faults, be more open to criticism, but being less correct it is often less cold. 126 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Roman esque The source of most carving executed before the middle of the twelfth century is found in Classic or in Byzantine art. Mention has already been made of twelfth-century Corinthian capitals, which differ but slightly from second-century ones, and of corbel capitals of the former date, which recall those of the sixth century. The farther one proceeds away from the centres of Classic and Byzantine art the less skilful and refined is the carving, and the greater the departure from old traditions. In Southern Italy and Sicily the carvers were undoubtedly Byzantine Greeks, and their work is distinguished by a delicacy which is never found more to the north. The Byzantine carvers of the sixth century confined themselves mainly to representations of foUage or to patterns CAPITALS 5 . EVSEBE , AVXERRE . 5 . V1TALE , R AVEM A. Fig. 86. formed of interlacing strands resembling open basketwork. Their descendants of the eleventh and twelfth centuries introduced animals more freely, especiaUy birds. The carving of the capitals in the cloisters of Monreale Cathedral, Sicily, is especially fine ; and birds, half hidden by leaves, pecking at bunches of grapes are frequent. In Southern France and Burgundy the feeling is some times the same, but the workmanship is rougher. The carving may have been executed by Byzantine Greeks, and its inferiority be due to the fact that the workmen who found their way to the west of the Mediterranean were less skUled than their brethren who found employment nearer home. On the other hand, it may well have been carried out by local men who tried to represent detail MURAL DECORATION, ETC. 127 with which they were only slightly acquainted. The mediaeval craftsman was a great traveller ; and many churches show evidence that at least two, and possibly more, schools of carvers were em ployed. The weU-known capitals in the cloisters of the Abbey of Moissac are probably all native workmanship, as, although they vary considerably in design, the executive skill shown is the same throughout. Some betray a striving after Byzantine ideals, but this, in many cases, is combined with Northern mythical repre sentations. In Burgundy there is great variety. In the church of Saulieu (c. 1119), a town which at one time was of considerable importance, the numerous capitals are marked by much free handling of the chisel and by considerable originality in design. Hardly Byzantine in feeling, they are decidedly unlike the usual Eomanesque carvings. In the Church of S. Eusebe, Auxerre, some of the capitals are quite ordinary, whilst others are decidedly the reverse. The capital illustrated resembles closely one in S. Ambrogio, Milan, and others in S. Mark's, Venice, S. Vitale, Eavenna, etc., and has the drilled holes and sharp acanthus leaves typical of Byzantine workmanship. In Germany, Nor mandy, and England the carving is always robust and strong, often coarse, but otherwise it differs little in design from carved work elsewhere. Eomanesque carved enrichments on arches, ribs, string-courses, Enrich- etc, are primaeval. The most common of aU, the chevron or zigzag, men s" was the favourite with the Egyptians, and the others also filtered through from the East. In the scalloped capital so common in EngUsh and German work of the early part of the twelfth century, the scaUops, or cones, with their dividing tongues, are direct descendants of the eggs and darts of Greek and Eoman times, which, in their turn, were derived from the lotus bud of Egypt. In one respect the Eomanesque carvers departed some what from tradition. They introduced symboUsm to a far greater extent than their predecessors, either Eastern or Western. It is possible that a symbolic meaning has in modern times been attributed in cases when none was intended by the craftsman — as a critic will often discover hidden meanings in a painting never dreamt of by the artist — but many carvings undoubtedly fulfilled a double object ; they added beauty and they told a story. With the change from the semicircular arch to the pointed, Carvings, new schools of carvers arose, chiefly laymen, not monks, who 115°-1250' discarded, to a great extent, Classic tradition and went direct to 128 A HISTORY OF ARCHITECTURAL DEVELOPMENT. nature for inspiration. This they rendered with considerable realism, never overstepping, however, the limitations which the material imposed. In some French carvings executed between 1150 and 1250 there is a certain dependence on old tradition, especiaUy in capitals. Those of Notre Dame, Paris, for instance, although full of freshness and immense vigour and thoroughly original in treatment, are undoubtedly descendants of Corinthian capitals. Almost down to the end of the fifteenth century capitals in France show traces of the influence of this Order — a compliment to the Eomans and a credit to French mediaeval carvers — which only disappeared when capitals became diminutive or were omitted altogether. Otherwise, the Frenchman, like the English man, went to nature. Whilst the source was the same, the results, however, are very different. Nobody who knows the work in both countries wiU ever mistake an EngUsh thirteenth-century carving for one from the Isle de France. French carvings are bigger in scale than English ones — even aUowing for the greater size of French churches — and the treatment, as a rule, is simpler and broader, probably because the stone generally used was hard and coarse. Most EngUsh carvings are smaU and deUcate by comparison, because they are executed in soft, easUy worked stone of fine grain. They are like Normandy ones, for which the same, or simUar, stone was used. In EngUsh capitals from 1150 to 1250, stalks spring from the necking of each, and, shooting up the bell, end in bunches of crisp, compact, stiff-leaved foliage. The same crispness and "tightness" and the predominance of trefoil leaves are equaUy noticeable in the bunches at the ends of cusps, in crockets and finials, and in the carvings of spandrils. 1250-1500. In the middle of the thirteenth century carving lost the stiffness which characterizes the earlier work in both England and France, and became as exact a rendering of nature as was obtainable in stone. In some cases realism was carried too far, and the work, although beautiful, is lacking in strength. The desire to be realistic imparts a curious crumpled character to much of the foUage, notably to crockets, as in Prior Crauden's Chapel, Ely. About the last quarter of the thirteenth century the carving in capitals ceased to spring from the necking and cluster under the abacus, and twined round the beU in a compact mass, the stalks being smaller and much less pronounced, and either hidden entirely or else tucked away under the foliage so that very MURAL DECORATION, ETC. 129 little of them showed. The effect was naturalistic, but it was neither so decorative nor so forcible as the earlier manner. The best carvings in England of this period are in Southwell Minster, Exeter Cathedral (in the corbels carrying the vaulting shafts), and in S. Mary's, Beverley, Yorks. After the Black Death considerable deterioration is apparent. The carving is cramped and knotty. It has neither the vigour and pleasant convention of the thirteenth century, nor the realistic beauty of the first half of the following one. Its angular stiffness — very different from the rounded stiffness of early Gothic — shows that the carvers were losing their skill, were neglecting the study of nature, and were settUng down into a conventional groove. The work improved again before the end came, and late fifteenth - century carvings often show considerable vigour. The most characteristic ornament in English work between Dogtooth. 1150 and 1250 is termed the " Dogtooth." It is fairly common also in Normandy, even as far south as Le Mans, where it occurs in the window over the doorway of the south tower of the cathedral, and it is met with occasionally in other French cathedrals, notably Laon. The usual statement is that this feature is derived from the "nail-head" of Eomanesque times, elongated and elaborated; but it is much more Ukely to be a modified version of the chevron, no longer continuous but spUt up into separate ornaments. During the above-named period the dogtooth was almost the only ornament the buUders permitted themselves. They placed it between shafts, half buried it in hollows, or aUowed it to stand out boldly in front of surrounding work. They used it in arches, piers, labels, string-courses, jambs of windows and doorways, ribs of vaulting — in fact, everywhere. Towards the end of the thirteenth century the dogtooth Ball disappeared and was replaced by the so-called "Ball flower," g°0wer' which consists of a calyx of three or four leaves, slightly open so as to show in the centre an unopened flower. Absolute monotony is avoided by varying the " hang " of the flowers, but at best the ornament is but a hard and unsympathetic rendering of nature, even after due allowances are made for the nature of the material it is carved in. It is most satisfactory when it alternates with a square four-leaved open flower, especially when, as often happens, the two are connected by a stalk so that a running VOL. 11. K 13° A HISTORY OF ARCHITECTURAL DEVELOPMENT. CAPITAL, SOVTHWELL, MINISTER, (c.1300). C DOGTOOTH. QIEVKD/1. CROCKET5 A/ID CAPITAL, (clls0)- (ciOSO). LI/1COLH CATHEDRAL (C.I200). C boss.s.mary, bvry s.ednwids. a tvdor rose Pig. 87. MURAL DECORATION, ETC. 131 pattern is formed. In the aisle windows of Gloucester Cathedral and Leominster Church the jambs, heads, mullions, and tracery are all peppered thick with ball flowers, the number of which the curious have counted and recorded.1 In the great circular west window of Chartres Cathedral the ornament carved round its openings (and in the labels over it and the side windows of the front) is a species of ball flower earlier in date and more beautiful than the EngUsh variety. The leaves are folded back, not brought forward, and the whole of each flower (not merely the centre) resembles a ball. In the Church of S. Pierre de la Couture, Le Mans (c. 1150), in the labels over the arches of the windows, are flowers which are also earlier than the English ones. In some of these the flower is quite open, in others only half open, and the effect produced by the variations is exceedingly good. Before the wars between the Houses of York and Lancaster Tudor the rose usually carved in English work was the simple five- ros6' et0, petaUed briar or wUd rose. When the Houses were united, to symbolize the union, roses were carved with ten and sometimes fifteen petals. This form, in consequence of the extensive use of it by Henry VII. and Henry VIII., has been termed the Tudor rose, and, with the portculUs, forms the most distinctive badge in the chapel of the first-named at Westminster. Few remains of twelfth-century glass exist; partly because Glass. it was not untU a century later that stained glass was employed extensively, and partly because early glass was rarely fitted into stonework, but into wooden frames, which were easUy removable. What Uttle there is shows that the designs were simUar to the designs for carvings, and that scroU work founded on the classic acanthus was the prevailing motive. In the thirteenth century figures were first introduced to any extent. As a rule, they are small, and are grouped together to represent scenes, sometimes biblical, sometimes pastoral. Each Ught consists of many of these pictures placed inside geometrical forms, such as circles, quatrefoUs, squares, etc. The effect is generally very dark, as there is little white glass, and often not much yellow, the prevailing colours being red, blue, green and purple, which are worked together in wondrous harmony and rival in colour the richest of Persian rugs. Amongst English examples of this type is the famous " Five Sisters " 1 Sharpe says that each window of Leominster Church has 820 ball flowers of different sizes. 132 A HISTORY OF ARCHITECTURAL DEVELOPMENT. window in the north transept of York Cathedral. In France similar contemporary windows are in the aisles of Chartres Cathedral, the choir aisles of Le Mans Cathedral, etc. In the clerestory windows of both, however, the design is different. The figures are large, and are no longer grouped and confined within geometric forms. In Chartres Cathedral each Ught has often only one large figure, although sometimes there are more. No attempt has been made to obtain uniformity throughout. The figures by no means all start from the same level, and some are much larger than others. In fifteenth-century glass, as a rule, the figures aU toe the line, are about equal in height, and the canopies above them alike in size, although they differ somewhat in detaU. There is no such exactness about the glass at Chartres. The canopies there, when they occur at all, are small and in absolute subjection to the figures. To these irregularities is largely due the unrivaUed reputation that the Chartres glass enjoys. If it has a fault it is that it is too dark, owing to blue being the predominating colour. In the clerestory windows of Le Mans Cathedral there is very little blue. Some of the figures, of which there are generally two to each light, one over another, are mostly yellow, whilst others are green or red. In fourteenth-century glass single figures, one to each light, are the rule, but they do not generaUy completely fiU the space. Medium in size, each stands out a mass of rich colour in a setting of almost clear glass, on the small lozenge-shaped panes of which a diaper pattern, usually in yellow, is pamted. Bound the outside of each light is a border of delicate patterns, in colour matching the central figure, which is separated from the muUions by an inch, sometimes less, of absolutely clear glass. This edging of white glass is of great value. It separates the glass from the stone and forms the finest frame imaginable. The fifteenth-century glass stainers followed fairly closely the workers of the previous one, but they frequently omitted the border, made the figures larger, and gave greater importance to the canopies over the heads of saints, apostles, and kings. These canopies are invariably copies on the flat of the carved canopies on tombs, tabernacles, etc., and although they are of some use in lengthening the design and fiUing the foliated heads of lights, they are often unnecessarily large and elaborate, and somewhat overpower the figures below. The artists of Chartres did without them entirely, or introduced them sparingly. That so many of the glass designers of the MURAL DECORATION, ETC. 133 fifteenth century could not, in the surroundings of their figures, tear themselves away from the designs of the mason is evidence of the decline in art which is apparent at that time in all the other crafts.1 Stained glass possessed such virility in the thirteenth and fourteenth centuries, and exercised so powerful an influence on architectural development, that as its use increased (as it undoubtedly did towards the end of Gothic art), one might have expected more independence on the part of the artists who pro duced it, and a special treatment for the accessories of their figures, different from that appertaining to carved stone and better suited to the material in which they worked. That in many cases they did not succeed in originating this constitutes the chief blot, perhaps the only one, on a great deal of otherwise beautiful glass of the late Middle Ages. No fault can be found with the figures themselves ; these compare favourably with the earlier ones, and evince also considerable advance in technical skill. Neither does aU fifteenth-century glass show extravagance in canopy work, nor can all be included under the above generalization. In some instances the design is arranged to fill a whole window with many figures, which are grouped in such a manner that, although separated from one another by mullions, they form one picture. In other examples the figures are smaller and there are many to each Ught, the scale and arrangement approaching nearer to thirteenth-century glass without the enclosing geometrical forms. Of this character is the east window of York Cathedral, which was executed by John Thornton, a glazier of Coventry, in 1405. York has also fine fourteenth-century windows in the nave, and the east window of Gloucester Cathedral is of that date. The finest of late windows in England are possibly those of Fairford Church, Gloucestershire (c. 1500), and King's College, Cambridge (c. 1530). Many parish churches in England contain fragments of old glass, and in some cases windows are completely filled with it ; but there is not the wealth in England that exists in France. In addition to Chartres, the Cathedrals of Bourges and Beims are fuU of fine glass, mostly early fourteenth century, and examples of the work of the century foUowing are exceedingly numerous. Modern glass is often held to be inferior to old in colour, but Colour in it is a question whether the best of the former cannot hold its g ass' own in this respect with the examples of the Middle Ages. Time 1 It is seen in monumental slabs and brasses. Fifteenth-century ones are far more masonic in treatment than any of the previous centuries. 134 A HISTORY OF ARCHITECTURAL DEVELOPMENT. and dirt are valuable allies. Although glass is a wonderfully tough material it is certain that it is acted upon to a small extent by the atmosphere. The acids in the air affect its surface, especially in south-west aspects, and roughen it, at times make small holes in it, forming something in or on which dirt can lodge. Much of the deepness of the blues and the richness of the reds in old glass is due to dirt, especiaUy in clerestory windows of large cathedrals wliich cannot easily be periodically cleaned. The term The term " stained glass " is, to some extent, a misnomer. In glass!"6 *ne nnest glass there is very Uttle staining or painting except on faces, hands, and flesh generally. The patterns on the rich robes of kings and bishops, and in the borders of fourteenth-century windows, are not in most cases pamted. The glass used for these is mainly flashed glass, i.e. one side of a piece of glass of one colour is covered by a thin film of glass of either another colour or a darker shade of the same, and the patterns are formed by eating away with acid the flashed coat, so that in places the original colour shows. Sometimes it is the ground that is eaten away, sometimes the pattern itself. This, there can be Uttle doubt, is the right and legitimate treatment for glass; and the more modern designers employ it, and the less they rely on applied pigments, the brighter is the future for this most beautiful and permanent form of colour decoration. CHAPTER IX. PLANS, ETC., OF EARLY CHURCHES IN FRANCE AND GERMANY. The development of the Eastern, or Byzantine, church plan from intro- the sixth century to the twelfth, and the gradual growth of the duotory- Western, or BasiUcan, plan, until about the beginning of the eleventh century, as exempUfied in certain Eastern countries and in Italy, have already been traced in the previous volume. No mention was made in that of early examples elsewhere in Europe, partly because these are neither numerous nor important, and partly so that the thread of continuity could be better preserved when deaUng with the work in other countries. The investi gations of archaeologists have somewhat disposed of the old theory that during the last few years of the tenth century church building was at a complete standstiU (owing to the vulgar super stition that the year 1000 a.d. was to see the end of the world) ; but it is still a fact that the number of churches commenced or rebuilt in the eleventh century was enormously in advance of the number in the previous one. The greatest activity was in the second half. Up to about 1050 there is little change to chronicle in either plan or general ordinance. Afterwards, advance was general and rapid throughout Europe, especially in the Western and Northern countries, and the plans show a uniform develop ment, and, in addition, special characteristics, which separate unmistakably the later churches from the earUer ones. The only countries which wiU be referred to in this chapter dealing with plans of early churches, i.e. those buUt between 850 and 1050, are France and Germany. Some contemporary churches in Italy have already been mentioned in Vol. I., and those of England wUl be dealt with in a subsequent chapter. Spain has few such examples, as, with the exception of the northern frontier, it was in the hands of the Moors until the eleventh century was far advanced; and in other countries, 136 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Basilicanplan. Basse CEuvre,Beauvais. Norway, Sweden, Bussia, etc., early work is almost non-existent, as Christianity, although it had taken root, had not yet borne architectural fruit. Most of the early French and German churches undoubtedly followed the typical Italian basilican plan. The majority have disappeared. They have either been razed to the ground entirely, or else have been so altered and added to, like S. Eemi, Beims, that their original designs are difficult to make out. But a suffi cient number remain to show that the most usual plan was a nave with single aisles, with, as a rule, one apse only at the east end ; a proof that the clergy and people, still simple in their devotions, clung to the primitive idea of only one altar. Some of the churches, however, are triapsal, and in a few examples the apses are divided from the nave and aisles by the unbroken transept, stretching from north to south, which occurs in many of the large early basilican churches in Eome, such as S. Paolo fuori le Mura, S. Maria in Trastevere (Vol. I., Fig. Ill), etc. According to Viollet-le-Duc, the original church of S. Denis, near Paris (c. fourth century) had an eastern transept of this type. The church known as the " Basse QSuvre," Beauvais (c. 997), three bays of which alone remain, gives a fair idea of the simpUcity of these early examples. In plan it follows ItaUan traditions, but such decorative features as it possesses belong more to the East, as is often the case in early eleventh-century churches in France. In this church (and in most other contemporary examples) the arcades inside are low, the piers being plain rectangular ones, and the arches, of course, semicircular ; the windows, one to each bay in the aisles and above the arcades, are large, because they were probably originally filled with pierced slabs, as in Byzantine churches ; and the space between the arcade and clerestory, wliich is considerable, left plain, although it was possibly originally painted. In these early churches this space is very rarely arcaded, as in the majority of churches of later date ; 1 and neither funds nor workmen were forthcoming to treat the wall surfaces with marble, as in S. Miniato, Florence, or with mosaics, as was customary in many parts of Italy. In the Basse 03uvre, more design is visible outside than inside. The arches of the side windows are coursed, two red tiles alternating with wide stone voussoirs; the west window has an archivolt of great delicacy, 1 In the monastio church of Montier-en-der (c. 998) there are galleries over the aisles, and consequently triforia to the nave, but this church is an exception. EARLY CHURCHES IN FRANCE AND GERMANY. 137 the pattern being a most unusual one, and the walls are faced with small squared stones, about 6 inches square, laid with great care. The plan of the church is Basilican ; but the alternating courses of the side-window arches, the archivolt of the west window, and the outside masonry of the walls all suggest Byzantine influence. One point worthy of mention in connection with early basilican Alter- churches in Germany is the combination of piers and columns, ^f^a which has already been referred to in the first volume, as existing columns. in such churches as S. Maria in Cosmedin, and S. Clemente, Borne (Vol. I., Figs. 112 and 114). In Minister Cathedral, and S. Michael's, Hildesheim (c. 1020), piers take the place of columns every third bay ; and at Driibeck (c. 1000) and elsewhere columns and piers alternate. All these churches are timber-roofed, were buUt, in fact, before the reintroduction of vaulting ; and the alter nation of large and smaU supports had consequently nothing what soever to do with vaulting requirements. A detail of some interest in connection with these churches, and others of contemporary or sUghtly later date in Germany, is that the columns are generally monoUthic, and, foUowing Classic precedent, diminish in diameter from base to cap, and are entasised. The plans and general ordinance of the churches mentioned Plans due above are much the same as contemporary and earlier churches in jj^S118'11 Italy, and show no signs of advance. But in addition to these zantine there are others in both France and Germany which are of greater iufluenoe- interest, because their plans show unmistakable evidence of the two great influences in force in the early centuries of Christianity, the ancient Boman and the somewhat later but stUl more inspiring Byzantine ; influences which are absent in most of the churches built after the eleventh century. These examples are, in fact, connecting Unks between the East and the West ; between the civilizations which had lost or were losing their force, and the new reUgious and artistic movements which, at the beginning of that century, were in their infancy. They are also especiaUy worthy of study in that, in many cases, they contain the germs of later developments, and thus help to elucidate some points the origin of which would otherwise remain obscure. France and Germany were in a backward state, except during Roman the time of Charlemagne, throughout the early centuries of the influenoe' Middle Ages, when art flourished in Italy and the East; and it is only natural, therefore, that many of their early churches 138 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Byzantineinfluence. should reflect strongly the characteristic traits of the stUl earlier perfected work. At the beginning of the eleventh century the direct influence of the architecture of ancient Borne was by no means dead, although centuries had passed since Eoman ideals dominated the civilized world. It was strongest, perhaps, in the part of France bordering on the Mediterranean, because the early colonists had introduced traditions of buUding there wliich had not been superseded by the artistic movements which effected changes in other lands. The Church of S. Pierre, Beddes (near Bedarieux), for instance, of the tenth century, is modelled on the baths of Diana, Nimes, and like that buUding is covered by a barrel vault, and has attached columns lining the walls of the interior. In Germany, Eoman influence crops up occasionally, and nowhere is this more marked than at Mettlach, where there is a tenth-century octagonal church. This is planned without aisles, but in the thickness of the wall are semicircular niches which recaU the Pantheon, and other similarly planned Eoman buildings. The oldest building in France showing Eastern influence is probably the Church or Baptistery of S. Jean, Poitiers, built SECTIO/1. AIX- LA- CM APELLE, (original desicyi). SCALE OF IO O IO 20 30 40 SO 60 70 BO 90 IOO FEET ' -1 1 1 1 I 1 I I I I I Fig. originally in the fourth century, but rebuilt and added to sub sequently. It has three apses, two at the sides and one at. the end. Byzantine feeUng is still more marked in the church built by Charlemagne at Aix-la-Chapelle (Aachen) late in the eighth century or early in the ninth. Later additions mask somewhat the original plan. This consists of an octagon surrounded by an aisle enclosed within a multangular outer wall. On one EARLY CHURCHES IN FRANCE AND GERMANY. 139 side is a porch, flanked by turret staircases, opposite which was probably a small chancel. The likeness between the plans of this church and S. Vitale, Eavenna (Vol. I., Fig. 153), is unmistakable, although the later church lacks the columned niches between the piers surrounding the central space which give the earlier one its distinctive beauty. At Ottmarsheim, Germany, is a church simUar in size and plan, although two centuries later in date than Charlemagne's church.1 The outside waU is octagonal, not multangular, but otherwise the differences are slight. Both churches have gaUeries arcaded in front. More interesting than any of the above is the Church of Germigny-les-Pre's (Loiret), said to have 5ECT1C71 LOOKl/NG EAST Fig. 89. been built at the beginning of the ninth century. In plan it consists of a square, which, in orthodox Byzantine fashion, encloses a Greek cross, the cruciform portion being divided from the corners by four piers. Apses project from the sides. Above the inter section rises a moderately lofty square tower, under which, in recent restorations, has been placed a dome, although it is doubtful if one was there originally.2 The arms of the cross are covered by barrel vaults, and the apses by semi-domes, the latter being 1 Brescia Cathedral (Vol. I., p. 254) has a similar plan. 1 The apses surrounding the church are shown as published by Cesar Daly in 1849. There was possibly also a western apse, and the east end may have been triapsal. Churches with cir cular east ends and oblong naves. S CROIX , /nomnAJovR. 140 A HISTORY OF ARCHITECTURAL DEVELOPMENT. considerably lower than the main arms. In plan and general ordinance the church, save for its side apses, is Uke S. Theo dore, Constantinople (Vol. I., Fig. 161), and numerous smaU churches in Greece. The building, however, to which it bears vthe closest resemblance is the old Church of S. Satiro, MUan (Vol. I., Fig. 168), the parent of wliich was probably the Church of S. Lorenzo in the same city (Vol. I., Fig. 154). Of the same type are S. Croix, Mont- majour (c. 1016), and S.Martin de Londres, both in Southern France. The plan of the former is a quatrefoU, with an added porch, whilst that of the latter is simUar, but instead of the western apse it has a short oblong nave. This type of plan was dropped through the succeeding centuries, but was revived by the architects of the Eenaissance, who realized its great possibilities. Next in interest comes a group of churches, also in France, with circular east ends and oblong naves, the best known being S. Fig. 90. CHVRCH AT CHARROVX 5CALE OF FEET. Pig. 91. Benigne, Dijon, Charroux, near the Loire, and Neuvy S. Sepulchre. As regards the last, although the circular portion dates from c. 1045, the nave is a later addition. If one could be quite certain that the plan of Charroux, as generally presented, is the plan of EARLY CHURCHES IN FRANCE AND GERMANY. 141 the original church, it would be one of the most interesting in existence. The central space at the east end is small, merely the width of the nave, but it is surrounded by three aisles, so that the diameter of this portion far exceeds the width of the western nave and aisles. There is more certainty regarding S. Benigne (c. 1001), although the entire western arm and the greater part of the east end have been rebuilt. Still, the crypt of the latter remains. This consists of a central circle sur rounded by a couple of concentric aisles, with' portions of an oblong chapel at the west end. There can be little doubt CHVRCH OF THE HOLY SEPVLCHRE. JERVSALEM. SCALE OF 10 o 150 FEET. Fig. 92. that the upper church was buUt on the same lines as the crypt below, and that is therefore interesting not only for itself, but as indicating the plan of the church proper. To what extent, if any, these churches owe their plan to Church of the Church of the Holy Sepulchre at Jerusalem is a point Ig6^017 that may be argued. UntU 1076, Christians were freely per- chre, mitted to enter Jerusalem on payment of a small legal toll. In l^^ that year the city came into the hands of the Seljukian Turks (a somewhat different race from that which previously possessed it), and it was the ill-treatment of pilgrims which followed that excited the wroth of Peter the Hermit and occasioned the first crusade. 142 A HISTORY OF ARCHITECTURAL DEVELOPMENT. M. Anthyme S. Paul states x that between the years 990 and 1000 pilgrimages to Palestine were general, and that one Foulgues Nerra, when he died at Metz in 1040, had visited Jerusalem four times. The church as it then existed was therefore weU known. But what was its plan ? Constantine's church, the first to be built there, was basilican. He built it to the east of, and detached from, a great apse, about 120 feet in diameter, which he placed round the spot where he believed the tomb to be. His work was destroyed ; but about 1040 his namesake, Constantine Monarchus, Emperor of the Eastern empire, built an ordinary circular church round the tomb, except that apparently only the aisle was covered, the central part being open to the sky. When the Crusaders reached Jerusalem in 1099, they at once commenced the rebuilding of the church; but, instead of following the Eoman Emperor's plan of a detached church, they built their chancel on to the circular church. The result was a western circular nave — the building of Constantine Monarchus — and an oblong eastern chancel. There can be little doubt that the church as rebuUt supplied the plan for many "Templar" churches throughout Europe, such as the one at Laon, in Northern France — the nave of which is octagonal — and our own four examples at London, Northampton, Cambridge, and Little Maplestead in Essex, the last being a century or so later than the others.2 But the churches of Dijon, Charroux, etc. — with the exception of an eastern chapel in the case of S. Benigne — show a reverse arrangement ; their east ends are circular, their naves oblong. It follows, therefore, that, being buUt, some a century, others half a century, before the first crusade began, they cannot have been influenced by the Holy Sepulchre Church as rebuilt by the Crusaders. The builders of Charroux, Dijon, etc., devised a plan which gave them a suitable nave, and a choir so arranged that the faithful could gather round the altar. It was this desire for space at the east end wliich occasioned the most important of the developments which will be described in the next chapter. 1 In his " Histoire Monumentale de la Prance.'' 2 Although the Templars may have taken the plan from Jerusalem, it had been employed before, and probably had a Byzantine origin, as mentioned in Vol. I., pp. 255-6. CHAPTER X. THE DEVELOPMENT OF CHURCH PLANNING. The chief alterations that took place in the planning of churches Intro- during the eleventh and following centuries occurred in the "uotlon- eastern arm. The nave remained much the same all through the great period of church-building. At first, many churches continued to have merely a single apse or three apses side by side, either as eastern terminations to the nave and aisles, or else opening out of a transept as before described. In others, how ever, a more marked tendency towards greater length and increased importance of the east end is manifest. Nowhere is this more noticeable than in Burgundy, Normandy, and England, the three countries in which the greatest building activity prevailed during the last decades of the eleventh century. The great church of Cluny, in Burgundy, commenced 1089, marks the culmination of the early developments. In its plan are found not only many of the characteristic marks of later French cathedrals, the chevet, double aisles to both nave and choir, etc., but, in addition, the long eastern arm and the smaller transepts east of the great transepts, which afterwards became characteristic of many English ones. Full development was reached in both France and England about the middle of the thirteenth century ; the few modifications made subsequently possess little importance. The changes in plan followed on much the same lines in all countries until about the middle of the twelfth century, irrespective of the fact that some large churches are without aisles, others are aisled throughout, whUst a few have aisles to the choir and none to the nave. After that time, considerable differences appear in the different centres, and are conclusive proof that each of the great building countries, England, France, Italy, and Germany,1 foUowed its own line and worked out its own solution of the problems presented to it. 1 Germany followed France in the thirteenth century, but afterwards was more independent. 144 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Causes for changes. Change in position of altars. The advances in planning and the modifications introduced are traceable to various causes. The most important of these are : (1) the need for additional altars, occasioned chiefly by the large increase in the number of clergy and monks ; (2) the insistence on more marked separation between clergy and laity; (3) the greater interest taken by the laity in matters of religion, and their greater influence; (4) Pilgrimages; (5) Eelic worship; (6) Alterations in rules relating to burial and baptism. The increase in size of the eastern arm, although following naturally, to some extent, on the increase in size of the monasteries and corresponding increase in the number of the monks, was, in the main, due to alterations made in the placing of the altars. In the case of a large monastic church a considerable number of PLAAi OF THE ABBEY CriVRCH OF 5. GALL. w4^E S I? ~r X a a — s b s — a a pa — "3- «3* |"<§>+ ' A ALTAR u=^ fr T. TOWER iii-ly ?4 a^ ^fTU^- Fig. 93. altars was essential. The one altar, or the three of the earlier, smaller church, was insufficient. UntU near the end of the eleventh century, it had been the practice when more were required to place them in the nave. In the Abbey Church of S. Albans, now the Cathedral, each pier of the nave had an altar on its western face. StiU stronger proof that many altars in the nave were customary is afforded by the existing plan on vellum of the church of the Monastery of S. GaU, Switzerland, built in the early days of the ninth century by, it is stated, the architect who designed Charlemagne's church at Aix-la-Chapelle. In this plan are shown, in addition to altars at the east and west ends dedicated to S. Peter and S. Paul respectively, altars down the aisles and the centre of the nave as well. These positions THE DEVELOPMENT OF CHURCH PLANNING. 145 were satisfactory so long as the nave, in addition to the choir, was given up mainly to the monks; but the arrangement presented many difficulties if the church of the monastery was also the church of the people, in which their representative, the bishop, had his chair, as at Canterbury, Durham, Norwich, Winchester, etc., or if the church were a cathedral pure and simple, unattached to a monastery, and served by secular canons. It also required modification in the ease of Cistercian monasteries, in which accommodation had to be provided for the numerous lay brethren who worked in the fields belonging to each monastery and were little more than agricultural labourers.1 But all mon asteries derived a large proportion of their income from the offerings of the faithful, this being especially the case when the reUcs preserved in the churches were special objects of veneration. The monks, therefore, did not desire to close their churches, and yet they desired privacy. To obtain this they removed the altars from the naves and erected them east of the crossings. The canons of the cathedrals followed suit. From this time onwards the separation of clergy and laity was complete. An altar was placed at the east end of the nave for the use of the latter, and the whole of the eastern part was screened off and reserved exclusively for the former, except on days of festival, when the entire church was thrown open to aU. The present arrangement in S. Albans Cathedral is probably not unlike the original one. The choir proper extends three bays into the nave, from which it is screened. To the east of it is the high altar separated from the chapels beyond by another lofty screen. To the west of the choir proper are now another altar, and seats for the clergy and members of the choir. In Gloucester Cathedral — a monastic church which only became a cathedral at the Eeformation — it is evident that the gaUery round the choir was set aside for the monks alone. Portions of it formed chapels; and there are still remains of five altars there — one on the south side, two on the north, and the remaining pair over the Lady Chapel. The arrangement not only ensured the monks complete privacy, but also economized space. In England, in most cases, the laity seem to have had access to the altars in the great transepts, but in France this is not 1 At Fountains Abbey, a Cistercian monastery, the nave was for the lay brethren, and their altar was placed at its east end. The aisles of the nave were screened off for the use of the monks. VOL. II. L 146 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Easternextension. so certain. When the choir was entirely east of the crossing, as in all later cathedrals, Exeter, SaUsbury, Wells, York, etc., or commenced under the crossing, as at Chichester, the transepts naturally appertained to the nave. Even when the choir extended westward of the crossing into the nave, effective separation from it and the transepts was easily obtained by means of stalls or screens. The extension westward of the choir is most marked in cathedrals to which monasteries were attached, or in abbey churches which were not cathedrals. In West minster Abbey the choir still occupies three bays of the nave, in Norwich Cathedral two bays, in Gloucester and Winchester Cathedrals one bay. Two plans of eastern extension were followed in the latter half of the eleventh century. The earlier and simpler one con- PLA/1 OP CAATERBVRY CATHEDRAL, A5 BVILT BY LAH- -FRA/1C, ArtD IA1 1174 SCALE OF FEET Pig. 94 (Willis). sisted merely in lengthening the chancel by the addition of two bays, or more, between the transepts and the eastern apses. The aisles were lengthened as well, although sometimes, as in the Abbaye-aux-dames, Caen, they are completely cut off from the chancel by walls. S. Georges, Boscherville, and the church at Cerisy-la-foret, Normandy, are typical examples of this simple elongation. In both, the aisles finish square, whilst the chancel of each has an apse. The transepts stand out beyond the aisles, and have apses or altars to the east, showing that additional altars were even then considered necessary. This was also the original plan of Canterbury Cathedral (c. 1070), and of S. Etienne, Caen (c. 1060). THE DEVELOPMENT OF CHURCH PLANNING. \\1 The chief drawback to the above plan was that circiUation a Eastern the east end was difficult. The aisles formed cul-de-sacs. This -^^^ was a very serious objection in churches where pilgrimages to shrines and relics were frequent, and where pilgrims came in vast numbers. The other plan, which was perhaps adopted in a few instances at the end of the tenth century, but did not become general for nearly one hundred years, solved the difficulty by providing an ambulatory behind the central apse. The apses to the aisles were swept away, and the aisles continued round the back of the high altar. It is impossible to say in which church this plan was first adopted, or even in which country it was first thought of.1 An eastern ambulatory, with a gallery over, occurs round the Church of S. Stefano, Verona, which Cattaneo, with some hesitation, ascribes to the tenth century. In France it seems to have been introduced early in the foUowing century; but in England not until about 1090, Gloucester Cathedral (c. 1089) being probably the first.2 As a matter of fact, an eastern ambulatory was not so much an innovation as an adaptation. The difficulty of circulation at the east end had been overcome in S. Benigne, Dijon, and in other similar churches described in the last chapter (see Fig. 91) ; and all that was really done in the churches now under consideration was to substitute an oblong with semicircular end for a complete circle. The eastern ambulatory plan was especially favoured in those churches to which large crowds were attracted, through their possessing the bodies of saints and martyrs, or other valuable reUcs, as at S. Martin, Tours, S. Denis, near Paris, S. Sernin, Toulouse, Charroux, etc., and Canterbury Cathedral, after the murder of Thomas a, Becket. The crypt of the original cathedral at Chartres, which stUl exists, built early in the eleventh century, has an ambulatory, which, it is almost certain, was continued above. The east end of Eomsey Church (c. 1120) shows an equally suitable although slightly different plan for providing necessary circulation. This is not strictly speaking an ambulatory, but the aisles are continued east of the chancel, and open into a passage behind the high altar. The ending is thus a square one, an important detail which will be referred to later. 1 In S. Lorenzo, Milan, c. 530 (Vol. I., Pig. 154), all the apses are surrounded by aisles. 2 Amongst English churches of the eleventh century the following had either apses or square terminations to the aisles : Canterbury, Chester, Durham, S. Albans ; whilst Norwich, Bury S. Edmunds, Gloucester, had eastern ambulatories. The chevet. 148 A HISTORY OF ARCHITECTURAL DEVELOPMENT. The outside waU of the ambulatory in some of the earUest examples forms an unbroken semicircle, as at Morienval (Oise) and S. Saturnin, Auvergne, but in most churches apses project in ROAASEY ABBEY CHVRCH. SCALE OF FEET Pig. 95. front of the curving wall and contain altars. These apsidal chapels at first were small, and, almost without exception, semi circular in plan. Between them, as a rule, were windows. In most early examples there are only three chapels, as at CHVRCH AT VIG/IORY 1 SCALE OFIO 0 50 1 l 1 I r 1 | 100 FEET. Fig. 96. Vignory (Haute-Marne, c. 1030), although sometimes there are five, as at Cluny, S. Sernin, Toulouse (see Fig. 157), and occa sionally four, as at N. D. du Port, Clermont-Ferrand. Thus was born the "chevet," the characteristic ending of nearly all Pig. 97. — Le Mans Cathedbal : East Exd. [To face p. 149. THE DEVELOPMENT OF CHURCH PLANNING. 149 the great churches and cathedrals of Northern France. The term is applied to any east end in which an ambulatory — sometimes consisting of one aisle only, sometimes of double aisles — is carried round the apsidal ending of a choir or presbytery, with radiating chapels projecting from it. The chevet soon lost its simple character; the chapels increased in number and in size, and, ceasing to be separated from one another, united, and formed a continuous band, sweeping round a central apse. Often they are continued down the sides as well. In Le Mans Cathedral there are as many as thirteen surrounding the choir, seven at the end, and the remainder at the sides. In Notre Dame, Paris, and Bourges Cathedral, the chevet is of the early simple kind ; at LE AAn5 CATHEDRAL «* 1 ¦ 1 I 1 l 1 1 1 '*'-#-i"-*'N -*-*- -*.-*. *--*£¦ ¦?--?---?- jjj /\ i /\ i ,/\ ! A. ! /\ : /\ $ A i A i /\ j /&#to 111a J'tM^- o 50 'ii 11111111 SCALE OF 100 150 1 1 1 1 1 I I FEET Fig. 98. Eeims (c. 1220) it is fully developed ; and at Amiens, Beauvais, and Le Mans it is carried stUl one step further. In the last-named (c. 1230) the projection of all the chapels is considerable, and they have again become separate. Between them there are windows ; probably because it was felt that, the chapels having greater projection than was customary, the windows in them would not give sufficient Ught to the ambulatory. The result is one of the most striking east ends in the world ; and all the more striking because the whole of it can easily be seen, which is by no means always the case in French churches. The eastern ambulatory plan was, in the fifteenth century, French abandoned in many churches in France of secondary rank. In ff^l 150 A HISTORY OF ARCHITECTURAL DEVELOPMENT. chevet plan. Thechevetin Eng land. Englisheast ends. the church at Pont l'eveche, and in S. Jacques, Lisieux (c. 1500), for instance, the aisles end square, and beyond them project the chancels which have semi-octagonal or semi-hexagonal ends. The chevet plan in its simple form was fairly general in English cathedrals built before the middle of the twelfth century, but after that time it disappears; the only examples of later date retaining it being the Abbey Churches of Westminster (c. 1250) and Tewkesbury. The reason for the chevet in the latter church is that it was buUt on the foundations of an earlier east end; and the reason for it at Westminster is that Henry IIL, who rebuilt the Abbey, was at heart a Frenchman, and although he could not force the workmen to build in the French fashion entirely, he could dictate the plan and general ordinance of the church. WhUst France was perfecting its chevet plan, England was busy throwing off the yoke of continental tradition. Even in NORWICH CATHEDRAL. SCALE OF Fig. 99. eleventh- and twelfth-century churches the eastern arm is con siderably longer than in most contemporary foreign examples, showing that the desire for length existed here from quite early days. Thus the bays between the crossing and the apse number four at Norwich, Durham, and Bury St. Edmunds (c. 1089-1096), the same number as in the great church at Cluny, whilst in Canterbury Cathedral, when remodelled in 1184, this number was doubled (see Fig. 94). This desire for length led to the adoption by later English builders of either one or other of two plans. There was not the THE DEVELOPMENT OF CHURCH PLANNING. \%\ same unanimity amongst them as distinguished their brethren of Northern France. In one plan, the choir, presbytery, etc., are continued the full height of the church as far as the extreme east waU, as at Ely, Lincoln, and, later, York ; in the other, the high portion of the church stops short east of the high altar, and retro-choir and chapels of considerably less height are built LI/1COL/1 ****, ¦1AA-..FV CATHEDRAL. *:.:¦*•: .¦.»::..>: $;;.-:»:^.^;;;j> w ";if ;£ >;-;*/ A*::* » »¦ » *. ..fr'.-*-* _i|C'*'fr »' « »'" r- - -it- -it— i .V, .V SCALE OF » IOO ISO 200 350 »o FEET SALISBVRY **JM» CATHEDRAL. r4-rj-r4r III! 1LM- :* -?: :*' .#• .'*' .»: 4*. _ ,^(|) .— • •* » * «'#'*.'•'-- ,'1*. -v. <«HI- -. ;»;. ;». -. ;*;. t .*: . ..?. \#» —" •'*(?,: •>.-,.«-- =.* = = -\»'.= '-.»,= -'.*.; '-.•?;¦ '*':. ¦rrtiitt ri f**3 rtlGM ROOP STOPS HERE- Fig. 100. out beyond, as at Salisbury, Exeter, Chichester, Wells, S. Albans, etc. In the latter examples, below the east window of the presbytery, which starts above the roofs of the chapels beyond, are arched openings, through which deUghtful perspective effects are obtained. At Chester there is only one opening, at Exeter there are two openings, at Salisbury and Wells three. At 152 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Com parison of Englishand Frencheast ends. S. Albans the openings are blocked by a high stone screen, which, together with the tall reredos behind the high altar, shuts off the space in which the shrine of the Saint is placed and blocks the vista completely. There will always be a difference of opinion as to which of the English methods is the more satisfactory inside, and whether either is superior to the French. In French cathedrals with double ambulatory aisles a certain amount of perspective effect is obtained, especially when the Lady Chapel has con siderable projection, as in Le Mans Cathedral. In those with single aisles the perspective is somewhat cramped, and the effect not so good. Moreover, in the majority of French churches the openings round the central apse are too numerous and too narrow, the piers too slender, and the arches disagreeably stilted. In English examples, charming though the result is at Salisbury, Exeter, etc., it is a question whether the better and more dignified course were not followed at Ely and Lincoln. The perspective may be more playful in the first-mentioned two cathedrals, but that hardly makes amends for the lack of the grandeur and the feehng of space which are so remarkable in the others. Externally, opinions are also divided. Most people wUl probably admit that outside the French east end surpasses both the English endings, but not all. To some its restlessness is displeasing. In their opinion the simple dignity of the east end of Ely has no equal in continental work. In the plan adopted at Salisbury, the low buildings east of the high altar give scale to the rest of the church, but they do not produce so pyramidal an effect as the ambulatories and chapels abroad. The curtailment in length of the high portion of the church is no drawback in English examples in which this plan is followed, because the ratio of length to height of the main structure is always amply sufficient. Such curtailment would have been fatal abroad, where the churches are so much loftier and proportionately shorter. The naves of continental churches are often as long as many of ours, but the eastern arms are rarely so. Between the crossing and the start of the apse there are only three bays at Eeims, four at Chartres, five at Amiens, and five at Cologne, to take four of the best known of French and German cathedrals ; whereas Salisbury and Exeter have seven bays from the crossing Photo: W. A. Mansell & Co. Fig. 101. — Ely Cathedeal: Choie and Lantebn. Photo : Frith & Co. Fig. 102. — Salisbdby Cathedbal: Choib. [To face p. 152. THE DEVELOPMENT OF CHURCH PLANNING. 153 to the return wall behind the high altar, York and Ely nine bays from the same place to the east wall, and Lincoln ten. The greater length of English cathedrals aUowed of transepts EngUsh of much greater projection on either side, and in addition, of ranseP s- smaller transepts about halfway between the great transepts and the east end. Eastern transepts have already been mentioned as one of the features of the church at Cluny. From Burgundy they traveUed to England, and appeared first at Canterbury (c. 1170), Lincoln (c. 1192), and Eochester (c. 1210), found a place in SaUsbury Cathedral, when it was commenced in 1220, and at Beverley Minster a few years later. The reason for the great projection of the main transepts and for the additional transepts was simply to provide sufficient altar accommodation. The French builders needed no eastern transepts, and could be content with giving their main transepts slight projection because their chevet plan provided aU they wanted. In England it was different. In SaUsbury Cathedral — which may be taken as the plan regarded by English builders as the ideal one, inasmuch as practically the whole was built at one time, and after Norman influence was past — fourteen altars were possible, six in the main transepts, four in the smaller ones, two at the end of the presbytery aisles, one in the Lady Chapel, and one, the High Altar, in the presbytery; all facing west, according to the traditional English custom. To what extent the people had access to the altars in the Side eastern chapels has already been referred to. It seems pro- c ape s' bable that the custom was not the same in all countries, nor even in cathedrals in the same country. Otherwise it is difficult to account for the fact that in many French cathedrals chapels are buUt out beyond the nave aisles the entire length of the western arm of the church, whilst in others there are none, as at Chartres and Beims. In cathedrals commenced before the middle of the thirteenth century these side chapels are later additions. None were buUt in France before 1240. The cathedrals of Laon, Paris, Amiens, Coutances, etc., were without them at first. It was not until some forty or fifty years after they were finished that the aisle walls were taken down and the chapels built out beyond, between the buttresses. At Laon Cathedral the old window-sills were cut through, and the ends stiU show in the walls; at Amiens Cathedral the outline of the buttresses and their set-offs are easily traceable from the chapels inside, 154 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Side chapels in England. Possiblereasons for side chapels. showing where the new work joins on to the old; at 'Coutances Cathedral the ends of the string, which formerly ran along under the windows inside, are still visible. These side chapels, one to each bay of the nave, are separated from one another by walls, and are, so to speak, distinct shrines. The altar in each is generally placed on the east side, so as to face west, but not unfrequently it is against the outside wall, and faces north or south. At Coutances, the side waUs above the altars are pierced with unglazed muUioned and traceried openings which agree in size and design with the side windows, a very good and light effect being produced. In later churches, i.e. in those of the fourteenth and fifteenth centuries, the side chapels are part of the original design, and were built at the same time as the rest of the church. Chichester is the one English cathedral with side chapels Uning the nave, but even these are somewhat different from those abroad. There are two on the south side, each two bays long, and three on the north, two of which are also two bays in length. All were added at the end of the thirteenth century. The chapels were divided from one another by walls, the upper portions of which were probably pierced, as at Coutances. Lincoln Cathedral has two thirteenth-century chapels at the extreme west end, one on each side, but these hardly come under the same category as the rows of chapels in French cathedrals. Why were these side chapels not required in England, and how was it that the necessity for them arose so suddenly in France? These questions are not easy to answer. The chief reason was probably reUc worship. BeUcs were undoubtedly prized in England from early times, but it seems certain that they were never so numerous here, and that they were not regarded as so essential as they were abroad. Or it may be that the Englishman always preferred congregational worship to the more private devotions to which side chapels lend themselves. If the exact extent were known to which the laity were admitted to the chapels of the eastern arm, in both England and France, some light might be thrown on the subject. It is known that the relations of the monastic orders and the people changed somewhat about the middle of the twelfth century, and that there was less antagonism between them ; also that the power of the laity and their interest in Church matters increased considerably about the same time, but it seems difficult to connect these with the Fig. 103. — Laon Cathedbal : Aisle Wall cut through for 'Chapels. Photo : Author. Pig. 104. — Coutances Cathedral : Side Chapels. [To face p. 154. THE DEVELOPMENT OF CHURCH PLANNING. 155 remarkable increase in the number of the chapels of a church, especiaUy as that increase took place about a century later. It may have been due to a compromise. The monks and canons may have said, "Leave to us the east end, and we will build chapels for you in the nave;" or if the people had been shut off from the east from the first they may have become strong enough to assert their rights, and demand consideration. In the absence of proof, however, this is all conjecture. A definite statement which could be proved would be welcome, because it would account for one of the differences between French and EngUsh cathedrals, which is now difficult to understand, and also for the exceptions to the French rule in France. Although English cathedrals have not side chapels of the Chantry kind common abroad, many have chantry chapels. These, how- c ape " ever, are different. Chantry chapels appear first early in the thirteenth century, but most of those now remaining belong to the late fourteenth or to the fifteenth century, when little else remained to be done in the way of cathedral-buUding. They are smaU chapels buUt either to surround tombs of bishops, abbots, etc., as at Winchester — where amongst others are the tombs of the great buUders, WiUiam of Edington and William of Wykeham, bishops in succession of the See — or else at the expense of private individuals, who also left money so that masses could be chanted daily over their tombs for the repose of their souls. Chantry chapels are sometimes at the end of aisles, but in cathedrals are rarely built out beyond the main fabric.1 There are two, however, on the south side of Lincoln choir, flanking the porch, and their small dimensions and deUcate detaU give scale to the mass of the cathedral behind them. They are mostly inside, and are placed under the arches between the columns separating the aisles from either nave or choir. The greatest number are at Winchester, and there are also several at WeUs, Gloucester, Tewkesbury, SaUsbury, etc. The custom of building these chapels continued general for some time after the Eeformation. At Christ Church, Hampshire, there are three — two in the south aisle, dated respectively 1525 and 1529, and the third, the largest, in the north aisle, buUt by the Countess of Salisbury about the same time. Although there is not a single instance in England of a Double cathedral or large monastic church being planned with double ais es- 1 In parish churches they are frequently excrescences, and account largely for the irregularities in their plans. 156 A HISTORY OF ARCHITECTURAL DEVELOPMENT. aisles — Manchester Cathedral is merely a parish church, with extra aisles added subsequently — in France there are several. Double aisles, of course, were no mediseval innovation. The three greatest early basilican churches in Borne, those dedicated to S. Peter, S. Paul, and S. John, were planned with them, because, owing to the limitations forced on the builders by lack of funds and constructive necessities, in no other way could the space necessary for large congregations be obtained. Double aisles were adopted in French cathedrals for much the same reason as brought the chevet plan into existence, namely, to provide space for pilgrimage processions, which might pass round a church without interfering with the worshippers in the central ZIOTRE DAME, PARIS. -r---^---^"\?^ ¦,!••".?.'"".?«"""- •.'""^^^» *s hi. ''' * X • *\ * ,'\ !•*''' ! ^' •!•! J£^-'-'id( 200 300 FEET. Fig. 106. plans. portion and in the side chapels. In some of the great French cathedrals, such as Amiens, Beims, Chartres, Le Mans, only the east end has double aisles; in others, Paris, Bourges, Troyes, etc., they flank the nave as well. Most of these churches have aisles to the transepts in addition, and these, coupled with the double aisles east and west, or east only, give that effect of spaciousness at the crossing so marked in continental churches, and especially noticeable at Amiens and Chartres. In Antwerp Cathedral additions have given the nave three aisles on each side, with unsatisfactory results. The width of the nave and aisles together in this church is practically equal to their length. Very different from the many aisled churches of Northern France, are the churches entirely without aisles of the South. It THE DEVELOPMENT OF CHURCH PLANNING. 157 is difficult to believe that they belong to the same country, to the same style of buUding, and to the same period. The aisleless plan of Southern France owes its origin to Eoman tradition. The Cathedral of Albi is the direct descendant of the basilica of Constantine. Both buildings have the wide, central-vaulted space, the internal buttresses taking the thrusts of the vault, and the side recesses. They differ only in the number and size of these recesses. In the basilica there are only three; in the church there are twelve, exclusive of those round the apse. The reasons for the difference are that the mediaeval builders rejoiced in the redupUcation of parts which the Eomans, except in the Colosseum and other simUarly designed buUdings, did not regard as so necessary, and were afraid, perhaps, of the dangers attending the construction of great squares of vaulting a few inches only in thickness. Albi Cathedral,1 the church of the Jacobins at Toulouse (destroyed), and some others in the neighbourhood, prove con clusively that large churches can be built without regular aisles, and yet possess aU the essentials of Gothic architecture. And yet the fetish of the aisled plan stiU holds the field. The majority of the clergy stiU clamour for it, as though the naye and aisle plan were reaUy symbolical of the Trinity and an aid to devotions, instead of being merely a structural device for subdividing an internal space (so that it can be covered over easily, and at small expense) which was practised by the Egyptians in their Hypostyle HaUs of Assembly, by the Greeks in their larger temples, and by the Eomans in many of their Halls of Justice. In mediaeval days the aisled plan had its circumambulatory advantages, as aUeady stated, but for modern congregational purposes aisles of any width are an anachronism. No one can seriously maintain that columns and piers are not obstructions to the service of the present time, but in barely one modern church out of ten are they dispensed with. The problem of how to provide large unencumbered floor spaces was solved at Constantinople in Justinian's time, and in many cases by the architects of the Eenaissance, because at both periods recourse was had to the Eoman plan of few and large points of support. This plan can be followed as easily in the twentieth century as it was in the sixth and sixteenth centuries. No doubt its treatment presents archi tectural difficulties, but these difficulties have been overcome before. 1 For plan and section, see Figs. 202 and 203. 158 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Cistercian influence. The churches of Albi, Toulouse, Perpignan, etc., are proof of this, and their example has in some cases been followed in recent times. They have, in fact, inspired some of the most successful of modern churches in England, of which may be mentioned two early examples only, Bodley and Garner's S. Augustine's, Pendle- bury, and Pearson's S. Augustine's, Kilburn. In some churches of Southern France, La Trinite, Angers, for example, only the nave is aisleless, the chancel has the aisle divisions, the result being that, although the latter is still a fair width, the narrowed chancel arch and different proportions of the east end give scale and height to the church as a whole.1 Until the end of the eleventh century the Benedictine order was all powerful in determining matters of church planning and ritual, but in 1098, at the very moment when the order was at its zenith, was founded a rival one, the Cistercian. The Cistercian order was started as a protest agamst the extravagance of both Benedictine Uving and Benedictine buUding; although the Abbey Church of Vezelay, which S. Bernard of Clairvaux, the most famous member of the order, declared in sweeping terms to be unnecessarily elaborate and over-enriched, appears plain indeed in comparison with later Gothic churches. S. Bernard, the Puritan of the Middle Ages, as he may be caUed, laid down stringent rules for his followers. " A church," he declared, " shall be of the greatest simplicity, and sculpture and painting shall be excluded — the glass shall be of white colour, and free from crosses and ornaments." He further stated that " no towers or belfries of wood or stone of any notable height shaU be erected," and he seems to have entertained a rooted objection to triforia, as they are gene rally absent from Cistercian churches. Inasmuch as an apsidal eastern ending was universal in Benedictine churches, he declared for a square ending. The extremely Fig. 107.- -Plan from V. de Honnecourt's sketch-book. 1 The modern church of La Trinite in the Place Blanche, Paris, has a happy combination of wide nave and narrowed chancel. The nave in this church has aisles, but they are subordinate. THE DEVELOPMENT OF CHURCH PLANNING. 159 interesting sketch-book, which has luckily been preserved, of ViUard de Honnecourt, a thirteenth-century French architect, shows a sketch-plan of a Cistercian church. Under it the artist has written, "Vesci une glize desquarie ki fu esgardee a faire en lordene d'Cistiaux.1 To what extent S. Bernard's mandate influenced the develop ment of church planning is difficult to state with certainty. The effect of it has been exaggerated by some writers. The buUders of the great French cathedrals and large parish churches, in most cases certainly, paid little attention to it, and until the Eenaissance upset their ideas and destroyed their traditions ABBEY CHVRCH of MAVLBRO/VM. IlMi-Mi*,*' — _ ..if'-.::*:;:*:::*:::*:.-.*:::*:::*:::* _ * ' .''™.N-',;.N''' 'kv. .¦.':: W ''•';'. '-¦'' i-\' a \' * \'-',,/:i'! 1 ' x- I ' :". ; - \ ¦ •/,'','¦ > ¦ >' " ;' ' - • ;' <>' v " ' f.'t #:::ff:::*::-*::%::*: 10 o 50 scale orj_u I I I I I Pig. 108. continued to develop their eastern chevet endings as already stated. But in churches built by monks of the order on the country side, on the Continent as well as in England, S. Bernard's instructions in the main were followed. The plans of Chiaravalle, near MUan, Las Huelgas, Burgos, Spain, Maulbronn, Germany, Fontenay, France, and Furness and BuUdwas, England, show a remarkable uniformity. They are not, however, new plans, but reversions to an old one. They foUow the early basiUcan plan of a transept at the extreme east end with chapels extending 1 In modern French, " Voici une eglise carree qui fut projetee pour l'ordre de Citeaux." All the sketches have been reproduced by Lassus and Darcel, with explanatory notes, in a book entitled " Album de ViUard de Honnecourt," 1858. i6o A HISTORY OF ARCHITECTURAL DEVELOPMENT. beyond it, the only differences being that the chapels are some times more numerous than was thought necessary before, and ABBEY OT S.MARY, BVILDWAS SALOP. mTBnnc < 'ntniunt'or ' aisit " wail ¦:::-«:: ¦.».:-.*.¦.-.:»-.::.* 99 • 0 n A VE m^,W^m-^m^rWmm- 10 0 10 I I I J 40 J J oi unit w 80 j i SCALE or FEET 162'- 2lff Fig. 109 (Talbot Brown). they are rectangular. The church of Ebrach, Germany (c. 1250), shows an interesting struggle between Benedictine traditions EBRACH ABBEY CHVRCH , GERAAAIY. *:^J.::>|::^^:;.^- » 1 > T*fe 10 0 50 1 | TOO 150 SCALE OF FEET. V Fig. 110. and Cistercian rules. The east end is really a chevet, but a chevet squared. To the influence of the order are due the THE DEVELOPMENT OF CHURCH PLANNING. 161 eastern endings of Fountains Abbey Church, with its seven altars, and Durham Cathedral with its nine altars, both additions of the thirteenth century.1 But it is a mistake to suppose that that marked feature of English cathedrals and churches, their eastern rectangularity, is due to the Cistercians. The square east end is an EngUsh tradition. It is found in pre-Norman churches at Bradford-on-Avon, Wilts, and Dover Castle ; in churches in Ireland and Scotland of early date; in the eleventh-century Cathedral of Eochester; in the Abbey Churches of Eomsey (see Fig. 95) and S. Cross; and in countless eleventh and early twelfth-century parish churches in England, such as Old Shore- ham, Sussex, Dareth, Barfreston, and Patrixbourne, Kent, Adel, Yorkshire, etc., built long before S. Bernard became a power. All that that leader did was to revert to the simplicity of the early Christian plan which had been general in Egypt and Syria in the sixth and seventh centuries. The early square east end is not found only in Great Britain. In European countries it was never entirely abandoned during the tenth and eleventh centuries, although it was partially, owing to the power of Eoman precedent. The Cathedral and S. Nicolo at Bari (see Fig. 123), the Cathedral at Trani, all three in Southern Italy, have square east ends, foUowing Eastern custom ; and many churches in France, Germany, and Belgium might be mentioned, built before the commencement of the twelfth century, of similar plan. Two important rites of the Christian religion, baptism and Baptism! burial, might, in their development, have been expected to influence considerably church planning, but they can hardly be said to have done so. In the days of the early Church the separate baptistery was a necessity, as adult baptism was general, and the unbaptized were not admitted into the Church itself. But early in the eleventh century, as infant baptism became more customary, and parish priests were permitted to baptize, fonts began to be placed in parish churches, especially in Eome. In other parts of Italy, separate baptisteries continued to be built well into the tweffth century, as at Parma, Pisa, Novara, Asti, etc. The latest example of a separate baptistery is probably at Bergamo, where, early in the fourteenth century, a little octagonal buUding of delightful design was placed in a small court alongside the Cathedral.2 Outside Italy, with the possible 1 See Fig. 177. 2 This building has, I understand, been pulled down and rebuilt elsewhere. VOL. II. M 162 A HISTORY OF ARCHITECTURAL DEVELOPMENT. exception of S. Jean, Poitiers, no buUding exists in Western Europe which can be stated definitely to have been built as a baptistery, and the original destination of even the Poitiers example is uncertain. For so important a rite as baptism one would have thought that special provision would have been made in churches after the transferance there of the font, and that a portion of each building would have been specially set apart for it. But in no country was this done (with a few possible exceptions), probably because, as any priest could perform the rite in any church, the ceremony had lost, to some extent, its previous great importance. In English cathedrals the customary position for the font is near the western doorway, on the south side — in Winchester and Chester Cathedrals it is on the north — but its original position in every case may have been different. At Ely it occupies a good position in the south transept at the west end, and at Peterboro' it is placed in a simUar position and has a bay to itself, but it is unUkely that the somewhat unusual western planning of these two cathedrals was due to baptismal requirements.1 In parish churches the font is generaUy at the west end of the aisle, near the south door, and this is undoubtedly its original position as a rule, although in churches with western towers it not unfrequently occupies the space under the tower, or immediately to the east of it. Burial. More provocative of change was the permission to bury inside churches. In the early days of Christianity the good old sanitary laws of the Eomans still prevailed and bodies remained outside. But the protest against cremation, the insistence on burial on consecrated ground, the wish of the deceased or of his relatives to secure the most sacred spot, and above all, the desire to pay special honour to those who had devoted their lives to the advancement of religion, or who had lost them in its defence, brought about the change. It was strongly resisted at first. In 563 the Council of Braga gave permission for burial in church yards " in case of necessity," but on no account within the walls of a church. This was the thin end of the wedge. A later Council at Mayence decided that "no one shall be buried in a 1 In Truro Cathedral, Pearson made a feature of the baptistery, placing it at the west side of the south transept ; and in some modern churches special treat ment is accorded to the font and its surroundings. THE DEVELOPMENT OF CHURCH PLANNING 163 church but bishops, abbots, or worthy priests, or faithful laymen," 1 a fairly elastic and comprehensive list ; and another Council, held at Meaux, left it to the bishop and presbyter to settle who should be accorded the honour. Once the custom was sanctioned it spread rapidly, aud soon attained large proportions. Incidentally, it occasioned one of the most effective features of early churches, viz. the raised chancel, and also led, a few centuries later, to the Uttle chantry chapels, already referred to, which, in England especiaUy, add so much to the beauty of many of the cathedrals. The early burials were in small eastern crypts, not much more Raised than passages, as in S. ApoUinare-in-Classe, Eavenna,2 Torcello chancels' Cathedral (Vol. I., Fig. 125), and S. Ambrogio, Milan, altars being placed above the bodies. But the size of the crypt rapidly increased until, in many churches of the eleventh and twelfth centuries, it occupies the space under the whole area of the chancel, which is raised several feet above the floor of the nave. The best known of raised chancels are in S. Zeno, Verona, and S. Miniato, Florence (Figs. 116 and 117). It is impossible to overrate the fine effect produced in these two churches by this plan. In many of the earlier examples, of which the Eavenna Church may be regarded as typical, although the chancels are raised, the crypts are not visible from the naves. In the two churches of Florence and Verona, however, just mentioned, the fronts of the crypts are open and arcaded, and the vistas from the naves down into the vaulted crypts and along the length of the chancels above are amongst the most effective features to be found in mediaeval church architecture. Eaised chancels were contrary to the rule of the Greek Church, and consequently in those parts of Western Europe which came under its influence, notably in Southern Italy, although crypts are frequent, they are either entirely sunk, or else the nave floor is raised as weU as the chancel, so that the floor throughout the church is approxi mately level. In Trani Cathedral, an example of the latter plan, the crypt extends over nearly the whole of its area. In more northern Europe, in Germany, France, and England, raised chancels with crypts under are far from uncommon, but the west ends of the crypts are, almost without exception, closed. That all 1 A. Ashpitel, B.I.B.A. Journal, 1860-1861. 2 As S. Apollinare is stated to have been finished in 549, the Council of Braga's ruling appears to have been a protest against an existing custom. 164 A HISTORY OF ARCHITECTURAL DEVELOPMENT. were so originally is somewhat doubtful. It is unlikely that the northern mind could not realize the beauties and appreciate the aesthetic advantages of the more southern plan; and in many cases in which no view of the crypt is now obtainable from the nave, it is possible that alterations and subsequent additions have destroyed the original design. Italy. CHAPTER XI. ROMANESQUE IN ITALY. The chaotic state of affairs throughout Italy during the eleventh and twelfth centuries was naturally reflected in the architecture of the whole country. The history of that period is a continuous account of internal troubles and external wars; of conflicts between pope and emperor; of state fighting against state; of the efforts of the Eastern emperors to retain their footing in Italy, and of the desire of the Saracens and the Normans to wrest the land from them, or from anybody else who happened to be in possession. The Eomanesque of Northern Italy, i.e. in and around the ^°^hern plains of Lombardy, had its rise in the ninth century, although little of moment that survives was built before a century or two later. The Teutonic race of Langobards, or Lombards, who were conquered by Charlemagne in 774, were unskilled in building-craft, and there is no evidence to show that their buildings possessed any of the features which afterwards were distinctive of Lom bard architecture. Many existing churches which were formerly attributed to them are recognized now as belonging entirely to the eleventh or twelfth century, and, amongst the few which are older, none is earUer than the time of Charlemagne, except, of course, those buUt by the Emperors of Byzantium before the Lombards invaded Italy. There is always a temptation to ante date a building, especially if documentary evidence be followed, as most churches are of old foundation, and occupy sites previously buUt upon. This is the case in Italy more than in other countries, owing to its long history and to the vicissitudes which it has encountered. Charlemagne's empUe at the commencement of the ninth century included Northern Italy, France, the northern half of Spain, and the greater portion of Germany and Austria. At his death it crumbled to pieces ; but, after a century or more of 1 66 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Central Italy. Southern Italy and Sicily. confusion regarding the boundaries of the different kingdoms which rose from the ruins, Otto the Great consoUdated the German empire, and in 951 became, through marriage, king of Northern Italy also. For the next two hundred years the German emperors were all powerful there. It follows, therefore, that most of the Eomanesque churches in Lombardy were built when the country was under the rule of the Germans, and consequently owe some of their robustness to the energy and vigour of that race. The old methods of building which lingered there were not ignored, but they were gradually modified. Some of the modifications introduced were improvements, designed to meet the exigencies of ritual requirements and somewhat different construction ; whilst others were simply retrograde, owing to the inferior skiU of the craftsmen employed. The new feeUng which, in the eleventh century, made such headway in Western Europe, and spread with such rapidity in Northern Italy, met with little success farther south. The Apen nines formed an effective barrier to its introduction. Under their shelter the earlier traditions held their own. Venice also remained faithful to them, owing to its position, to its connection with the East, and to its independence. The Church of S. Mark is a proof how little it was influenced by Western developments. The papal territory was under the sway of the popes, and in Eome the basiUcan plan remained supreme. In Central Italy, or Tuscany, Byzantine influence in art matters was extremely strong, owing to the fact that Pisa, which for two or three centuries was the most important town in the district, like Venice, carried on a large trade with the East.1 Florence was of little note until the destruction of Fiesole (the old Etruscan city on the heights above the town), in 1125, drew the people from the hUls to the plain, and Lucca and Pistoja foUowed the lead of Pisa. The position in Tuscany, however, was more vul nerable, and northern methods of building were able to obtain some foothold there. Apulia, which embraced the greater part of Southern Italy, with Bari on the Adriatic as its capital, was from 871 until the middle of the eleventh century a Byzantine colony. Frequently 1 As evidence of the close connection between the Eastern empire and Pisa, it may be mentioned that when towards the end of the eleventh century the work on Pisa cathedral was at a standstill owing to lack of funds, the Eastern emperor came to the rescue, and found the money. ROMANESQUE IN ITALY. 167 harassed by the Saracens, the Eastern emperors managed to hold their own until 1040, when a new foe appeared in the shape of a band of Normans under Eobert Guiscard. For about the next hundred years Apulia was ruled by Eobert and his descendants.SicUy has had a more adventurous history than any other portion of ItaUan territory. From the early days of the Eomans and Carthaginians it was the seat of war, and later times brought it no peace. Its isolated position rendered it particularly open to attack. In 535, it came under the rule of the Eastern emperors, and remained in their possession until the advent of the Saracens, who in 827 landed and conquered the greater part of the island, although some towns held out until the end of the tenth century. The Saracens' reign over the whole island, however, was short lived, as the Normans from Apulia, after two or three expeditions, drove them out in 1090. CivU wars with various claimants to the throne foUowed, and Sicily may be said to have had but brief intervals of peace until quite recently. The architectural result of the changes and conflicts mentioned The title above is that there is a vast difference between the eleventh and esc,ue » twelfth century work of Northern Italy and that of the rest of the applied to country. In the north the churches of that period are mostly work. fuUy developed Eomanesque, and possess aU the traits customarily associated with work described under that head, such as ribbed and vaulted naves, clustered piers, and triforium galleries ; whereas further south, when these features do occur, they are found in conjunction with others which belong strictly to another and earlier art movement. The hyphened heading, Byzantine-Bomanesque, has been coined for the work, and possesses the advantage of sug gesting the main sources from which it was derived, but it does not express aU. In Apulia, and especially in Sicily, the Normans and Saracens both left their mark. The churches of Tuscany differ in many important essentials from those of Apulia, and, moreover, are by no means aU designed on similar lines ; and Sicilian churches possess certain marked characteristics which are absent from those of the mainland. In no other country is the archi tecture such a mixture as in Sicily. Nearly every church shows traces of Byzantine, Eomanesque, and Saracenic methods of con struction and decoration, mingled together in a most fascinating manner, and mixed with detail which only the Normans could have supplied. It is better, therefore, to group all under one broad 168 A HISTORY OF ARCHITECTURAL DEVELOPMENT. heading, as, if any other plan be adopted, half a dozen different groupings would be necessary to ensure correct division. Noethekn Italy. Milan was the centre of architectural advance in Northern Italy from the ninth to the twelfth century, and it was there, and in the neighbouring towns, that Italian Eomanesque was first fully developed. The most interesting churches are S. Ambrogio, Milan, S. Michele, and S. Pietro in Cielo d'oro, Pavia, and Novara Cathedral. All are vaulted throughout, have clustered piers, and, with the exception of S. Pietro, the large open triforium gaUeries common to twelfth-century work in all countries. The vaults are not in aU cases the original ones, in fact it is open to question Fig. Ill (Dartein). whether they are in some cases even copies of them, but there can be no doubt that these churches were designed to be vaulted.1 S. Am- Of the above-named, S. Ambrogio is the finest, and probably brogio. aiso ^ne earliest. According to Cattaneo, the east end dates from the time of Archbishop Angilbertus (824-859), the nave was rebuUt in the second half of the eleventh century, whilst the atrium and northern campanile followed at the beginning of the twelfth. The original nave of Angilbertus was of the ordinary basilican type, and the same length and width as the existing one. It had fourteen bays, whereas the present one has eight ; or, if one counts by the number of vaults, only four, as each bay of vaulting covers two bays of side arcading. In basilican churches the bays are always numerous and narrow, and the columns comparatively slender. In ItaUan Eomanesque ones a return was made to the old Eoman 1 See p. 75 for a discussion regarding the date of the S. Ambrogio vaults. Photo : Brogi. Pig. 112. — S. Ambrogio, Milan. Photo: Brogi. Fig. 113.— S. Ambrogio, Milan. [To face p. 168. ROMANESQUE IN ITALY. 169 plan of few and large piers (with smaller supports in between), which had been abandoned for some centuries, because in no other way could the vaults have been satisfactorily supported. The bays, in consequence, are few and wide. Here is the main reason for the differences in proportion and appearance between the two types of church. Each has its merits and demerits. In basilican churches, when the bays are arched, the side openings could have been wider, but the effect would not have been so good, as the scale which re- dupUcation of parts gives would have been lost, and the continuous timber roof or ceiUng would have occasioned an appearance of weakness below. In a Eomanesque church, massive support for the vaults was the main essential ; and although intermediate columns between the main piers possessed many advantages, unnecessary multipUcation of them would only have encumbered the floor space, without adding materially to the stabUity of the church. In S. Ambrogio nave there are three great piers on each side, with smaU, unimportant piers in between.1 The height from the floor to the crown of the diagonal ribs is over 60 feet, the crown of the transverse arches being about 10 feet less, as the vaults are domical. This is no great height considering the width, if comparison be made with Gothic churches, but it is sufficient. The proportions, in fact, are those of old Eoman vaulted build ings — the basUica of Constantine, for instance, being 83 feet wide and 120 feet high — and it is doubtful if the mediaeval men ever improved upon them. The piers and transverse arches are of stone, the diagonal ribs and side arches of brick, with pieces of stone inserted irregularly. These may be part of the original design, or be due to subsequent alterations.2 The charm of the church is due, in a large measure, to its Lighting lighting. This is ample ; although standing at the west end, and looking east, hardly a window is to be seen, except those in the apse. There are a few aisle windows, but the piers hide them from most positions. The west windows are large, but they are shadowed by the gaUery outside. The windows at the back of the triforium galleries give some assistance, but they are, of course, invisible 1 There is little difference in superficial area between the columns of the destroyed nave and the piers of the existing one, notwithstanding the decrease in the number of supports, and that half of the present piers are little bigger than the original columns. 2 The church has been extensively restored, so that the original treatment is uncertain. The third bay of vaulting from the west end has been entirely rebuilt in modern times. 170 A HISTORY OF ARCHITECTURAL DEVELOPMENT. West front,St. Am brogio. Externalopengalleries. from the floor of the nave. The easternmost bay of the nave is brilliantly lighted by a ring of windows in the octagonal cupola above, which throws a flood of light on to the altar and its baldachino. Altogether, the effect produced by the apparent absence of windows is surprisingly good. The west front of S. Ambrogio differs in some respects from other contemporary churches of North Italy, inasmuch as it is pre ceded by a cloistered atrium, which is returned at the east end to form the narthex of the church. Over the narthex is a gallery with three large openings in front. A single gable of flat pitch spans the whole front; the descendant of the Classic pediment. The treatment is very similar to that adopted in some of the Syrian churches mentioned in Vol. I. The simple gable is typical of Lombardic churches, even when, as in S. Michele, Pavia, the roofs behind do not agree with it. The roofing of the semi- domes over apses at the east end of these churches occasioned a little bit of rational construction which led to a pretty feature out side. This is the open AP3E OF -S.VINCELNZO- IN-PRATO, MILAN. IXT-" CENT. Fig. 114. gallery under the eaves, which was afterwards adopted at west ends, round turrets, and along the sides of churches. To support the timbers of the fiat-pitched roof protecting the semi-dome of an apse — for in Eomanesque work this was never aUowed to show outside — it was necessary to carry up the wall of the apse above the springing of the semi-dome. No great strength was required, and so this wall was pierced with openings which, simple at first, as in the apses of S. Ambrogio and S. Vincenzo-in-Prato, Milan, of the ninth century, soon became of great richness. At S. ROMANESQUE IN ITALY. 171 Michele, the openings are grouped in pairs by narrow pilaster strips or by strong clustered pilasters. In some other examples they form a continuous gallery, the openings being separated from one another by marble shafts. Germany copied the feature from Italy, and in certain parts of France, in Auvergne, and farther south, the design is frequently 5AM AUCHELE, PAV1A. ELEVATION OF EAST END. PLAN Fig. 115. met with. In Pisa and Lucca it was elaborated and dupU- cated to such an extent that the arcaded gaUery front is the distinctive characteristic of Pisan work. An interesting instance of Cupola the survival of Byzantine tra ditions is afforded by the cupola towers, either over the cross ing, when the church, like S. Michele, Pavia, or Parma Cathedral, is cruciform, or over the eastern bay of the nave when there are no tran septs, as in S. Ambrogio. These cupolas are, with few exceptions, octagonal, and are covered by flat-pitched roofs. They are even more common in Germany than in Northern Italy, and there are many in Southern Italy and in certain parts of France. They take the place of the Byzantine dome, and, like it, are in most cases surrounded by windows which, as in S. Ambrogio, admit a flood of light to the The same arcaded gallery treatment as They are not so striking east end of the church.1 to apses is generally carried round them as our English central towers, but, especially in Germany, where they often group with other towers and turrets, are very effective. 1 There are frequently domes underneath, which they cover and hide from view outside. 172 A HISTORY OF ARCHITECTURAL DEVELOPMENT. S. Zeno, Verona. Other churches. Porches. The church of S. Zeno, Verona, built in the first half of the twelfth century, has all the characteristics of a vaulted church ; sturdy clustered piers alternating with Ughter columns, and, on the outside, more strongly marked buttresses than are customary in Italy, and yet apparently it was never vaulted. Transverse arches, however, were evidently thrown across the nave from each of the big piers, starting from capitals which are now halfway up the walls. One of these arches still remains towards the west end ; the others were removed when the church was heightened early in the fourteenth century, and the choir remodelled. To this date belongs the present timber ceiling of the nave, a coffered barrel vault of unusual design, with tie-beams from side to side.1 S. Zeno has no triforium ; in this respect it differs from S. Ambrogio and other churches in Lombardy. Farther south, at Parma, Modena, Piacenza, etc., the somewhat later churches follow sometimes Lombardic methods, sometimes Byzantine. Thus, the west front of Piacenza Cathedral has a single gable, as at S. Michele, Pavia. The cathedrals at Modena and Parma have triforia, but in both churches they are small and very different in appearance from the spacious galleries of S. Ambrogio. Both churches also have clerestory windows. The Cathedral at Novara has a fine atrium in front of the west end. It is a thousand pities that this most effective addition to the approach to a church was discarded in later mediaeval days, and has not been revived in modern times. All architects are agreed as to the necessity for designing the approaches to and surround ings of a house, and what is necessary for a house is still more appropriate for a church. It is a blot on the Gothic shield that the dignified tradition left by the Eomans, and handed down by the early Christian builders, was ignored. The ritual reason for an atrium may no longer exist ; the aesthetic advantage of it is as great and as pressing as ever. A feature in all northern churches in Italy, and in a few southern ones as well, is the projecting porch. Such porches are sometimes two storeys high, as in the Cathedrals of Verona and Piacenza, but are more often of one storey only, as at S. Zeno, Verona (Fig. 71). They are vaulted, and possess, as a rule, two peculiarities. One is that the vault is carried on detached columns, standing well away from the wall, which could not 1 The following ohurches have similar roofs : S. Permo Mag'giore, in the same city, S. Stefano, Venice, and the Cathedral of Aquileja, not far from Trieste. Photo : Alinar i. Pig. 116. — S. Zeno, Verona. Photo: Brogi. Fig. 117.— S. Miniato, Florence. [To face p. 172. ROMANESQUE IN ITALY. 17 1 possibly stand if it were not for the iron tie above them, which prevents the vault from spreading; and the second is that the columns rest on the backs of animals. Whence came the latter device is difficult to say. In ancient Assyrian work, executed centuries before, pilasters against a wall often start from lions' backs in a similar way; but if any connection ever existed between these and the much later Western work, the links have been lost. The more likely hypothesis is that the idea originated north of the Alps — perhaps in columns of wood — where the love for the grotesque was strong, as the early carvings show. In Southern Italy, windows are sometimes treated in a similar manner. At the sUl level of the east window of Bari Cathedral, for instance, carved elephants boldly project and carry shafts which support a rich archivolt. On the south transept of the same building there are two windows with somewhat similar corbels, and in other churches of Bari, Tram, etc., further instances are to be found. Tuscany. In Central Italy greater variety in design is noticeable than in Lombardy, because no one influence was paramount. In Borne, old traditions stiU held the field, and there is little difference between the churches buUt in the eleventh and twelfth centuries and those of the ninth and tenth. In Pisa, Florence, Lucca, Pistoja, etc., the plans also remained basilican, but more traces of Byzantine feeUng are visible. In addition, the virility of the northern Eomanesque was such that it could not be confined entirely within its boundaries. > Even close to Pisa there is one church, S. Pietro-a-Grado, the outside of which is purely northern in design, whilst the plan shows no trace of northern influence, and old materials are freely used inside. Much the same may also be said of the two churches, S. Pietro and S. Maria, at Toscanella, near Viterbo, about fifty mUes north of Eome.1 In S. Miniato, Florence (c. 1013), built on a hUl above the s.Miniato, town, the three influences, Basilican, Byzantine and, to a less extent, northern Eomanesque meet. In many respects it is an ordinary basiUcan church, except that its crypt is larger, and its chancel more raised than in earlier examples. Old columns 1 Toscanella was an independent town of some importance, and it showed its independence in its architecture by ignoring, to some extent, the work immediately to the north and south of it. Florence. 174 A HISTORY OF ARCHITECTURAL DEVELOPMENT. and capitals — some of the latter much too small for the columns below — are extensively used; over the nave and aisles are open timber roofs ; and in its general plan, construction, and proportions, there is little to distinguish it from the earlier churches in Borne. It was never intended to be vaulted ; in fact, it could not be, and SECTION and PLAN S.MINIATO, FLORENCE. IO l_ o JL IO _l_ 20 I 5CALE OF FELT FOR SECTION. 5CALE OF FEET FOR PLAH. Fig. 118. every part of the design proclaims that fact. But — and herein Ues the main difference between this church and ordinary basiUcan churches — transverse arches span both nave and aisles and divide the upper part of the church longitudinaUy into three divisions, as though there were vaults between. The nave and aisle ceilings are, in consequence, not continuous as in the earlier examples. To ROMANESQUE IN ITALY. 175 carry these transverse arches, piers are introduced every third bay. These are not plain rectangular ones, like those in S. Clemente, Eome, etc., but are quatrefoil in section. One quarter of each on the nave side is carried up above the crown of the arches, forming a half-round pUaster, to support the main transverse arches, whilst from the quarters in the aisles spring the side transverse arches, starting from the same springing Une as the arches of the arcades. Byzantine influence in Tuscany is shown mainly in the Byzantine decoration of churches. The Tuscan architect decUned the con- mfluenoe' structional mixture of brick and marble, which produces such excellent effects in many churches in Northern Italy, and employed marble only. Sometimes he used it in alternating bands of light and dark — a method first used at Pisa, and afterwards carried to extremes in the. later cathedrals of Siena and Orvieto — or else Fig. 119. he covered his carcase walling with thin veneers of different- coloured marbles, arranged in simple geometrical patterns. The absence of triforium gaUeries — Pisa Cathedral is the only important church in which they occur — afforded excellent opportunities for the latter treatment inside, which, as at S. Miniato, Florence (Fig. 117), for instance, were not neglected. In addition, circles, triangles, and lozenges, formed of marble tesserse (sometimes of glass mosaic) were inserted in the plain marble facing outside; notably in the tympana of arches, between windows, etc., giving pleasant variety, and producing jewel-like effects. Hardly a church in Pisa, Pistoja, and Lucca, has not some of this work, which was also a favourite with the builders further south, as the west front of Troja shows. But the most marked features of Tuscan churches outside are Galleries. the arcadings and open galleries. The whole of the ground storey is generally arcaded, the arches springing from pilasters of slight 176 A HISTORY OF ARCHITECTURAL DEVELOPMENT. projection or from shafts of marble touching the walls. The arcading is carried across doorways, but emphasis is generally given to these by raising the arches above them. This is the rule in the churches of Pistoja, and was also adopted in the west front of Troja. The open galleries above, at Pisa and Lucca, often consist of three or four tiers across a west front, and sometimes along the sides of a church as well. They are seldom divided into bays by pilasters, as in Lombardy, but are continuous. The result is somewhat flat and monotonous in consequence, notwith standing the fine effects of light and shade produced by the arched heads and slender detached shafts. It was, perhaps, as a corrective LOWER PORTION OF WEST FRONT, I TROJA CATHEDRAL. I Fig. 120. to monotony that in the west front of Pisa Cathedral the architect did not place the bays of the arcading and galleries central with one another. The nineteen openings of the lowest gallery come irregularly over the seven bays of the ground floor arcade, and in the two top galleries eight openings are placed above nine in the gallery below (see Vol. I., Fig. 179). A somewhat similar arrangement occurs along the side of the Church of S. Giovanni, Pistoja. This is an instance of the arcaded method of design carried to extremes, which is only relieved from absolute mono tonous ugliness by the fact that the arches of the topmost storey do not centre with those below. The windows of Pisa ROMANESQUE IN ITALY. 177 Cathedral show also how unnecessary exact centering was con sidered by the Pisan architects. The clerestory windows are placed in alternate bays of the external arcading, but this arcade bears Uttle relation to the internal divisions of the church. The same may be said of the aisle windows, which come sometimes behind columns and sometimes between them. It was left for later architects to discover, with doubtful results, that everything in a church must be arranged symmetrically. One of the most deUghtful architectural pictures in the world Pisa. is that formed by the cathedral, baptistery, leaning-tower, and PISA CATHEDRAL. HALF PLAN OF CALLERY. HALF PLAN OF GEP FLOOR. \—-r B:z:SE:r®: : :a:::.s:-:js::^"| ::Br::B; :;^fc ;:b-™ ?V--;B=-fSsr"X= = =a'"~-ffl;"_E!:"'0'"rB'z--'s'— 3i^ SCALE OF FEET Fig. 121. campo santo at Pisa. The effect these buildings produce is not owing so much to their intrinsic beauty, since each has many faults, as to the way in which they group and harmonize with one another. The cathedral was commenced in 1063 and consecrated in 1118. It consists of nave and choir, both with double aisles, and wide, projecting transepts with single aisles and an apse at the end of each. The aisles throughout are vaulted, the rest of the church being covered by timber roofs. Over the aisles is a triforium which forms a high, wide gallery round the church. VOL. 11. N 178 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Exterior : Pisa Ca thedral. Refine ments. One peculiarity inside is that the transepts are cut off from the crossing by similar arcading to that of the nave, and by the triforium galleries above, which are carried across their ends. This plan shows that not only were transepts an innovation Uttle understood at Pisa when the church was built, but also the desire of the architect to adhere to the basilican plan of con tinuous arcades.1 The result is excellent; there is no feeling of weakness, such as is often apparent in mediaeval churches when the arches at the crossing are far higher than those at either side; and in few churches in Europe are there such fine perspective effects as can be obtained here, looking along the transepts across the church. The effect of the transepts outside is as unsatisfactory as the effect inside is the reverse. Having only single aisles, and being in all respects narrower than the nave, their roofs are lower than the nave roof, and cut into and abut awkwardly against the rest of the church. Neither is the oval-shaped dome over the crossing a beautiful object. The best part is the west end, which has considerable dignity. Its lines are truthful, inasmuch as they agree with the section of the church behind, which is more than can be said for most ItaUan western facades. But its virtue in this respect is a drawback. Nothing is more difficult to treat satisfactorily than a high central gable with half-gables lower down at the sides. The Greeks were wise in not attempting it; they made the ambulatories of their temples the full height, so as to allow of a single pediment at each end. The architects of S. Ambrogio, Milan, and other churches in North Italy, followed the same plan, although not always so truthfully. The builders of basilican churches got over the difficulty by masking the aisles by a narthex ; and the later mediseval ones by the introduction of western towers, or by screens as at Wells and Lincoln. Western towers, however, are rare in Tuscany, in fact, throughout Italy, and very few churches in Pisa and the neighbourhood do not follow the example of the Cathedral. Along the flanks of Pisa Cathedral the lines of the string courses, etc., are not straight, but rise and fall so uniformly that it is difficult to believe these curves are unintentional.2 In 1 The idea may possibly have been taken from S. Demetrius, Salonica (Vol. I., Pig. 130), where a somewhat similar plan is followed. 2 Measurements showing these curves are given in Cresy and Taylor's ROMANESQUE IN ITALY. 179 other churches in Italy, and notably in Tuscany, other curves have been recorded by Mr. Goodyear. Many of these there can be little doubt are due to settlements, earthquakes, thrusts of arches and vaults, or to alterations made from time to time which have disturbed the equiUbrium of the building. The fact that indifferent construction has, in some cases, forced piers out of the perpendicular, or careless setting out has led to walls being not in straight alignment, makes the proving of intentional inclina tions, either outwards or inwards, very difficult. Some of the simpler curve-and-line phenomena, such as the diminution and entasis of columns, the entasis of spires, the "batter" of towers and buttresses, are refinements admitted by every one. They are directly traceable to Classic traditions, preserved and handed down by Byzantine and Italian workmen from the Greeks and Eomans; and they were retained because the artistic sense of the mediaeval builder realized their value in correcting optical illusions. The north transepts of S. Fermo Maggiore, Verona, has brick pUasters at the angles which are carried up above the gable as square pinnacles. These pilasters are 41. inches wider at the bottom than at the top, the diminution in width being gradual, and the sides slightly curved. About halfway up are small closers.1 The desire to improve the internal perspective of a church by such devices as narrowing either one or both ends, dropping the height of a vault, decreasing gradually the width of the bays of an arcade, etc., were practised undoubtedly by both Byzantine and mediaeval buUders, although not universally by either. These devices are not confined to any one country. In Pisa Cathedral the easternmost arch at the crossing is some feet lower than the western one. In the beautiful Uttle early Gothic church of Montreal, Burgundy, the transverse arches of the nave are stilted, but the chancel arch is not stUted, and consequently the chancel vault is 3 or 4 feet lower than the nave vault. As regards the curves in plan of outside walls and the curves inside of walls and piers there is much room for scepticism as to whether they were in many cases intentional. That some outside walls curved, and curved uniformly, is undoubted. Poitiers Cathedral is an example. The walls here not only curve, but " Monuments of Pisa,'' published 1829, and by Mr. Goodyear in The Architectural Record (New York) of January, 1898. 1 I noted this diminution and rough entasis when at Verona in 1896. 180 A HISTORY OF ARCHITECTURAL DEVELOPMENT. the church is narrower at the two ends than in the middle. The logic of the curves in piers, described by Mr. Goodyear, is not always clear. Some piers curve outwards, others inwards. If the curves are intentional, the builders cannot have been all of one mind ; if unintentional, they can be accounted for in many ways. In the Eomanesque work of Italy and Southern France they can reasonably be expected as definite refinements introduced by the designers. In those countries old traditions may easUy have survived; but to believe in such intentional refinements in the later Gothic work elsewhere requires more proof than has yet been forthcoming. The subject is of considerable importance, but space does not permit of its being argued here.1 Apulia. The churches in Apulia, and Southern Italy generaUy, have suffered terribly from neglect, from Ul-judged restorations, and especially from unfortunate additions and alterations made in a particularly rococo manner in the sixteenth and seventeenth centuries. Notwithstanding this, they deserve more attention from the student than they generally receive. Their fittings wUl compare in beauty and interest with those of Eome and Florence ; the pulpits in the Cathedrals of Sessa and BaveUo, with their inlays of glass mosaic, being especiaUy noticeable. No part of Italy is so rich in bronze doors. This branch of industry, which was doubtless reintroduced here by the Byzantines, seems to have been practised by the local men to a considerable extent. The finest doors are in the Cathedrals of Troja, Trani, and EaveUo ; those in the two last-named being the work of Barisanus of Trani, and were executed about 1180. They are of soUd cast bronze, after the old Boman and Byzantine method, and are very different from the S. Zeno, Verona, ones, in which bronze plates and moulded and pierced bronze rails and stiles are naUed to doors of wood. The carving is generally far superior to that in other parts of Italy, owing to the district having been for so many years under Byzantine rule. In most cases it is extremely delicate and refined ; but occasionaUy it is quite the reverse, although the design is much the same, and may have been by Norman carvers, 1 The above was written before Mr. Goodyear held his exhibition in Edinburgh in 1906. At the moment of going to press, his reply to Mr. Bilson appears in the November number of the R.I.B.A. Journal. Readers are referred to this for arguments in support of his theory regarding the curves in Gothic ohurches. ROMANESQUE IN ITALY. 181 PIERCED MARBLE SLAB. BARI CATHEDRAL. F. M. s. PLAN. who copied with a heavy fist the lighter fingers of Greek artificers. This is most marked in a frieze round the top of the octagon cupola which covers the square at the crossing in Bari Cathedral. Byzantine influence is also conspicuous in the substitution of flush bands of glass mosaic for moulded dripstones over windows. Over each semicircular- headed window of the crypt in the east end of S. Nicolo, Bari, is a concentric band of glass mosaic, flush with the stonework, instead of the usual project ing hood-moulding, with squares of the same material let in above. The jewel- like effect is very telling, and all the more pleasing since it comes as a surprise. The pierced marble window- slabs of Bari Cathedral are amongst the latest of their type, and there are many others in most of the churches of Bari and Trani, those of S. Gregorio, alongside S. Nicolo, being especially numerous and beautiful. The churches on the Adriatic, or eastern, side are, as a rule, Ex- larger than those on the western, or Neapolitan, side. The most amPles important of the former are the Cathedrals of Trani and Bari (c. 1100), S. Nicolo, Bari, commenced 1087, and the Cathedrals of MoUetta, Canosa, and Troja, the last being built soon after the town was founded by the Greek prefect Bugianus, in 1017. S. Nicolo is the most characteristic. The west end is severely plain ; the east end a straight wall which masks the towers at the angles and the apse in the centre. The transept, as at Trani, is continuous from north to south. One peculiarity in the church is that the columns of the three western bays of the nave are coupled, the inner columns carrying transverse arches, which reach as high as the string-course below the triforium. These columns, although old, are manifest additions, but it is some what difficult to say why they and the arches were added. The most reasonable hypothesis is that they were inserted to strengthen the side walls, which are of good height.1 Pig. 122. S. Nicolo, Bari. 1 In Trani Cathedral the columns are also coupled, but there are no transverse arches to the nave. The outer columns carry the aisle vault. [82 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Outsidearcading. The best feature is the arcading of the aisle walls outside. The arches do not spring from pilaster strips or from slender shafts, as in Tuscany, but from wide buttresses projecting about 8 feet.1 In Trani and Bari Cathedrals there is similar arcading, although the projection is a few feet less. The boldness of these buttresses is all the more curious because a very little break is all 5. NICOLO, BARI. .SECTION AND PLAN. Fig. 123. that is necessary to ensure a deep shadow in the South. There are similar arcades, but of less projection, along the aisle waUs of the Cathedral of Issoire, Auvergne (see Fig. 153), and round the 1 In S. Nicolo, the aisle walls have been brought forward to the face of the buttresses to provide side chapels inside, but these are evidently additions, and no attempt has been made to bond the new work with the old. In Trani Cathedral the buttresses still show their full projection, and the arcade extends from west to east in an unbroken sequence of many bays. ROMANESQUE IN ITALY. 183 upper part of the chancel of S. Philibert, Tournus, Burgundy, but none in other parts of Italy. They cannot be due to Norman influence, because although contemporary churches in Normandy and England have buttresses of considerable width, these only project a few inches, and-the space between them is never arched. That the buttresses of Bari and Trani were not regarded as mere ornaments — no ItaUan ever did regard a buttress as an ornament, he conceals it whenever he can — and as proof that their statical value was weU understood, it may be pointed out that along the transept walls, against which no vaults exercise thrusts, there are no buttresses, only pilasters of slight projection to carry the arches.1 Above the outside arcading at Bari and Trani was evidently originally an open gaUery, as in, Pisan work. In S. Nicolo the gallery exists, but the openings are now walled up. In Trani Cathedral the gallery has been entirely swept away. In nearly aU the churches in and about Bari, the extrados of Arches. an arch is rarely concentric with the intrados. The latter is generaUy semicircular ; the former approaches closely to horse shoe form, the voussoirs being longer at the top than at the sides. This method of construction (a sound one it must be admitted) is not confined to any one part of Italy, although it is more general in the centre and south than in the north.2 The true horseshoe arch is more frequent in buildings round Naples than on the eastern side, showing that Saracenic influence was stronger there. It occurs over windows and doorways in the churches of Caserta Vecchia, Benevento, etc. Scattered about Southern Italy are several churches which Domed differ from those already described, inasmuch as they are domed. ° urc es- The Cathedral at Canosa has five domes — two over the nave, one over each transept, and one above the crossing. In plan it is therefore similar to many Eastern churches. The Cathedral at Molfetta, and the Church of S. Maria Immaculata, Trani, have three domes each, the central one being the highest. The aisles have quadrant vaults. At Trani, the central dome appears outside as a low octagonal tower, covered by a roof, whilst the side domes 1 In the Byzantine church of Daphni, near Athens, each transept has projecting buttresses, and the space between them is arched. It is not unlikely that Bari and Issoire took the idea from Greece. 2 Neither is it confined to any one period. The pointed arches of later mediaeval work are frequently built in the same way, and in the buildings of the Renaissance a pointed extrados and a semicircular intrados are common. 1 84 A HISTORY OF ARCHITECTURAL DEVELOPMENT. are completely hidden. The reason for domes in these and other churches of similar design was probably that the workmen, being MOLFETTA CATHEPRAL. LO/1GITVDIJNAL .SECTION. Fig. 124. Greeks, knew how to construct them, whereas the ribbed vault had not yet penetrated so far south. Various influences. Sicily. The influence of the Saracens accounts largely for the distinctive peculiarities of SiciUan churches, although theU construction is, in the main, due to Byzantine Greeks. In addition, the Norman element, which is often manifest, forms a link of interest to Englishmen. It is interesting to note how these influences react on one another. Thus, although the dome is a Byzantine feature, the stilted form used in Sicily comes from farther east than Con stantinople ; and although the lofty marble dadoes that line the walls in many churches recall Byzantine methods of decoration, much of the detail introduced into them is Saracenic. The mosaics that often cover the domes and upper parts of walls are unmistakably Byzantine ; but these are found side by side with stalactite-vaulted ceilings, which only Saracenic workmen could have made. The ceiling over the Cappella Palatina, Palermo (c. 1132), and the whole of the interesting porch, or open-sided entrance hall, in the Palace of La Zisa, Palermo (c. 1160), with Photo : Alinari. Pig. 125. — Cloisters, Monreale Cathedral. Photo : Alinari. Fig. 126. — Monreale Cathedral. [To face p. 184. ROMANESQUE IN ITALY. 185 its fountain and stalactite vaults, are the most marked examples of Saracenic work in the island. The crestings or battlements which crown some of the buildings outside may, with equal certainty, be ascribed to the Saracens. They present forms unknown in both mediaeval and Byzantine churches. The jambs of doorways and windows are often a curious mixture of Greek, Saracenic, and Eomanesque detail. Their mouldings are very shaUow; and pilasters, carved with acanthus leaves and other running patterns of Greek origin, are more common than shafts. The members of the arches are of correspondingly slight pro jection, and zigzags, lozenges, and other enrichments generally associated with Norman work, are frequently carved side by side with more classic detail. These, however, are as Ukely as not to be due to the Saracens, from whom the Normans borrowed — not necessarily from SicUy, but from Palestine — much of their ornamentation. Other arches and jambs have the heavy " billet " ornament — each billet a separate stone — which is characteristic of many of the churches in Cairo.1 An interesting instance of conflict between Norman and Cloisters, Saracenic ideas is afforded in the cloisters of Monreale Cathedral. Monreale The arches are moulded in true northern fashion, with a some- drai. what rough semicircular member as an inner order, and are stUted and pointed. The shafts supporting the arches are extremely deUcate coupled shafts, each alternate pair being inlaid with coloured mosaic, while the intermediate shafts are plain. All the capitals are exceptionally well carved, and the variety in design is most remarkable. A single oblong abacus covers each pair of capitals, but is nothing Uke wide enough to support the inner order of the arches above, which consequently overhangs. This shows conclusively that two sets of workmen were employed; one set designed the arches, the other the shafts and capitals, and no one made any attempt to bring the work of the two into relation with each other. All the arches of SicUian churches are pointed. The pointed Pointed arch was undoubtedly used in Sicily long before it came into arohes- 1 Good examples of this at Palermo are in the Cathedral, and round the lower windows in the tower of the' church of La Martorana. MONREALE CATH EDRAL CLOISTERS. H- ABACVS -a o r 2 1 1 J__ J Scale p M. S oFTeet Fig. 127. 186 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Stiltedarches. Detachedshafts. Inlays. general use in more western countries ; but it does not follow that there is any connection between this fact and the introduction of the pointed arch elsewhere. Although the Normans conquered Sicily in 1090, they had something else to do for the next thirty or forty years than bother themselves about the shape of arches, or, in fact, about building at all. No work of theirs was finished before 1130; Cefalu Cathedral was commenced in 1132, and Palermo and Monreale Cathedrals not until 1170 and 1174 respectively. There were undoubtedly plenty of native pointed arches on the island when the Normans landed, and these would have familiarized them with the form ; but adventurers — for so the Normans were — intent on conquest, would be less likely to observe architectural detaU than pilgrims to holy shrines eager to note everything they saw. In Palestine, pointed arches were more numerous than in Sicily ; and if the pointed arch were an importation, which is doubtful, it more likely came from there than from farther west. Most of the arches besides being pointed are stUted, Uke most early ones in France, especially in the south. The Saracens always stilted their arches, whether semicircular, horseshoe, or pointed, and the form used in Sicily is certainly due to their influence, although it is possible that their influence did not extend elsewhere. Very curious is the way in which stumpy, detached shafts are used at the corners of apses and other recesses inside most Sicilian churches. They are set back in reveals, and not only add colour and interest to the design, but also help to give scale. There are several such at the east end of Monreale Cathedral and in the entrance hall of La Zisa, Palermo. In S. Cataldo (see Fig. 129) the reveals remain in which they were placed, although the shafts have been removed. A favourite method of decoration on outside walls, which remained popular for centuries in Sicily for both domestic and ecclesiastical work, is a kind of inlay in which the stone is cut away and lava or black composition inserted. The lava some times forms the background to a pattern, sometimes the pattern itself. The outside of the Cathedral of Monreale is particularly rich in work of this description, and in Palermo Cathedral the walls of the porch, sacristy, and other portions of the building are covered with diaper patterns cut in the stone, but so slightly sunk as to be little more than incised. These lines were originally ROMANESQUE IN ITALY. 1S7 fiUed with black composition which has almost entirely dis appeared; the result being that, except in a strong light, the patterns are hardly visible. The Cathedral of Monreale, near Palermo, has by far the most Monreale striking interior of any Sicilian church. Its magnificent per- ^ai.6" spective is largely due to the way in which the eye is carried upwards towards the east end to the grandly designed mosaic head and shoulders of Christ which fill the semi-dome of the apse. The arch at the east end of the nave— which, by-the-bye, is not stUted — springs from about the level of the crown of the arches of the nave arcade, but the arch and vault east of the crossing, and the semi-dome beyond, are much higher. The different springing- MONREALE CATHEDRAL Fig. 128. lines are marked by simple abaci covered with mosaics. The choir vault and semi-dome over the apse, like aU the arches in the church, are pointed. The roofs over the nave, crossing, and transepts are timber, and the effect would certainly have been finer if there had been a dome at the crossing with vaults to north and south, like the one to the east. The capitals of the nave columns are finely carved, although their appearance is somewhat spoilt by the heavy cushion dosserets above covered with mosaics. Counterchanging is freely employed in the mosaic decoration on the waUs, soffits of arches, jambs of windows, etc. The figures are always either white or many colours on a background of gold, but wherever conventional patterns are used, as in the window jambs and arches, the patterns are in gold and the ground A HISTORY OF ARCHITECTURAL DEVELOPMENT. Cefalii Ca thedral. Churches. in colour. The panelled marble dado, 23 feet high, round the aisle walls, has its rails and cresting in coloured glass mosaics. The outside of the church is very striking on all sides. On the south side are the cloisters ; at the west end is a porch, sand wiched between the two western towers, which stand almost clear of the church; on the north side a loggia, an almost unique position for one ; and at the east end are stately apses decorated with intersecting arches (aU of which are pointed), the upper tiers being carried on detached marble shafts. In the Cathedral of Cefalii, Eomanesque feeUng is more apparent than at Palermo or Monreale, although it is the earUest church of the three. The arches of the nave are of two orders ; moulded square abaci take the place of the mosaic dosserets of Monreale ; at the east end is a triforium ; and the chancel and south transept are vaulted. The eastern portion of the church is much higher than the nave, and probably was built later. The walling outside is stuccoed, evidently the original treatment, as it is of rubble. The contrast between the stucco and the ashlar buttresses, windows, arcading, etc., is very striking, and may be recommended to those who have cleared the plaster off outside walls of churches in Eng land and caUed it restoration. At the west end the plan is the same as at Mon reale. Between the two projecting towers is a porch, and in front is a fine open court, surrounded by a wall, which recalls pleasantly the atrium of earlier churches. In addition to the three cathedrals already mentioned, there v 'V V. \v. 5. CATALDO. PALERMO V- f. Fig. 129. Photo: Alinari. Fig. 130. — Cepalu Cathedral : East End. Photo: Author. Pig. 131.— S. Giovanni degli Eremeti, Palermo. [To face p. 188. ROMANESQUE IN ITALY. 189 are a number of small churches throughout the island worthy of study. The most noticeable are in Palermo, and include the CappeUa Palatina, in the Castle, and the churches of S. Giovanni degU Eremeti, La Martorana, and S. Cataldo. AU were buUt by command of the Norman kings, but as the workmen employed were either Saracens or Byzantine Greeks, Northern characteristics are hardly perceptible. They are more thoroughly Eastern than the cathedrals, and most of them are domed. S. Cataldo (c. 1161), although shorn of its mosaics, is the most interesting, chiefly because in its present state its construction can easily be studied. The plan is nearly a square — but the two steps across prevent this from being too apparent — and is divided by columns in the same way as the smaUer churches of Greece, etc. (Vol. I., Fig. 174). The aisle apses at the east end are formed in the thickness of the wall, but the central apse shows outside. The three squares of the nave are covered by domes rising from octagonal drums, carried on squinch arches. The drums are circular in plan outside and there is no break between them and the domes, so that the latter have the bulbous appearance of Cairene ones. The oblong divisions of the aisles have intersecting vaults, without ribs, beautifully built with courses parallel to the ridges. The main difference outside between this church and simUarly planned Greek ones is that there are no transepts. The aisles are one height and are nearly as high as the nave, excluding, of course, the central domes. CHAPTER XII. ROMANESQUE IN GERMANY. Intro- The wealth of Germany in early Eomanesque churches has been duction. referre(i to in a previous chapter, and untU the middle of the twelfth century the Germans well held their own with other nations in the importance, beauty, and originaUty in design of their buildings. After that they made little advance. They were handicapped by their own creations. They could not shake them selves loose from the shackles of the style which, if they had not originated, they had undoubtedly done a great deal to foster. The extent to which the workmen had lost the faculty of design, and spirit of independence so essential for artistic advance, is shown in the church at Andernach, built at the beginning of the thirteenth century. It is a remarkably fine example of architecture, big in scale and well proportioned, but considering the hundred years or more which separate it from other equally fine churches in the country it is evidence of the stagnation which unfortunately had set in. Other churches of about contemporary date show a similar retention of forms and proportions which had long been abandoned elsewhere. An undue craving for novelty and change is no doubt a curse, and one that has done more harm to architecture than to any other art ; but the legitimate desire to improve on what had been done before, which animated the French and EngUsh buUders at the end of the twelfth century, was a laudable ambition. When in the middle of the following century the Germans awoke to the necessity for change, they had sunk too deeply into the rut of tradition to be able to extricate themselves unaided. They there fore turned to France, and in the fine churches of the north found models which they imitated with more or less success. But there is Uttle that is characteristically their own in their thirteenth and fourteenth-century work. The fame of German church architecture rests mainly on the early efforts of the eleventh century, and on the fine buildings of the following one. apses. ROMANESQUE IN GERMANY. 191 At the beginning, the Germans, like all other nations, owed much to Italy, but their dependence on that country was, from the first, coupled with a boldness and strength and sense of fine scale and simplicity in design largely their own. They used ornament sparingly, but what little they used is good. The forms of their capitals and the designs of their carvings are based on Byzantine work, but have a distinct character ; less refined, no doubt, than the work further south, but more in keeping with the soUd construction of their buildings. One of the most characteristic traits in German churches Double is the double apsidal ending; one apse being at the east end, the other at the west. Double apses were by no means unknown in earUer churches elsewhere, but in many cases they are due, as in the church in the ValpoUcella district near Verona,1 to the change in the orientation of churches, which resulted in the building of a new apse for the altar at the east end, and the retention of the original chancel apse at the west end to serve as porch or as baptistery. Some early Christian churches in the East were buUt in the first instance with double apses, but the plan of these was probably due to special ritual requirements which can hardly have existed in Germany many centuries later.2 The generally accepted reason given for the two apses in that country is that the eastern apse was for the abbot, or prior, and monks, the western one for the bishop and the people. Each is stated to have had its own choir. If this reason is the correct one the arrangement architecturally was a far finer one than that general elsewhere, which placed the people's altar at the east end of the nave and accorded it no specially treated surroundings. It was, moreover, a return to the ancient custom, altered in the fifth century, which allowed the priest to face the east without obliging him to turn his back on the altar. Whatever the reason for it, it is a tribute to the artistic power of the early German builders — perhaps, also, to the freedom of the people from monkish rule. In some churches, it is true, there is no western apse — Speier Cathedral is the largest exception to the general rule — but in most examples the apse at the west end is equal in size and importance to that at the east. In Mainz Cathedral the western 1 See Cattaneo's " Architecture in Italy," Fig. 30. 2 In the plan preserved in the Monastery of S. Gall, double apses are shown (see Fig. 93). 192 A HISTORY OF ARCHITECTURAL DEVELOPMENT. arm is far larger than the eastern. It was buUt (c. 1200) about one hundred years later than the rest of the church, and although this may account partly for its extent, its size and beauty are also proof of the determination of the laity in the thirteenth century that their altar and their representative should have noble surroundings.1 There is never more than one apse at the west end, although not unfrequently there are three side by side at the east end, as at Laach Abbey Church. These provided space for the additional altars required by the monks, and further accommodation could be obtained in the crypts, which, although by no means universal, LAACH ABBEY CHVRCH. aggg^BE ,--.-.V — ¦?. A ' i\ ' A ' A ' /> ¦• ,'. ^.'"•¦ii.' \L xi.' ^ \Lf '-^i' ^ % -*v;.t.v.:*.v;*v>*-.V- SCALE OF 'P.? i i i i gP too _L FEET. Fig. 132. Transepts. are not uncommon under the eastern choirs, as at Mainz Cathedral. The crossing in front of each central apse, west as well as east, is generally covered by an octagon tower, or cupola, carried on squinch arches, which rises above the roof, and on either side of it, in most churches, are transepts. These are invariably of slight projection, if they project at all. There is more compact ness about the plans of German churches as a whole than is generally found elsewhere. The main result outside is that there is little to show which is the west end and which is the east. Each end has its apse, its pair of transepts, as a rule, and its cupola ; whilst the unbroken line of the roof over the nave forms a connection. Considerable ingenuity is displayed in varying 1 The western altar is the one now used for service. ROMANESQUE IN GERMANY. 193 the designs, so that of no one church can it be said that the east and west ends are alike. And yet, notwithstanding the striking effects which are frequently produced, a feeUng of dis satisfaction is not altogether unnatural. One misses the marked differences in treatment between west and east fronts which are so fascinating in many of the churches of other countries. The most effective feature of the German church externally Towers is generally its skyline. This is owing to the fine balance which ^rets the two large cupolas give to the central main roof, and to the subordinate towers or turrets which are appendages to nearly all the large churches. The Germans understood thoroughly the art of grouping these to the best advantage, and of placing them so that they should assist, and not conflict with, the larger features. There are generally two turrets at each end. In some WORMS CATHEDRAL. ^I.A.y^*-*!^ ISO ^Sj^B wMd m,'.?\ ,7, f° \°° * 1 1 4 SCALE OF FEET. Fig. 133. cases they stand out in bold projection from the transepts, as in Mainz Cathedral and the Church of S. Michael, HUdesheim ; in others they flank the apses, as in the Cathedrals of Speier, Worms, and Bonn. In other countries circular towers are rare ; in Germany they were common in all the early centuries of church building. In the parchment plan preserved in the mon astery of S. Gall, two such towers are shown at the west end (Fig. 93). They appear as though they were detached, but this may be merely a defect in draughtsmanship.1 One result of the double apse plan is that, as a rule, the Lateral entrances to a church are at the sides, and not at the west end. entrances. 1 Some of the Ravenna towers are detached, but they are in different positions. The tower of S. Apollinare Nuovo is somewhat similarly placed, but this is attached to the south-west angle of the aisle wall. VOL. II. 0 194 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Open galleries. Pilasters, etc. There are exceptions, of course, when there is no western apse, as at Speier Cathedral, and in some cases also, as at Trier, Mainz, and Laach, there are doorways at the west end opening into the aisles. But the central western doorway is rare, not only in Germany, but also in churches in Switzerland built under German influence. It often finds no place even when there is no western apse, as in Zurich Cathedral. By its omission the Germans deliberately lost an opportunity for architectural display, and also for fine scenic effect. A central west doorway affords an excuse for the exercise of the architects' and the sculptors' skill, and at the same time provides the only truly dignified entrance to a church. It might have been thought that the Germans would have transferred their energies to the lateral entrances, but, with , a few exceptions, these, so far as can be judged from the original ones which remain, were small and unimportant. The transepts are too near the ends to have entrances through them, and the main doorways therefore, in most examples, face one another about the middle of the church. From Pisa and Lucca the Germans borrowed the arcaded galleries, especially noticeable in those towns. But they used them in moderation, and consequently with better effect. They never allowed them to degenerate into merely decorative append ages. Galleries are sometimes formed in the thickness of western walls, as at Trier Cathedral, where there are two, one above the other, and occasionally along the sides, as at Speier, but they are generally only under the eaves of apses and cupolas. In these positions they are especially effective, and are, moreover, particularly appropriate for reasons already given (see p. 170). The walls outside are, with the exception of the gaUeries, treated very simply and with great severity. The thin pilaster strips, and arched, corbelled string-courses of Lombardy are general, especially round the smaller towers and turrets; but otherwise there is little attempt at decoration. In this respect there is much greater resemblance to the architecture of Northern Italy than is found in the Eomanesque work of any other country. In S. Philibert, Tournus, it is true, the design is much the same as at Mainz, Speier, etc., which is accounted for by the fact that Burgundy belonged to the German Empire in the eleventh century ; but elsewhere in France and in England the pUasters are generally of different proportions, more structural and less merely decorative, and the string-courses of another character. Plwto : Frith & Co. Fig. 134.— Worms Cathedral; East End. [To face p. 194. ROMANESQUE IN GERMANY. 195 The east end of Worms Cathedral, with its octagonal tower Worms behind, shows well the circular turrets, the galleries under eaves, dral.6" and the pUaster treatment just mentioned. Moreover, it is a striking example of the desire of the Germans to obtain variety in the ends of their bi-apsidal churches. The square outside encloses a semicircular apse, and the whole east end is most happily grouped. In many large churches of the twelfth century the aisles only Vaulting. are vaulted — and not always those — the naves being covered by timber roofs. When vaulting was first employed for naves is somewhat uncertain. The three largest of the Ehenish Eomanesque churches, Speier, Mainz, and Worms, may have been intended to be vaulted from the first, but if so the probability is that the vaults were simple intersecting ones, without diagonal ribs. The nave vaults of Mainz and Worms are now ribbed, the ribs being moulded, and the transverse arches plain. In all three churches the vaults are very domical. The aisle vaults of Mainz Cathedral have no ribs, only transverse arches. This, coupled with the fact that on the nave side there are no projections from which either diagonal or wall ribs can spring, seems proof that the nave vault of this church is a later addition. The transverse arches are sUghtly pointed, an additional reason for doubting that the vault is the original one. In Worms Cathedral diagonal ribs may have been intended from the first, as the piers are large, and there are projections. These, however, are not conclusive proof of this, as there are similar ones at Speier, where the vaults, which date from the middle of the twelfth century, have no ribs. In nearly all the churches, late as well as early, the fine plan of one big bay of nave vaulting equalling two bays of aisle vaulting is followed. The early builders preferred this plan because of the fine effect which it produces, and not because they wished all bays square. They wanted their vaults to be big and to look big, and they relied on the subdivision of the arcade at the sides to give scale and to emphasize the result they were striving for. Most of the vaults are quadripartite, but in some cases where the vaulting is a subsequent addition, as in S. Maria, Cologne, the sexpartite plan is adopted. The vaults of this description in the Church of the Apostles, in the same town, are more satisfactory than is customary. None of the bays in either naves or aisles of the three large representative churches mentioned above is exactly square — those of Mainz Cathedral, in fact, are distinctly oblong. 196 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Absenceof tri forium. The grand simplicity inside of the three great Ehenish churches is one of their chief beauties. There are no triforia ; in fact, the Fig. 135. general absence of galleries from the majority of German churches —an early exception is at Gernrode (c. 1000) — suggests that, what- SPEllER CATHEDRAL SCALE OF Pig. 136. ever may have been the case elsewhere, there was no use for them in Germany. Desire not to increase the internal height a* Ei C5 ROMANESQUE IN GERMANY. 197 may have had something to do with their omission. The nave of Speier Cathedral is about 46 feet wide and 108 feet high. This is the taUest of the three. In Mainz Cathedral the width is the same, but the vault is some 20 feet lower. In Auvergne, as akeady mentioned, they introduced triforia, but did without upper windows. In the churches now under consideration they adopted the opposite plan. Opinions, doubtless, will always be divided as to which is the better, but the buUders in both Germany and Auvergne showed their good sense in not attempting to have both features. Arcade, triforium, clerestory, is no doubt a natural and jiroper division in a church of French loftiness, but in the earUer work, in which a more modest height sufficed, it means the cramping of either the top storey or the bottom one. In the large cathedrals of Switzerland, Bale, Zurich, and Lausanne, which, however, are more Lombardic than Teutonic, there are triforia, but the arcade suffers in aU cases.1 The Germans, although content with a moderate height inside, Arcades. were by no means averse to the effect which height gives. This is shown in the treatment of the nave arcades of Speier, Worms, and Mainz Cathedrals. There was evidently a strong desire to obtain verticaUty by continuing the piers above the springing of the arches. In aU three churches the piers are alternately large and small, the latter being perfectly plain rectangular ones, with attached shafts to carry the transverse arches on the aisle side only. In Speier Cathedral there is a similar shaft on the nave side as well. The design in the three churches varies slightly. At Speier and on the south side of Worms Cathedral,2 the face of the pier is carried up straight to the springing of the clerestory windows, arches being then turned from one pier to the next. The only break is the abacus moulding from which start the arches below, which are set back a few inches. In Mainz Cathedral the piers are not carried up so high, and the arcade really consists of two arches one above the other, the lower arch being set back, and the upper arch coming underneath the clerestory windows.8 The 1 In the thirteenth-century churches of Limburg-on-the-Lahn, Andernach, etc., galleries are also introduced — in the former there is an upper triforium as well — but in these churches the influence of Northern Prance was beginning to make itself felt. 2 The view is looking west. 3 In Jedburgh Abbey Church is a somewhat similar arrangement, but there are columns instead of piers, and the treatment is more architectural, although the British example is on a smaller scale (see Pig. 223). 198 A HISTORY OF ARCHITECTURAL DEVELOPMENT. effect in aU is exceedingly fine. It may be urged that it is somewhat bare. Those who feel that may have the consolation of knowing that the plain surfaces offer magnificent opportunities for decoration, and that probably they were painted originally. Those who are satisfied with fine, simple architectural lines and do not crave for ornament will be content with the churches as they are. Western The western facades of some of the churches which are not biapsidal suffer from the fact that the central west doorway is omitted. This gives a meaningless appearance to the whole elevation, as it lacks an important feature on which the eye can rest. In nearly all the later churches, however, and in some of the earlier, the omission is rectified. But if the churches lack a western doorway, they are seldom without western towers. Speier Cathedral is almost the only large church without a western apse in which they are omitted. In smaUer churches without them the western wall often has no gable, but is carried up above the ridge of the roof behind and is straight at the top. S. Castor, Coblenz, is one of the best known churches with western towers without the apse in between, but its front suffers from the fact that the towers are too close together. The nave behind is of con siderable width, and there is therefore no excuse for cramping. As it is the towers are partially behind the nave, with the result that the aisles with their roofs project beyond them with unsatisfactory results. This fault was remedied in the later churches of Ander- nach and Limburg-on-the-Lahn. Laach Abbey Church (c. 1112), besides being a remarkably good example of straightforward honest building, has one distinctive feature which is always welcome — a western atrium (see Fig. 132). This helps to make its entrance front one of the most effective in Germany, as the apse, central tower, and side transepts, with their circular turrets at the ends, rise above the low roof over the cloisters round the court, the cloisters adding scale to the church behind. Trier. Trier Cathedral is noted for being a remodeUing of an old Eoman building, and for having the biggest vaults of any Eomanesque church in Germany. The original buUding was a square, enclosing a cross (as in Byzantine churches), the arms of which were probably the same height as the vault over the centre. This accounts for the unequal spacing of the bays, a pecuUarity which was happUy followed in the western part of the church when that was built about the middle of the eleventh century. The ROMANESQUE IN GERMANY. 199 Eoman building was not pulled down, and in the walls of the existing church are brick voussoirs of the old arches and the capitals of four pilasters which may have belonged to the great piers at the angles of the central square. The large bays are about 53 feet square. When the vaults were built is somewhat uncertain, but probably not before the end of the twelfth century, or the beginning of the thirteenth. The vaulting is slightly domical ; the transverse arches dividing the large bays from the narrow ones are pointed and of considerable width, and the diagonal ribs are, by comparison, exceedingly thin. The vaults resemble so closely those of S. Pierre, Saumur (see Fig. 164), and the Churches of Poitiers, etc., that it is not unlikely aid was obtained from Southern France, TRIER CATHEDRAL. Fig. 138. where the workmen had already had considerable experience in the building of vaults of similar size. The vaults of the transepts, to the north and south of the eastern square, are the same height as those of the nave, the other side vaults being lower. The church thus retains its cruciform plan, although both the western and eastern arms have been lengthened. Altogether, whether viewed from the east or the west end, the interior is one of the most striking in Europe. The number of churches with apsidal endings to transepts as Trl: well as to choirs is unusually large in Germany. At Cologne east end. alone there are three — S. Maria in Capitolio, S. Martin, and the Apostles Church. S. Quirin, Neuss, and Bonn Cathedral are two other examples, but in the case of the latter a long chancel has been added, which destroys the original effect at the crossing. 2oo A HISTORY OF ARCHITECTURAL DEVELOPMENT. None of the above has aisles round its apses except S. Maria, and consequently this church is by far the most interesting of all. Mention has in previous pages been made of other churches elsewhere of similar plan, and their possible derivation from S. Lorenzo, Milan, and other Byzantine churches, has been referred to. The nave of S. Maria has to some extent lost its original character, as the vaulting is an addition, but the eastern portion is much the same as it was when built some time in the eleventh century, except that the decoration is new. It occupies the site of an older church, the foundations of which may have influenced its plan. The transept arms are covered by barrel vaults, the apses by semidomes, and the crossing by a low dome carried S. MARIA 1/1 CAP- -ITOLIO, COLOG/1E.; "\:& '"ff'-'ra i:::'0— .K-"hPh« wf" ,™. .W, ,*,"/.^"\'-^-":7 ¦;*;;^^4*7-%m fc-T ^^::^:.j^:$ffi ^fc ifcafefc^fej^i i:::_B- b '¦ -. : Hfpal l!"";'®v .Q>imi\ In IO D ' I 1 50 too 1,1 1 SCALE OF FEET. Fig. 139. Tournai Cathe dral. on true pendentives. Tapering columns with cushion capitals surround the apses and separate them from the aisles. To the little square chapels to the north and south, beyond where the aisles of transepts and chancel meet, much of the fine perspective effects at the east end are due. The whole of the eastern part is raised a trifle above the nave, and under the chancel is a well- lighted crypt, approached now by steps down from the two transepts. There is nothing that calls for special mention in the exteriors of these churches except the tower over the crossing of S. Martin's. That is about 40 feet square, without the octagonal turrets which come one at each corner and add greatly to its apparent width. Perhaps the finest cathedral in Europe in which the triapsal ROMANESQUE IN GERMANY. 20 1 Fig. 140. 202 A HISTORY OF ARCHITECTURAL DEVELOPMENT plan was adopted is at Tournai, in Belgium. The original chancel apse has been destroyed, and the choir belongs to a later date than the rest of the church, but the transepts remain practically un altered. The effects inside are very fine, but finer still is the result produced outside by the central tower with its lofty spire, and the four towers, also with spires, which flank the northern and southern apses. CHAPTER XIII. ROMANESQUE IN SOUTHERN FRANCE. The churches of the eleventh and twelfth centuries in France Intro- exhibit almost as much variety as contemporary ones in Italy, notwithstanding that the influences which caused the differences in the latter country hardly existed in the former. The Saracens had been driven out of France long before the churches now under consideration were commenced, and although here and there a detaU may suggest a Saracenic origin, the resemblance, in most cases, is probably accidental. Byzantine influence is more marked, but most of it was indirect, and was due to tradition more than to the employment of foreign workmen. The Eastern emperors had no footing in France, and to what extent, if any, Byzantine craftsmen entered the country is a debatable question. In the first haU of the eleventh century, before the native buUders found their feet, the Eastern plan, no doubt, was frequently foUowed, as many of the earlier churches show, but as the wealth and influence of the great Benedictine monasteries increased, and the people manifested more interest in religion, the reUance on foreign models decreased. The fine domed churches of Aquitaine are to some extent an exception, but only one, S. Front, Peri- gueux, is unmistakably Byzantine in plan. This is undoubtedly a copy of S. Mark's, Venice. The characteristics of the other domed churches may in part be due to Greek or Venetian colonies in the neighbourhood,1 or to the merchants of like origin who traded in those parts, but they are more likely the result of local traditions of buUding which had lain more or less dormant during the dark ages. Of Classic influence there are unmistakable traces, especially in Southern France. In the North, the remains of work of the ancients were few and unimportant ; in the South, Greek 1 Sharpe, in his " Domed Churches of La Charente," states that there was a large colony of Venetians and Greeks at Limoges. 204 A HISTORY OF ARCHITECTURAL DEVELOPMENT. buUdings, as well as Eoman ones, were plentiful. The monu ments of Classic art stiU existing there are probably but a hundredth part of what were standing in the days when the church-building movement was at its zenith. Eight hundred years ago these exercised enormous influence. The detail of the South is almost entirely based on Eoman models. It is true that some of the capitals are of the "corbel" type first used in the East (see Vol. I., Fig. 149), and that the diaper of different coloured stones — so marked in the volcanic Puy de Dome district, and existent also in all countries — had its origin in Eastern architecture ; but most of the capitals are free renderings of the Eoman Corinthian, and diaper work occurs in many of the earUer churches in France, such as S. Jean, Poitiers, S. Geneixrax, etc., and had become a vernacular trait long before the tweUth century. The fluted pilaster, so prevalent in Burgundy and Provence, is not Byzantine — in fact, it is seldom, if ever, used in Byzantine architecture. It is a Eoman feature preserved in Southern France, owing to the numerous remains there of Classic art and the conservative instincts of the people. To the same causes are due many Corinthian capitals which might belong to the second century a.d. ; mouldings of a deUcacy unknown in contemporary work in Northern France, England, or Germany ; and the carvings on the mouldings, the eggs and darts, acanthus leaves and guilloches, and fret patterns, differing but sUghtly from those executed by Eoman workmen in the later days of the empire, which had descended straight from Eome without Eastern modifications. But the abundance of ancient remains in the South only partially accounts for the differences between the churches of Northern and Southern France. The people of the South were of a different race from the people of the North. The language they spoke was different. Moreover, the King of France, in early days, could claim complete control over but a small portion of what is now France. Burgundy, at the beginning of the eleventh century, was part of the German Empire ; the dukes of Normandy and Brittany were to all intents and purposes their own masters ; in both South and North were other dukes and counts who practically ruled as they pleased their large domains. They acknowledged the king at Baris as their suzerain, but suzerainty in feudal days was a very doubtful quantity. In the latter half of the twelfth century more than half of France belonged to, or ROMANESQUE IN SOUTHERN FRANCE. 205 was dependent on, the King of England; and although the beginning of the thirteenth century saw a vast extension of the French king's territory and influence, a large portion of the country stUl remained semi-independent. No wonder, therefore, that, apart from differences of race and language, apart, too, from the presence or absence of Classic remains, the South differed from the North, and that in the South especially there is a variety in contemporary examples greater even than that to be found in Italy. An analysis of the Eomanesque churches in Southern France Analysis. shows that, although there are weighty points of difference between them, in some respects there is agreement. For instance, the aisleless plan of church was used throughout to an extent unknown elsewhere. It becomes rarer as one proceeds north wards, towards the Loire, but the two most important churches in Angers, the Cathedral and La Trinite, are of this type. Small churches without aisles are common to all countries, but in Southern France many large churches — churches with a span of from 50 to 60 feet — consist of nave alone. Some of these are covered by intersecting vaults ; others by barrel vaults ; a group are domed. The three methods of covering were employed irrespective of whether a church was aisled or not. The barrel vault is found everywhere, in large churches as well as in small ; on the Loire, in the far South, and especially in Burgundy and Auvergne. The ribbed intersecting vault did not penetrate into the southernmost parts until the twelfth century was nearly over, but on and about the Loire are some of the finest early examples of it in France. An irregular line drawn from Tours to Bordeaux passes through the country possessing the most notable of domed churches, but this method of covering was not confined to this one district. No satisfactory hard-and-fast division can, in con sequence of so many exceptions, be made dividing the churches of Southern France into vaulted or domed, into aisled or non- aisled. As, however, the covering, in these examples, presents more characteristic treatment than the plan, it is followed in preference to the other in this chapter. Perigueux may be taken as the centre of the domed churches Domed of France. In this town there are two, S. Etienne, built in part omirones- about 1050, and the great five-domed Church of S. Front, which, excepting portions of an earlier church which form a kind of narthex at the west end, was rebuilt after a fire in 1120. 206 A HISTORY OF ARCHITECTURAL DEVELOPMENT. The dome covering the original square bay of S. Etienne, like most of the domes of the district, is carried on true pendentives. 5. FRO/IT. PERICVEVX Fig. 141. The arches under it are slightly pointed, but otherwise it is thoroughly Byzantine in its plan and construction. The simi- S.FRO/^iT, PER.1GVEVX. section across transepts Fig. 142. larity between S. Front and S. Mark's, Venice (see Vol. I., Fig. 164), as regards plan, construction, and dimensions cannot ROMANESQUE IN SOUTHERN FRANCE. 207 be accidental. S. Mark's is a copy of the Church of the Apostles at Constantinople, pulled down by the Turks in the fifteenth century, and it is therefore possible, although extremely unlikely, that the architects of S. Front may have gone farther east than Venice for their inspiration. The rebuilding of S. Mark's was begun in 1063, and by the end of the eleventh century the fame of its plan, novel to Western eyes, had spread far and wide. Why the plan was adopted in this one church at Perigueux, and not in other domed churches of the district, is impossible to say. No records exist of Byzantine architects or workmen having been employed upon it ; and the influence of the Venetian traders, who no doubt sang the praises of their church wherever they went, can hardly in itself have been sufficiently strong to account for the adoption of a plan essentially Eastern. The only reasonable supposition is that as the dome was already recognized in the district as the natural covering for a church, the suggestion of a plan which would give an opportunity for five domes was eagerly adopted. The domes of S. Front, Uke those of S. Mark's, come one over S. Front, each arm and one over the crossing. All are carried on pointed ?f ri" transverse and longitudinal arches, wliich spring from heavy square piers, pierced with narrow semicircular headed openings. There is a certain clumsiness about the way in which the domes start from the pendentives which suggests that the builders were not thoroughly at home with the construction they had adopted. The domes are set back behind the face of the pendentives. In other churches of the district, mostly a few years earUer in date, Angou- leme, Fontevrault, etc., they spring naturally from the face of the pendentives as in Greek churches.1 S. Front lacks the arcading between the big piers which (as already stated in Vol. I., p. 244) helps to give scale to S. Mark's, and partly owing to this omission, aud partly to the absence of mural decoration, it does not look so large as its prototype. One refinement of the Eastern church is also wanting. All the domes of S. Front are the same diameter, about 40 feet, and there are no detached shafts against the piers, as in the transepts and choir of S. Mark's, to diminish the span at these places, to form pleasant breaks in both waU and vault, and to vary the size of the domes. This in itself suggests the absence of a Byzantine architect ; for no one who had 1 In S. Etienne, Perigueux, and Cahors Cathedral (c. 1100) the domes start as in S. Front. 208 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Aislelessdomedchurches. Angou- leme Ca thedral. worked on S. Mark's could have failed to notice the fine effect the breaks produce, or would have omitted them in any subsequent church for the design of which he was responsible. The lighting of S. Front is mainly from the side walls, and one misses the rings of windows in the domes which add so much to the charm of S. Mark's. There are a few windows in the domes, but they are very small.1 S. Front can hardly be said to be without aisles, although they are merely the width of the great piers supporting the domes, and are in no way cut off from the rest of the church. Most of the other large domed churches in the neighbourhood must be termed aisleless, notwithstanding that there are inside buttresses of sufficient projection to allow in most cases of strong longitudinal arches from west to east. These buttresses take the thrusts of the transverse arches, which together with the longitudinal ones sup port the domes (with the assistance of the penden tives), as in churches of Eastern plan. In some of the smaller churches such as Mou- thiers, Berneuil, etc., only the square near the centre of each church is domed, the nave and chancel being covered by barrel vaults. In all the larger ones, however, Angou- leme, Fontevrault, Cha- tres, SoUgnac, Souillac, etc., the naves, which, exclusive of the bay at the crossing, are from two to four bays long, are domed. The Cathedral of Angouleme (c. 1101-1119) may be taken as the typical example. In plan it is a Latin cross. The transepts 1 The church has practically been rebuilt in modern times, and much of its interest has gone in consequence. For particulars of the alterations made in the so-called restoration, see Mr. R. Phene Spiers, "Architecture: East and West," article on S. Front of Perigueux, etc. Fig. 143.— Church at Souillac. ROMANESQUE IN SOUTHERN FRANCE. 209 have considerable projection, as the squares ending them (over which are towers) are separated from the crossing by a bay on each side A AAtGOVLEME BFTH CATHEDRAL. p t iiS 1 • :'' ' il n ,' 'i'l !!' ''! Ii .1 '• .1 •ii si " ii> /1' m - ¦*'-''^>^» I ' v- - ' '^j s-. -' iJ~v - --'-it lar-s^iJtjk ^^^^^----^-^^^^^^^^^^^^^^^^ m^J ~— 1^— > iRi TJ^y IO 0 50 100 1 1 1 1 1 1 1 1 f.-:-! SCALE OF FEET. ArtGOVLEnE CATHEDRAL . SECTI0rt,0/1E BAY OP /1AVE Fig. 144. which finishes with an apse. The apses round the chancel are some what unusually placed, the arrangement being similar to that in Issoire Cathedral, Auvergne, except that there is no ambu latory aisle. The whole of the church, with the exception of the chancel and narrow transept bays, which are barrel vaulted, is domed, but only the dome over the cross ing shows from the outside, the nave domes being hidden by the covering timber roof. In the last fact lies one of the main differences between Angouleme Cathedral and other churches of similar plan, and Perigueux Cathe dral. In S. Front all the domes show ; and this gives it that Eastern look which the others lack. The covering timber roof also prevents any vol. 11. P Fig. 145. 210 A HISTORY OF ARCHITECTURAL DEVELOPMENT. possibility of windows in the domes, which, however, does not much matter in an aisleless plan because plenty of light can be obtained from the sides. In Cahors Cathedral, however, the domes which cover the two great squares of the nave are both visible outside, and are carried on moderately high drums. The above churches prove beyond a doubt that the aisleless plan can be employed on a large scale with excellent effects both out side and in. That the naves in these examples are domed is, so to speak, an accident. They could just as well be vaulted, like Angers Cathedral, which has a similar plan (see Figs. 47 and 162). The scheme of few and large points of support is the old Eoman one, and one cannot help feeling that it is superior to the more Northern ABBEY CHVRCH AT F0/1TEVRAVLT SCALE OF Fig. 146. Gothic plan of multipUcity of piers. The want of columns between the piers, which give scale in Eoman and Byzantine buildings, is not felt so much because their place is taken by the arcading on the side walls, which produces much the same result. In width, Angouleme Cathedral is about 53 feet between the walls and 37 feet between the piers. One of the objections to the aisleless plan is the danger of the chancel being unnecessarily wide. At Angouleme this is avoided by narrowing it, so that it is only a trifle wider than the space between the piers. At Fonte- vrault the arrangement is a still happier one. The square at the crossing is much narrower than the nave, and is separated from it by an arch of fair proportions with narrow openings on either side. Eastwards is an apsidal-ended chancel, round which is an ROMANESQUE IN SOUTHERN FRANCE. 211 ambulatory out of which open chapels. The result is a com plete eastern chevet joined on to an aisleless nave, any effect of ,0F"S. H I LAIR E, |W POITIERS. J~ Fig. 147. 212 A HISTORY OE ARCHITECTURAL DEVELOPMENT. Domed aisled churches. S. Ours, Loches. Barrel-vaulted churches. incongruity, either outside or inside, being avoided by the pro jecting transepts. Of the many domed churches with aisles only two need be mentioned, Le Puy Cathedral (Auvergne) and S. Hilaire, Poitiers, because these, besides being remarkably fine churches, illustrate two different methods of carrying domes without employing true pendentives. In S. Hilaire the domes are irregular octagons in plan, and the corners of the squares below are merely covered by rough squinch arches like those over the crossing in S. Ambrogio, Milan. This is also the usual method adopted in the many churches in Italy, Germany, and France which have domes over the crossings only, the naves being either barrel-vaulted or groined. In Le Puy Cathedral, in many of the bays, a more graceful method is adopted. The domes are well raised, and in the corners below, starting from the string-course which runs round each bay above the level of the top of the transverse arches, are niches covered by semidomes. The outer angles of the niches are enriched by shafts, and corresponding shafts flank the corners under the centre of each semidome. It is the pseudo-pendentive plan already noticed when dealing with the later churches of Greece (see Vol. I., Fig. 143), with the addition of shafts which improve greatly the design.1 A church that marks the conflict between the predilection of the Southerners for a rounded dome and the preference of the Northerners for a steep-pitched roof is S. Ours, Loches, to the south of Tours. The nave of two square bays is similar in plan to the domed churches farther south, such as Cahors Cathedral, and to the rib-vaulted, aisleless churches a little to the north, of which those at Angers are the most remarkable. It seems probable that, when the church was first planned, either vaults or domes were intended. But neither form was built. Each square of the nave is covered by a stone pyramidal spire, which is octagonal in plan inside and out. There is no inner ceiling or shell ; the corbelling for each octagon is treated decoratively, and it and the stonework above are visible from inside. The pair of spires are sandwiched between two towers, each of which has also a spire. The sky-line is a most striking one, and the combination of four spires, and otherwise no roof, is absolutely unique. Probably the most interesting of barrel-vaulted churches in 1 A somewhat similar design comes under the dome over the crossing of S. Philibert, Tournus, but the shafts there are differently spaced. Photo : Author. Fig. 148. — Le Puy Cathedeal : Dome over one Bay op Nave. [To face p. 212. ROMANESQUE IN SOUTHERN FRANCE. 213 France, of the first half of the eleventh century, is S. Philibert, Tournus, the nave and narthex of which were finished about 1019, the choir being a century later. The reason for the large narthex is referred to later (see pp. 222-3). The main point of interest to consider now is the vaulting of the church. In the nave, instead of rectangular piers or stumpy columns, as in most contemporary Fig. 149. churches, there are fine, lofty, cylindrical columns, built with many small stones to each course. Their diameter is greater than the thickness of the wall above, and from the segments project ing beyond the wall on one side spring transverse arches across the aisle, whilst on the other side the segments are carried up above the crown of the longitudinal arches to carry transverse arches across the nave, Much the same arrangement exists in 214 A HISTORY OF ARCHITECTURAL DEVELOPMENT. the church at Carcassonne.1 The chief peculiarity of S. Philibert, however, which differentiates it from all other barrel-vaulted churches, is that the nave transverse arches support barrel vaults which run transversely from north to south. The advantages of the plan are many. Each vault forms an abutment to the vault THROVCH flARTMEX. THROVGH /1ARTHEX & ffAVE. SECTIONS THROVGH rtAVE Fig. 150. on either side of it, and it is only at the east and west ends that precautions have to be taken to resist the thrusts. At the west end of this church there was no difficulty, as the narthex is two storeys in height. What the method of resistance at the east end 1 See Fig. 151. The same plan, elaborated, is followed in the later Gothic cathedrals in France, which, with few exceptions, have supports considerably thicker than the walls they carry (see Pig. 19, Reims). ROMANESQUE IN SOUTHERN FRANCE. 215 was originally is uncertain, as the transepts and choir are addi tions of the early twelfth century. At the crossing now is a dome, over which is a tall tower, the solidity and weight of which are more than sufficient to counteract any thrusts at this point. In the lower storey of the narthex, the central portion is vaulted with intersecting vaults, but the aisles are covered by barrel vaults which run from north to south like those of the nave, and like those in the old basiUca of Constantine, Eome. The central part of the upper narthex is roofed with a barrel vault running east and west, and the aisles by continuous quadrant vaults. It is a little curious that the method of vaulting adopted in the nave of S. Philibert was not more generally followed elsewhere, before the craze for great height set in. It provides perfect lighting, as the high side windows are almost invisible to any one standing in the nave, and it also produces proportions which must deUght all except those whose one idea in a church is height. The barrel vaults hardly show, unless one deliberately stands under a bay and looks up. The same method was adopted in a few later examples, but only over aisles, not over naves. Over the side galleries of Notre Dame, Paris, were originally barrel vaults running north and south, but these were swept away during alterations made some fifty years after the church was first built. In England we have similar examples over the aisles of Fountains Abbey, Yorkshire (c. 1150), and although the vaults are now destroyed, their haunches remain, and the original design is clear. There are no earUer examples in Europe than at Tournus, but as this mode of vaulting appears to have been a not uncommon one in Persia, it is possible that the idea came from the East. In the Tag Eivan, which is ascribed by M. Dieulafoy to the sixth century, barrel vaults, carried on very wide transverse arches, run north and south. The proportionate width of the transverse arches in the Tag to the spaces between them is much the same as in domed buildings, and it would be interesting, to know why the typical Eastern covering was in this and other similar buUdings abandoned for the vault, Other barrel-vaulted churches can be grouped under three heads. In the first, the churches have lofty side arcades supporting nave barrel vaults, which are generally pointed. They have no triforia or clerestories. Some have aisles, others merely deep recesses separated from one another by internal buttresses. Of the aisled type, S. Nazaire, Carcassonne, with its aisles covered by semi circular barrel vaults, is amongst the earliest (c. 1096). The 216 A HISTORY OF ARCHITECTURAL DEVELOPMENT. abbey church of Fontfroide and S. Trophime, Aries, are similar, but the aisles of both churches have quadrant vaults, and S. Trophime has small clerestory windows under the nave vault. The Cathedrals of Avignon and Orange, the abbey church of Montma- jour, and N. D. de Nantilly, Saumur, amongst others, have no aisles, but the side recesses are covered by semicircular barrel vaults. Their scale is very fine. The Mont- majour Church is 43 feet across, Orange Cathedral 45 feet, in both cases exclusive of the side recesses, which in the latter church add over another 20 feet to the width. Churches under the second head, mostly in Auvergne, have large triforium gal leries over the aisles, but no clerestories ; whilst those of the third have the three customary divisions, and belong mainly to Burgundy. Auvergne. There is probably no district in Europe in which the churches bear so close a resemblance to one another as in the old province of Auvergne.1 All might well have been designed by the same architect and built by one set of workmen. The most important are N. D. du Port, Clermont-Ferrand (c. 1080) (probably the earliest), the churches of Chamalieres, Orcival, Polignac, S. Nectaire, the last, although small, being one of the most perfect, and the Cathedrals of Issoire and Brioude, built in the first half of the twelfth century. Many of their characteristics they share 1 The Cathedral of Le Puy is an exception, but then it is the southernmost of all, and seems to-have been built under other influences. S.riAZAIRE, CARCASSOA1A1E. Fig. 151. ROMANESQUE IN SOUTHERN FRANCE. '-17 with churches of similar construction outside the province, such as S. Sernin, Toulouse, S. Etienne, Nevers, and the abbey church at Conques, but others are distinctively their own. In plan, they are all aisled, have thoroughly developed eastern chevets — that of Chamalieres has been altered inside — and, as a rule, projecting transepts on the east side of which are apsidal chapels. The chevets are remarkably well arranged, and their effect outside ISSOIRE CATHEDRAL SCALE OF FEET. 20 40 60 iO PLA/1. Fig. 152. js in all cases most striking. At S. Nectaire there are three apses, at Clermont-Ferrand four, at Brioude and Issoire five, but at the last-named the central apse is a square one squeezed in between two semicircular ones. The naves are covered by semicircular barrel vaults, with transverse arches at somewhat irregular intervals, and the aisles by intersecting vaults with, as a rule, transverse arches but no ribs.1 Over the aisles are spacious 1 In the church of Polignac, near Le Puy, the aisles are barrel-vaulted as well as the nave, 218 A HISTORY OF ARCHITECTURAL DEVELOPMENT. galleries ceiled by quadrant vaults which take the thrusts of the central vaults. There are no clerestory windows,1 but the churches are perfectly lit. The windows in the aisles are large, and a certain amount of light filters through the triforia openings from the small windows in the outside walls. The crossing and aisles alongside are as well lighted as the crossing of S. Ambrogio, Milan, but by a different method, and one that is peculiar to Auvergne. The four arches at the crossing are kept low, the dome over the crossing itself starting far above them. Over the chancel arch and in the walls on either side of it, above the arches of the aisles, are windows. The side windows are possible because the aisles dividing the transepts from the crossing are of great height, far higher than the nave and chancel aisles. They are covered by quadrant vaults, which start from the top of the arches at the ends of the barrel vaults of the transepts. Arched openings over the north, west, and south arches at the crossing, on a level with the windows, act as borrowed lights and help to distribute the light. The effect produced inside by this arrangement is very striking, and outside it is still more remark able. The crossing itself is generally covered by an octagonal cupola, and the quadrant arches over the aisles of the crossing by lean-to roofs which butt against it, and rise well above the other roofs of the church. The lean-to roofs, together with the cupola and chevet, give an unequalled effect of breadth to the eastern facade. Outside The covering in Auvergne churches is generaUy stone, in big slabs, bedded on the sloping top of the vaults. The ridges are also of stone, elaborately pierced and carved, and many are two or three feet in height. The roof over each eastern apse at Issoire and Brioude stops against a small gable, which prevents it from cutting unpleasantly into the ambulatory roof. At the apex of each gable is carved a Greek cross. These crosses, the elaborate ridges, the delicately carved capitals of the many shafts round the apses — those of Chamalieres are perhaps the best — and, above all, the diaper inlay of lava and red or white volcanic stone 2 in the gables and round the upper part of the apses are all so reminiscent 1 In the church of S. Etienne, Nevers, which belongs architecturally to the province, although it is some distance north of it, there are clerestory windows, which come down unpleasantly close upon the arches over the triforia openings, and would have been better omitted. 2 The richest diaper work in Auvergne is in the Cathedral of Le Puy (the centre of a volcanic district), especially in the cloisters attached to the church. detail. Photo : Author. Pig. 153. — Issoiee Cathedeal: Bast Front. Photo : Author. Fig. 154. — Issoiee Cathedbal : Apses at Bast End. [To face p. 218. Photo : Author. Fig. 155. — Issoire Cathedeal, looking East. [To face p. 219. ROMANESQUE IN SOUTHERN FRANCE. 219 arches. ISSOIRE CATHEDRAL. TREFOIL HORSESHOE ARCHES 1/1 TRIFORIWV. of Byzantine work that one cannot help suspecting that either the designers must have been Greek by birth, or else that they had been trained in an Eastern school. The arched recesses outside along the aisle walls of Issoire Cathedral are so similar to those at Bari and Trani as to suggest also a connection between, or like descent for, the buUders of Auvergne and those of Southern Italy. One curious trait found in many of the churches, generally Trefoil inside, is the trefoil or cinquefoil arch, horseshoed at the springing. One hesitates to ascribe this to Saracenic influence, but it is difficult to account for the feature in any other way, unless, like the arched re cesses, it filtered through from Southern Italy.1 In Issoire Cathedral all the openings in the triforium on the north side have trefoil heads, except in one bay, where they are semicircular ; whilst on the south side the very reverse is the case, all the heads are semicircular except in the one bay facing that in wliich the heads are of that form. Other churches of similar plan and construction outside Barrel- Auvergne are S. Sernin, Toulouse (c. 1090), its Spanish sister, lurches S. Iago, Compostella, the Abbey Church of Conques, S. Isidoro, farther Leon, Spain, etc. They differ mainly from the others in not south- having the central arrangement described above, and in the greater importance of their transepts. The similarity between the Spanish examples and those on the other side of the Pyrenees is easUy accounted for. No churches of any size were built in Spain untU Toledo was recaptured from the Moors in 1085, and the Spaniards, having no school of building of their own, naturally sought architects and workmen from Southern France. The 1 Street, in his " Gothic Architecture in Spain," shows a sketch of an extensively foliated arch inside S. Isidoro, Leon, Spain, which must have been suggested by some Moorish example, but the Saracenic occupation of the country of course lasted later in Spain than in Auvergne. Fig. 156. 220 A HISTORY OF ARCHITECTURAL DEVELOPMENT. S. Sernin, Toulouse. church at Conques is in some respects finer than S. Sernin, although not so large. It is barrel-vaulted throughout, except at the crossing. This is covered by an octagonal lantern, with a window on each of its sides, which throws a flood of light into the church. Its chevet has a fine appearance outside, but it lacks the decorative additions which give such distinction to the chevets of Issoire and Brioude. S. Sernin is the largest of barrel-vaulted churches in France (excepting the Abbey Church of Cluny, now practically de stroyed). The nave is narrow, being less than 30 feet wide, but then it has double aisles, which make the total internal width Bur gundy. S.SERA1I/H, TOVLOVSE. 10 o so 1 1 1 I I I I I SCALE OF FEET Fig. 157. of the western arm not far short of 100 feet. The transepts have single aisles on both sides, and aisles at the ends as well, like the Cathedral of Winchester. The crossing was probably originally covered by a low octagonal lantern tower or cupola, as at Conques, but at a later period the present telescopic tower, with spire over, was added, which necessitated strengthening the piers below it. The effect outside is striking, but hardly compensates for the damage done to the interior by the narrowing of the openings at the crossing. Burgundy was a large province, and its vicissitudes, coupled with its extent, account for the variety of design in its churches. Belonging to the German empire in the eleventh century, it did not become an integral part of France until some centuries Fig. 158. — Vezelay Abbey Chueoh, looking East. [To face p. 220. ROMANESQUE IN SOUTHERN FRANCE. 221 later, and for many years the Dukes of Burgundy were the equals of kings. But the real rulers of the duchy were the monks. The abbots of Cluny, Citeaux, and Vezelay kept almost royal state. TheU energies were unbounded, especially in the direction of church building. To them is mainly due the superiority of Burgundian churches at the end of the eleventh century and commencement of the tweUth, but to their conservatism must also be attributed the failure to take advantage of the develop ments which later took place elsewhere. From Sens in the north, to Autun and Tournus in the south, through the hilly district in the centre, known anciently as Le I I } 10 O JO 100 1 1 1 I I I I I 1 I I I SCALE OF FEET TME ABBEY CHVRCH OF VEZELAY, BVRGV/1DY. Fig. 159. Morvan, which boasts Vezelay, Saulieu, Semur, Avallon, are many churches of the early tweUth century of great interest and con siderable variety. The majority of these are barrel-vaulted ; the principal exception being the fine Abbey Church at Vezelay, which has the intersecting vault throughout.1 Here, in 1146, the French king, at the bidding of S. Bernard, announced the second crusade ; and from here, in 1190, our own Bichard I., and PhUip Augustus of France started for the Holy Land. The nave of Vezelay was consecrated in 1104. Its ten bays Vezelay. are oblong, and the transverse arches and wall ribs are semi circular ; all being stilted, the latter especially so. There are no diagonal ribs to either nave or aisles. Contemporary aisle 1 Except in the choir, which is later work. 222 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Narthex. vaults are common enough, but there are few nave vaults of this date, and particularly few oblong ones. The narthex of three bays was buUt 1128-1132, and there the vaults have pointed arches, but no diagonal ribs, except in the eastern bay.1 The absence of diagonal ribs is one of the peculiarities of Burgundian intersecting vaults. Their omission at Vezelay is natural, as the church is early in date, but their absence from much later vaults in the duchy is proof of the conservative spirit which reigned there. The church of Pontaubert, near Vezelay, built, it is stated, by the Templars towards the end of the twelfth century, is one instance out of many. It is vaulted throughout, and, although the central bays are over 17 feet square, there are no diagonal ribs. Burgundian churches differ from aU others in Western Europe in the importance of the narthex. The narthices are of two types : AVTVM CATHEDRAL WESTERN PORCH &\ io 10 _L_ Pig. 160. open porches, as at Autun, Paray-le-monial, Beaune, Semur, N. D. de Dijon, etc. ; enclosed, forming ante-churches, as at Tournus, Vezelay, and Cluny (see Figs. 150 and 159). Of the former, Autun (c. 1160) is by far the finest. The central bay in front is exactly double the width of each side bay, and the porch is two bays deep. Much of the effect it produces is doubtless due to the striking beauty of the sculpture in the tympanum over the great central 1 Viollet-le-Duc says that these ribs are additions. ROMANESQUE IN SOUTHERN FRANCE. 223 doorway (see Fig. 83), but, in addition, its proportions and di mensions are truly noble. The reason for these striking additions to the west ends of Burgundian churches is not quite clear. The narthex in the Eastern church, and in early churches in Italy, was for those who could not, by the laws of the Church, be admitted into the Church proper, but the regulations regarding admission in the Western church were different. The institution of infant baptism and the abandonment of adult baptism had destroyed the probationer class, and the rules against the entry into churches of evil-doers were not so strict as in earlier -days. Other reasons must therefore be sought ; and these may be found in the large number of pilgrims who, on days of high festival, flocked from afar, and in the need that existed for having some spot under shelter where they could be marshaUed. It hardly seems likely that even the enclosed ones actually housed the pilgrims, they were too smaU for that, but it is quite possible that they were used for this purpose when other sources failed, or in special cases. The enclosed narthex is generally attached to churches belonging exclusively to a monastic order; probably because the monks guarded zealously the right of entry into their churches. The open narthex is more commonly found in cathedral churches into which the people had the prescriptive right to enter. Whatever were the reasons for the narthex, they ceased to exist after the thirteenth century. The open porches of French and English cathedrals are very different from the great western additions of either Vezelay or Autun, and are merely architectural features. S. PhUibert, Tournus, the earUest of Burgundian barrel-vaulted Barrel- churches, has already been described. The other barrel-vaulted vaulted . j-iii . onurches. churches 01 the duchy are totally different in design, and are mostly one hundred years or more later in date. The first to be built was the great church at Cluny, which was begun in 1089, the nave being finished about 1130, and the narthex some fifty years later. Its total length, including the narthex, was 580 feet, which is about 50 feet longer than Winchester Cathedral. The peculiarities of its plan, the double aisles, double transepts, and eastern chevet have aUeady been mentioned. The greater part of the church has disappeared. The Cathedral of Autun, con secrated in 1132, is a copy of it on a smaller scale, and from it one is able to reaUze perfectly the design of the parent church. AU the principal arches and the barrel vault of the nave are pointed, as was the case at Cluny, aud the arches are stilted, as 224 A HISTORY OF ARCHITECTURAL DEVELOPMENT. was general in all Southern examples. The vault is divided into bays by a transverse arch over each pier. The aisles are covered by intersecting vaults without diagonal ribs. The fluted piers, n. 35 -O CE/1TRE i TO CE/TTRE ACROSS /1AVE TO rtEXT PIER AVTW1 CATHEDRAL, OriE BAY OF A1AVE. THE STO/1ES ARE ALL VERY LARGE AMD KEYSTOflES V5ED m ALL CASES I SCALE OF FEET F'M-3 I59S Fig. 161. or rather piers with fluted pilasters on their face, the carving and design of the capitals, the use of large stones and of keystones in the arches, pointed as well as semicircular, the delicate ROMANESQUE IN SOUTHERN FRANCE. 225 contours of many of the mouldings, the design of the triforium, with its fluted pilasters and semicircular-headed arcading, all proclaim the strong influence exercised by the old Boman remains in the town, and contrast curiously with the pointed arches and vault.1 The Cathedrals of Paray-le-monial and Beaune are practically identical in design with Autun. Beaune Cathedral has the same detached shafts immediately under the vault ; nearly the same arrangement of triforium and clerestory ; the same fluted pUasters as vaulting shafts ; the only difference of any importance being that the arches of the arcade spring from three-quarter columns instead of from fluted pilasters. The main differences between the churches of Burgundy and those of Auvergne, apart from the shape of the arches, are that the former have clerestory windows and the latter none ; and that in the Burgundian church the triforium is unimportant and the openings in it smaU, as it is essentially a bUnd storey, whereas in Auvergne it is of considerable importance. The work in Burgundy is a step in advance of the other, fine though that is, notwithstanding its reUance on old detaU. One thought occurs to one in studying it, and that is, how curiously extremes often meet; how transitional work produces much the same effect at the end of a movement as at the commencement. At Autun, Beaune, etc., the new order is beginning to take the place of the old ; in many of the early Eenaissance churches in France — S. Eustache, Paris, S. Michel, Dijon, for example — the new order, grown old, is being in its turn supplanted by a revival of what in the twelfth century was a disappearing tradition. Autun Cathedral and S. Eustache are wonderfully alike in many respects, notwithstanding that during the four hundred years that separate them, one of the greatest of all architectural styles had budded, blossomed, and withered. The fame of the great churches of Northern France, commenced Bib- in the first haU of the thirteenth century, has somewhat prevented churches fuU justice being done to the slightly earlier vaulted churches bordering on the Loire. Everything has been judged by Amiens ; and because these churches have not its great height and daring construction they have been counted as inferior by the student of architecture.2 And yet they have points in their favour, especially when judged by modern requirements, which the other lacks. 1 The two gateways at Autun are amongst the most interesting of Roman ones. 2 As well complain because the dachshund has not the greyhound's length of leg. VOL. II. Q 226 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Angers Cathe dral. They are of moderate height, and yet do not look low. They are sensibly planned and sensibly lighted, although in the latter respects, they cannot compare aesthetically with the type of church of which S. Ambrogio, Milan, and Issoire Cathedral are examples. They belong in plan and general proportions to the same class as the domed aisleless churches akeady described, the main difference being that they have ribbed intersecting vaults instead of domes. The most interesting are the Cathedral and La Trinite, Angers, S. Pierre de la Couture, Le Mans (all c. 1150), S. Pierre, Saumur, and S. Eadegonde, Poitiers (c. 1170). All these have aisleless naves. Angers Cathedral has transepts, which are A/1GER5 CATri- EDRAL.(s.mavrice). r^ ''i |fl ;tv iff ¦; * i \ m w / ' \ W L' l _ _ jjm ^ : *\T i. \ ' ? 'i i \i <* i, ii -¦ ' v I. i ' i ^ IT- — T_"7TP I, V 1 , £ I: \i/ j TT ' ' ^ tW 1 -' : ^^ i ii ^ ¦ io o t.j_ ^ is ^ I. ^ ' \ ' i >• / \ ./ i \/ I' \/ A ' /^ " / "> I1 / * ll ' \ / * ll ' >• il ' \ SO 100 1 1 1 1 1 1 _ >\ — I1 7l&- ,W ' \ i' /'i?vO-iI .' v l1 / ll v \ ^jW SCALE OP FEET. Fig. 162. additions, whilst the chancels of La Trinite, S. Eadegonde, and S. Pierre are earlier than their naves, and date from the eleventh century. The nave of Angers Cathedral, owing to its fine scale and simplicity of design, is exceedingly striking. The vault is very domical, like all the early vaults in France, and the transverse arches, diagonal ribs, and the longitudinal arches against the side walls are strong and bold. Each bay of vaulting is nearly square, the span being over 50 feet, and the height from floor to apex about 80 feet. These are no mean dimensions. Of course abut ment was easy. There were no aisles over which the thrusts had to be transmitted, and the supports could be any size the architect chose to make them ; in fact, the whole of the continuous waU ROMANESQUE IN SOUTHERN FRANCE. 227 could indirectly be utilized. The difficulties to be surmounted were therefore far less than in the later vaults, carried on thin piers and stiffened by flying buttresses. And yet their accomplish ment in the middle of the twelfth century was a wonderful achievement. The men who designed Angers Cathedral and other similar churches, S. Pierre de la Couture, Le Mans, La Trinite, Laval (another church of the same type), must have had a trace of Eoman greatness in their composition, and a keen ' distaste for everything mean and petty (see Fig. 47). Along the side walls of Angers Cathedral (and the two other churches just mentioned) are big, pointed arches, one to each bay, which produce a much better effect than the smaller arcading in the later transepts and choir, which is simUar to that in S. Pierre, Saumur (see Fig. 164). Over the arches, partly carried by them and partly corbeUed out, is a gallery, which runs round the church below the pairs of semicircular-headed windows.1 It is a pity that the transepts were ever added ; they detract from, rather than add to, the effect inside. The faults in the church are that the chancel is the same width as the nave, and that there is no break between the two. The pleasant feeling of mystery, the sensation that there may be a surprise in store, and the finer perspective effects which a chancel arch gives are in consequence lacking. La Trinite, Angers, is superior to the cathedral in this respect. La The openings on each side of the chancel arch are mere slits, and Trini^.Angers. LA TRiniTEr A/1GERS A. HALF PL A/1 .THROVGH SIDE APSES. B. DO. ABOVE DO. '£_ C. STO/1E SLOPE . SC Fig. 163. might have been treated more architecturally, but they are right in idea, as three openings look better than a single one. The vault 1 In the Le Mans church the windows are somewhat smaller, and over eaoh pair a circular light has been subsequently pierced, but the circles are no improvement- Saumur. 228 A HISTORY OF ARCHITECTURAL DEVELOPMENT. is sexpartite, clumsily built, and some 10 feet narrower than the Cathedral one. There is very little difference, however, in the width of the two churches inside, if recesses are measured, as on each side of La Trinite are semicircular apses, or chapels, in the thickness of the lower part of the wall. The upper part is set back outside, and above the slope over the semi-domes of the apses are strongly projecting buttresses, to take the thrusts of the main arches and ribs, and smaller ones for those of the intermediate transverse arches.1 s. Pierre, There is no great disparity in size between S. Pierre, Saumur, and S. Eadegonde, Poitiers, and the Angers churches, but there is a marked difference in detail. This is most noticeable in the vaulting. The transverse arches are still bold and sturdy, but the diagonal ribs are of the most slender description. Strong trans verse arches are a structural and aesthetic necessity when the vaulting is domical, and they possess also the great advantage of emphasizing the bays of the vault. It is extremely doubtful if the later builders, in their pride of craftsmanship, made any improvement when they reduced these arches to much the same size as the other ribs. Many a vault, both English and French, of the thirteenth and fourteenth centuries looks flat and monotonous, owing to the absence of strongly marked divisions. Moreover, the effect of length is lost. The eye wanders over the vault without really noting its extent. It is the transverse arches that make barrel vaults so impressive ; and possibly it was a know ledge of this that led the Burgundian architects, when building their intersecting vaults, to dispense with diagonal ribs altogether, more than ignorance of their structural value. They preferred that nothing should weaken the effect which the transverse arches produce. The thin diagonal rib, consisting merely of a single three-quarter round member, found in S. Eadegonde and S. Pierre, is not confined to these two churches, but occurs in nearly aU the contemporary domical-vaulted churches of the district. Another local peculiarity is the ridge rib, of same section and size as the diagonals. The presence of this feature is all the more curious because ridge ribs were seldom employed in other parts of France — never in early work — and the EngUsh vaults in which they appear are some fifty years later in date.2 The vaults over the 1 La Trinite has been so extensively " restored " that it is practically ruined inside. 2 Similar ridge ribs occur in the Cathedral of Trier, Germany. S. PIERRE, 5AVMVR . I ,.jp€ii ^itfillii? rOS-111 !H "Vi^l*" '/L<# 1 Writ- «r*c T J\ ! rWsz Fig. 164. 230 A HISTORY OF ARCHITECTURAL DEVELOPMENT. chancel of the Church of S. Serge, Angers, are probably the most domical in existence, and their shape is the more remarkable as they were not buUt until the thirteenth century. Existing chancels saved both S. Eadegonde and S. Pierre from the mistake made in the Cathedral at Angers, and each has its chancel arch. In the former church, this is flanked by narrow openings which are better managed than in La Trinite, Angers. For many reasons these two churches are more worthy of careful study than others which have a far greater reputation. W<^ The west fronts of Eomanesque churches throughout France are, as a rule, no richer than any other parts. In many cases there is a pair of towers, but these emphasize, rather than modify, the general plainness. At S. Etienne, Caen,1 the design is sim- pUcity itself; between the towers two tiers of three round-headed lights come over a doorway which could not weU be plainer. At the Abbaye-aux-Dames, Caen, more attempt is made at ornament, but the work is later, and the design of the rest of the church is somewhat less austere. The Cathedral of Angouleme has a front of exceptional merit, and the west end of N. D. la Grande, Poitiers, is covered Avith carving and sculpture. But simUar enriched west ends are rare. The towers, when they occur, seldom project more than a foot or so, and there are no examples having the bold projection which gives such distinction to the fronts of Monreale and Cefalu Cathedrals, SicUy. The original west front of Eouen Cathedral had apparently three arched recesses, somewhat simUar to those of Lincoln Cathedral; but the design is difficult to determine, as much of the earUer work is hidden by later additions. In Burgundy the west ends are more elaborate than elsewhere in France. A great deal of the richness against which S. Bernard thundered is found in the entrance doorways. At Avallon there are three which, for beauty of execution, are hard to beat. At CharUeu the carving over the doorways is not only exceedingly fine, but it is full of a truly remarkable classic feeling and deUcacy. The open porches of the Duchy have already been referred to. In Southern France there are many fine western porches of earlier date than these. The most notable are at S. Gilles, and at S. Trophime, Aries. But these are more the result of special effects of sculptors, probably imported for the purpose, than of the designers of the churches. In concentrating their 1 For Normandy churohes, see Chapter XV. Fig. 165. — Le Puy Cathedral : West Front. . .. _ ¦ --:- -" v Plioto: Author. Fig. 166. — S. Michel de l'Aiguille, Le Puy. [To face p. 230. ROMANESQUE IN SOUTHERN FRANCE. 231 efforts on one part, and not frittering them over a larger area, these artists showed sound judgment. The exceeding plain ness of the surrounding walls of both churches acts as a foil to and emphasizes the richness of the porches. The church at S. GlUes boasts three doorways side by side, and the design is more ambitious than that of S. Trophime. The most remarkable west end in Auvergne is that of Le Puy Cathedral. The church owes its imposing appearance to the com manding position it occupies on a steep slope near the top of a hill, and to this is due the unique arrangement of its western facade. The triple entrance porch, which occupies the entire width of the church, does not stand in front of the nave, but is underneath it. The approach to it is by a straight, wide stair case of many flights (each flight of about twelve steps), which continues under the porch, and then turning to right and left, leads on one side to the cloisters, on the other direct to the church above. The result is one of the most striking entrances in Europe. The peculiarities of the interior have already been referred to in this chapter.1 1 Le Puy possesses other original buildings ; not the least, although the smallest, of which is S. Michel de rAiguille, perched on top of a lofty needle of rock which rises abruptly from the plain. In plan it follows the outline of the summit, and is like a snail-shell. CHAPTER XIV. EARLY ROMANESQUE IN ENGLAND. Intro- The early work in England, executed some time before the uc lon- Norman Conquest, which commonly goes by the name of Anglo- Saxon, is an offshoot of the great Eomanesque style quite as much as the later Norman is. The examples of it are neither numerous nor, in an architectural sense, important, but are nevertheless exceedingly interesting, owing, in many cases, to certain traits they possess which are absent in later churches. Some were buUt before the eleventh century, others in the sixty odd years before WiUiam I. landed, and a few in the decade or two following. Two examples belonging to the last period are the churches of S. Mary and S. Peter, Lincoln, which are built at the foot of the hill on which stands the cathedral. No examples in England above ground can boast any great antiquity. None are contemporary with the old basilicas of Eome and Eavenna, or the great church at Constantinople. Christians some of the Britons were in the days of the Eoman occupation, but when Eome, attacked on all sides, recalled her legions in 411-418, the country fell into the hands of the Saxons, the people returned to paganism, and any churches which may have been built were destroyed. The foundations of the church in the Eoman city of SUchester, Hants, show that it consisted of nave, aisles, transepts, and narthex, and that at the end of the nave was an apse. In Ireland, Christianity continued after the Eomans left England, and, moreover, flourished exceedingly.1 In the sixth century the Irish church was second to none, and the fame of its teachers extended over the greater part of North western Europe. S. Columba converted Northern Scotland, S. Columban settled in Burgundy, and when driven from there passed with S. Gall and other monks into Switzerland. The number of churches abroad dedicated to S. Columba, S. Chad, and S. Gall, 1 Whether the Irish Churoh owed its origin to Borne or to farther East is a debatable point whioh can be left to ecolesiologists. EARLY ROMANESQUE IN ENGLAND. 233 etc., and the numerous old Irish manuscripts which have been discovered in Switzerland, Germany, etc., show how widespread was its influence. The second conversion of England to Christianity dates from the mission which Gregory the Great, Bishop of Eome, sent to this country in 596. In the following year, Augustine, MelUtus, Paulinus, and some forty other monks of the Benedictine order landed in the Isle of Thanet. Their success was immediate. In a few years the whole of England was once more Christian. The Eoman monks and their adherents early came into collision with the foUowers of S. Columba. The great point of difference between them, the keeping of Easter, was settled in their favour by King Oswy of Northumbria, at the Synod held at Whitby in 664, and after that the power of the Church of Eome increased, and that of Ireland declined. Christianity, however, was not to prevaU unmolested in this country for long, as towards the end of the eighth century the Danes commenced the series of raids wliich culminated in the submission of the Saxons to Sweyn in 1013. His son, Canute, received baptism, and then church building recommenced. The old Saxon chronicles state that the Danes " everywhere plundered and burnt as their custom is," and the last foray of Sweyn's seems to have been unusuaUy sweeping in this respect. Not many churches escaped the fire, and of those that were fortunate, a number were pulled down fifty years or so later by the Normans in order to be rebuilt on a larger scale and in more sumptuous fashion. WhUst some Saxon churches were of stone, the greater number Timber were buUt of wood, and were consequently easUy fired. Only c uro es" one church remains in England to give an idea of what these timber churches were like. That is at Greenstead, in Essex. The outside waUs are low and are built of halved trunks of oak trees placed verticaUy side by side, the rounded half being outside, with fillets of wood inside to cover the cracks. Without asserting that any of the original timbers remain — although possibly nearly all the existing ones are original — the church may be taken as a fair example of Saxon method of buUding. Another method was doubtless logs or balks laid horizontaUy, the ends of the logs crossing one another at the corners, in the same way that chalets, barns, etc., are built in Switzerland, and in other countries where wood is plentiful, at the present day. Owing more to the ravages of Sweyn and his predecessors than Existing to time, one cannot expect to find many entire churches in this ohurebes- 234 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Bradford-on-Avon. country earUer in date than the beginning of the eleventh century, and possibly there is not one. The crypts at Hexham and Eipon are undoubtedly older and probably date from 671-678. The greater part of Escomb Church, Durham, is also apparently about contemporary with these. Portions of S. Martin's, Canterbury, may well be older still, but the church has been much altered. The difficulty in determining dates is compUcated by the fact that windows and doorways have sometimes been re-used in later waUs, and that in some walls, which are evidently early, all the openings are later insertions. Eoman bricks in bands, and herringbone coursing, whether in brick or stone, prove nothing, as both were used after the Conquest quite as much as before. There was more church buUding in the seventh and eighth centuries than in either of the two following ones, and an attempt is made by archaeologists to differentiate between eighth-century detail and detaU of the tenth. Their arguments are not always entirely convincing and are too minute to be discussed here. Satisfactory reasons can be advanced why some churches cannot be earUer than the eleventh, but as regards an anterior date, much is conjecture and proof is difficult. The most complete early church in England is S. Lawrence, Bradford-on-Avon, Wilts, but there are no distinctive features by which its date can be fixed. BRADFORD- O/l- AVO/1 CHVRCH, Ipokch.B WILTS. In plan it consists of a nave and chancel with a large porch on the north side, and there are indications that a similar porch also existed on the south. The opening be tween nave and chancel is only 3 feet 6 inches wide. Its exterior is more archi tecturally treated than any other example, and the walls are faced with ashlar through out. The upper part is arcaded, the arches being carried on short pilaster strips, many of which are roughly fluted. The arches and pilasters are not built independently of the walling in between, but are formed merely by cutting away the face of the latter to a depth of about 2 inches. The masonry is very good, and this suggests that the church belongs more likely to the first half of Pig. 167. Photo : E. K. Piideoux. Fig. 168. — Bbadfoed-on-Avon Chubch : East End. [To face p. 234. EARLY ROMANESQUE IN ENGLAND. 235 ment. EARLS BARTO/1 TOWER, WEST EHD. the eleventh century than to the end of the seventh, the period which some would Uke to give to it. Its design, so unusual in England, may have been suggested by an Uluminated missal or a carved ivory, as on both arcading was a favourite motif ; or it may have been the work of foreigners. In any case the original inspiration is clear. The sixth-century churches of Eavenna, such as S. ApolUnare in Classe (Vol. I., Fig. 117), are decorated with simUar arcading in brick, and these form the connecting links by which this treatment of walls can be traced back to Imperial Eome. There is much more certainty regarding the dates of towers Pilaster which are ornamented with long, thin pilasters and rough cornices, treat~ or string-courses, as at Earls Barton, Barnack, Burton-on- Humber, etc. These cannot possibly be earlier than the tenth century, and more likely belong to the first haU of the eleventh. Their central posi tion at the west end is EngUsh ; but their design is Lombardic, and came either from there direct or through Germany or Burgundy. The majority of tenth-century towers in Lom- bardy — there are not many — are exceedingly plain. The earUest example, which is also the richest, is that of S. Satiro, MUan, which, according to Cattaneo, was built in 879. The pUaster and string-course method of ornamentation found on it did not become general in Lombardy, Ger many, and portions of France until one hundred years later. It is absolutely impossible that the examples of it in England can be earlier than the examples in those countries ; and the towers mentioned, and portions of walling in which similar Pig. 169. 236 A HISTORY OF ARCHITECTURAL DEVELOPMENT. features occur, are more likely to belong to the time of Canute, or his successors in the eleventh century, than to have been buUt in the tenth. The richest is at Earls Barton. There are pUasters of slight projection at the angles, and in between are rounded pilasters similar to those on the face of the north tower of S. Ambrogio, MUan, built 1129. This treatment has, of course, nothing whatsoever to do with joinery construction. It is in no sense an imitation of it. It is masonry construction ; and is especially valuable when, as in the majority of English examples, it frames in walls of rough rubble, possessing Uttle bond, which need strengthening at intervals. " Long A halo of mystery has been thrown over the so-caUed " Long a°d „ and short " bonding found in so many Anglo-Saxon churches. This, whether at angles or on the face of a waU, has been regarded as precious, and as peculiar to this country. As a matter of fact, in many countries and at most periods, from Italy in the days of the Eoman empire downwards, the people formed their quoins in this way whenever the stone of the district was such that it split up naturally into smaU rubble, with occasional bigger stones which could be used as stiffeners without squaring or chiselling. In Sussex and other parts such rubble and stones come out of small quarries daily now. All over Normandy simUar long and short bonding is found in garden walls and waUs of outbuildings, where- ever, in fact, the work is rough. There, as a rule, the short stones go through the thickness of a wall, the long ones only halfway ; pilasters not on an angle having two long stones, one behind the other. It is ordinary common-sense rough buUding, and is general where workmen are unskilled, or when the work is not of sufficient importance to justify the expense of labour on the stone. Sur rounding the precincts of the Abbeys of Jumieges and S. Georges de Boscherville ; at Caen, and in the country round ; at Chartres and elsewhere, are many such walls. The objection to this method of building is that the long stones are not generally on their natural bed, but that does not seem to have had a bad effect in Anglo-Saxon work. Long and short bonding has one virtue, it differentiates Saxon work from Norman, and is therefore some guide to the date of a building. But that is merely because the Normans were better masons, and had learnt at Caen to quarry bigger stones and to cut and square them.1 1 Saxon masonry is sometimes excellent, as at S. Lawrence, Bradford, and King Odda's Ohapel, Deerhurst, already mentioned, but the latter was built, as the Photo ; Autk Pig. 170.— Wall between Caen and Palaise. [To face p. 236. EARLY ROMANESQUE IN ENGLAND. 237 The pilaster strips on Sompting Church, are 9 inches wide, and project about 2 inches. The bonding stones are from 3£ inches to 6 inches high, and the vertical stones from 1 foot 6 inches to 2 feet 9 inches high, the last being the height of the bottom stone. The actual stones composing the pilasters are rather more than 9 inches wide, but portions are cut back to the face of the walUng, and their ragged edges now show. These were originaUy hidden by the stucco which covered everything except the straight-sided projecting strips. The brief historical sketch on a previous page shows that in Square the seventh and eighth centuries there were rival churches in eastend' England. This is of some importance, because it bears on one of the most interesting features of English church architecture — the square east end. In Vol. I. it was pointed out that although square east ends are common enough in Egypt, Syria, etc., all early churches in Eome had one or more eastern apses ; and in preceding and subsequent chapters of this volume the preference shown during the Middle Ages by aU continental nations for an apsidal ending is proved conclusively. This preference never extended to England, except during the century following the Norman Conquest ; and the reason why it did not may be sought for in the first place in the early churches of Ireland. These foUow the plan of the buildings in which the Christian primitives held their services. Like them they consist of two rooms (except the oratories, which are merely single cells), the larger one for the congregation, the smaller one for the altar and clergy. Both are rectangular ; no early church in Ireland has an apse. There is no need to speculate as to whether the Celts thought out this plan for themselves, it required no great effort of the imagination to do so, or had it suggested to them by others ; the only point of im portance to us is that these Irish examples show that the customary ending for a British church at the east was square. Existing pre- Norman churches with square east ends in England are double the number of contemporary ones with apses. Apses were probably reintroduced by S. Augustine — the earUer Eoman church at SU- chester had an apse, — and his church at Canterbury on the site of the present cathedral is stated to have had two apses, one at the east end, the other at the west. It was only natural that he and his followers should advocate the plan customary in Eome, dedication stone shows, " in the fourteenth year of the reign of Edward, King of the English," i.e. in 1056, when the intercourse with the Normans was very close. 238 A HISTORY OF ARCHITECTURAL DEVELOPMENT. General plan. Propor tions. WORTHCHVRCH, and if more of the churches built by them remained it would probably be found that all had the apsidal ending. After their death, however, the people returned in most cases to the traditional British form. Amongst existing churches the following have square east ends : Bradford-on-Avon, Wilts ; Wittering, Northants ; Escomb, Durham ; Dover Castle, Kent ; Breamore, Hants ; whilst Worth, Sussex ; Brix- worth, Northants; and one or two others have apses. Some of these apses, like Byzantine ones, are multangular outside and semicir cular inside. At both Brixworth and Wing, Bucks, the outer face is seven-sided. Below the apse of the latter is a small five-sided crypt surrounded by and inner faces are SCALE OF Fig. 171. an ambulatory. At Worth the outer concentric. The majority of churches have only nave and chancel, whilst some have a western tower in addition. The latter has akeady been mentioned as an early EngUsh tradition. The churches of Worth and Dover have transepts, which are narrower than the nave in each case. Churches with aisles were not uncommon, although few retain thek original ones, Wing being perhaps the only example. Augustine's large church at Canterbury had aisles, and existing side arches in the nave walls of Brixworth show that this church was also aisled. One peculiarity about most of the examples under consideration is that their naves are exceedingly lofty and narrow, proportions which exist also in Irish churches. In Deerhurst Church the nave is 38 feet long, 21 feet wide, and 38 feet high to the wall plate ; the chancel being of the same width and height with a length of 20 feet. The dimensions of the Bradford-on-Avon Church are : length of nave 25 feet 2 inches, of chancel 13 feet 2 inches ; width of nave 13 feet 2 inches, of chancel 10 feet ; height of nave to waU plate 25 feet 3 inches, of chancel 18 feet. The nave of King Odda's Chapel, Deerhurst, is 16 feet wide, the chancel 11 feet EARLY ROMANESQUE IN ENGLAND. 239 wide, the height of each being only 17 feet. In the three largest Saxon churches in England, Brixworth, Worth, and Dover, the height, although considerable, cannot compare proportionately with that of either Bradford or Deerhurst, but there is one pecu liarity about them worth noting : the nave of each is 60 feet long. The total internal length of Brixworth (excluding the tower) is 117 feet, of Worth 95 feet, of Dover 120 feet. The nave and chancel of Brixworth and the nave of Worth are 30 feet wide, the chancel of the latter being 21 feet wide. Dover has a nave 27 feet wide, and a chancel 18 feet 6 inches wide. These are all very fair dimensions — Westminster Abbey is only 35 feet wide — and although all three examples probably date from the eleventh century, it shows that large churches were built before the Conquest. Moreover, they look large, Dover especially. There is a bigness and simplicity about the chancel arch and arches to the transepts in that church which are particularly telling. The arches are unmoulded and of two orders ; the inner order being nearly the whole width of the wall, whilst the outer order has a projection of only 2 or 3 inches. The resemblance between these arches and early ones in Lombardy of two orders — those of S. Ambrogio, Milan, for instance — is very great. Saxon windows were probably not glazed, or, if glazed, the windows. glass was fixed in wooden frames. In many cases the openings were filled by pierced stone slabs, such as have already been described when dealing with Byzantine work, and Eomanesque churches in South Italy. At Barnack Church, Northants, there are two especiaUy good examples, the pattern being of the interlacing character which was such a favourite with Byzantine craftsmen. The windows themselves are sometimes semicircular headed, sometimes triangular headed, the latter being a common form with the Eomans, as the Baths of Cara- calla and other buUdings testify. The jambs in some cases are splayed on both sides, but more often are quite straight. When two or more lights are side by side, they are divided from one another by shafts in the centre of the waU. The capitals generally are of the corbel type already described as common to Byzantine and Eomanesque work (see Vol. I., Fig. 149), and have sUght projection at the sides, Fig. 172. 240 A HISTORY OF ARCHITECTURAL DEVELOPMENT. but considerable projection in front and at the back in order to reach from the outside to the inside face of the wall. This form remained the customary one in Italy, Germany, Southern France, and Switzerland through the eleventh and twelfth centuries, but the Normans abandoned it and generally placed their shafts near the outside face and carried their inside face independently of them. For that reason the presence of such capitals in England is a fairly certain sign that the work is pre-Norman, but cannot be absolutely relied upon. The two light side windows in the nave walls of Worth Church are particularly valuable examples, because Saxon windows in such positions are rare. Their jambs are straight, and so are the jambs of the tower windows of Sompting Church, Sussex. Most Saxon windows are in towers, in the beffry stage ; those in the Earls Barton tower having five lights. The shafts dividing the Ughts were generaUy turned in a lathe, and well turned too. They are considerably entasised and are numerously banded. The Worth ones are an exception, they being roughly worked by hand. Anglo- Mention has frequently been made in this chapter of the work— resemblance between pre-Norman work in England and con- Boman- temporary or earUer work abroad. All early churches in this esque' country, whether buUt by the followers of S. Augustine, by the Saxons before the Danes arrived, by the Danes after their con version, or by Edward the Confessor, are offshoots of Eoman tradition, or belong to the great school of Eomanesque buUding which prevailed throughout Europe during the early centuries of the Middle Ages. That they possess some characteristics not generally found elsewhere does not make them any the less Eomanesque. Other countries have other characteristic traits of their own. Whether the feeling came direct from Italy, or filtered through Germany and Burgundy, is immaterial. Those countries were certainly nearer to England, but the germ first appeared in Italy, and maturity had not been reached anywhere when the EngUsh churches were built. The connection between Eome and this country was fairly close in the days of Alfred the Great, and later under Canute. Both kings paid pilgrimages to Eome, and even if they did not bring back workmen with them, they must have returned with ideas. Foreign craftsmen are stated to have been imported to assist in the building of some churches, Monk- wearmouth amongst others. Who these workmen were is some what uncertain, as the term Franks applied to them might mean EARLY ROMANESQUE IN ENGLAND. 241 Germans, Burgundians,' or French. In Italy, in the eighth and ninth centuries, a large number of carvers were Byzantine Greeks, and it is possible, but not probable, that some of these men found their way to this country. The world was surprising small in those days notwithstanding the difficulties of travel. A design itself proves little regarding the nationality of the worker of it, because many designs which in their origin are charac teristically Byzantine — the interlacing one which appears in the window slabs of Barnack Church, for instance — had been adopted by ItaUan, German, and French workmen, and had become almost universal everywhere. The skill shown in the execution of a design is a surer guide, and the roughness of English work suggests a local carver. In connection with that, brief reference may be made to the many so-called Eunic crosses in Great Britain. That the patterns on these can have been originated in the north seems impossible. The claim for great antiquity formerly put forward for these crosses has been abandoned by most authorities, who place them mainly between the eighth century and the tenth. But almost identical patterns were common in Byzantine work of the sixth, as numerous carvings at Constantinople, Eavenna, and Venice testify. There is no suggestion from any one that Eunic crosses were carved by foreigners, but it may be suggested that they were copied from carved ivory crosses which might easily have been brought to this country in considerable numbers. Some of the figure sculpture of the South of France is probably founded on carvings in ivory executed by Byzantine workers. These men were far more proficient in the carving of beautiful patterns than they were in representing the figure, and examples of their skill in this respect travelled far and wide. There are very few important old churches in Switzerland and Southern France, for instance, which do not possess some specimens of their art, and in Italy, of course, examples are still more numerous. The carving on the jambs of the west doorway of Monkwearmouth Church is Byzantine in feeling, although it is rough. Only one church in England may be stated to possess a feature which is Teutonic, that is Sompting Church, Sussex. Its tower finishes with four gables, from the top of which starts a spire of shingles. This feature is certainly not Italian, and may well be derived from Germany, where there are many similar towers and spires of all periods (see Fig. 77). VOL. 11. B CHAPTER XV. ROMANESQUE IN ENGLAND AND NORMANDY. Intro- Foe nearly a hundred years after the battle of Hastings the duction. architecture of England and Normandy proceeded on identical lines. The movement began in Normandy, and sixteen years before Hastings, Edward the Confessor, with the aid of Norman craftsmen, had commenced his Abbey at Westminster. There can be little doubt that the Normans obtained their inspiration originally from Lombardy ; but they soon introduced special traits which differentiate their work from contemporary buildings in Italy, Germany, and other parts of France. They certainly did not bring the germ of their architecture with them from the North ; and they can hardly have obtained it from the neighbouring royal domain of France, which in the first half of the eleventh century had, to a great extent, lost the pre-eminent position in the arts which it enjoyed in the time of Charlemagne. The nationality of the two prelates who exercised such influence over WiUiam I. and his successor, supplies the clue to its origin, Both were Italians ; Lanfranc came from Pavia, Anselm was a native of Aosta. Nor- The earliest existing church in Normandy with architectural mandy. pretensions is probably that of the Abbey of Jumieges (c. 1050), and its ruins show that it was also one of the largest. Columns and clustered piers alternate, and although the nave was only covered by a timber roof, the aisles and gaUeries over were vaulted. Next come the two finest of the Caen churches, S. Etienne, or L'Abbaye-aux-Hommes, commenced 1066, and La Trinite, or L'Abbaye-aux-Dames, which, according to M. Enlart, was built between 1062 and 1066. The dates of the other early churches in Normandy are not known with any certainty. Bayeux Cathedral was consecrated in 1077, but a fire, some thirty years later, left little of the original building. The Abbey Church of S. Georges de Boscherville, one of the finest in the country, and the Abbey Church of Lessay, the design of which inside resembles Photo: Author. Fig. 173. — Jumiei.es Abbey Chubch, from the South. [To face p. 242. ROMANESQUE IN ENGLAND AND NORMANDY. 243 somewhat La Trinite, Caen, date from about 1100. Other churches at Cerisy-la-foret, Ouistreham, the port for Caen, Bernieres-sur-nier, etc., were probably built during the last five and twenty years of the eleventh century, or in the first quarter of the foUowing one. One of the most interesting of Norman churches is S. Nicolas, Caen, commenced 1083. The apse at the end of the chancel is covered by a very steep pitched roof of stone which rises high above the existing roof behind, and the apses on the east side of the transepts are covered in a similar fashion. The roofs of the two latter have a much better appearance than the central one, as they stop against the transept waUs under the eaves course. At the intersection rises a low tower, gabled on the east and west ends — an excellent way of ending a tower when funds do not permit of its being carried high. At the west end of the church is a narthex porch of three arches — an unusual feature in Norman work. Interesting though the above-named churches are, they cannot, England. with the exception of Jumieges Church and S. Etienne, Caen, com pare in size with the cathedrals and great monastic churches built in England during the last thirty years of the eleventh century and the following two decades. The extent of the list is astounding. Until the outburst of building in France a hundred years later, no country at any period could boast so many important churches in course of erection at one time. Little seems to have been attempted before 1070. The Conqueror and his nobles were otherwise engaged, and the priests had to accustom themselves to new surroundings, choose their sites, and make the necessary preUminary arrangements. But when once buUding operations began, there was little further delay. The following list contains the names of most of the large cathedrals and churches, and the approximate dates when they were commenced ; the letter B. signifies that the church was a monastic one of the Benedictine Order ; the letters S.C., a cathedral served by secular canons ; and A.C., a church belonging to the Order of Austin Canons, which had been founded in 1061 : — 1070-1080.— Hereford, S.C. ; Canterbury, B. ; Lincoln, S.C. Eochester, B. ; S. Albans, B. ; Winchester, B. ; and York, S.C. (practically no remains above ground). 1080-1890. — S. Paul's, London, S.C. (practically no remains) ; Ely, B. ; Gloucester, B. ; Malvern, B. ; Tewkesbury, B. ; Worcester, B. ; Pershore, B. ; St, David's (Wales), S.C. 244 A HISTORY OF ARCHITECTURAL DEVELOPMENT. 1090-1100.— Bury St. Edmunds, B. ; Chester, B. ; Chichester, S.C. ; Christ Church, Hampshire, S.C. (and later A.C.) ; Durham, B. ; Norwich, B. 1100-1110.— Sherborne, B. ; Southwell, S.C. ; CarUsle, A.C. ; Bangor, S.C. 1110-1120.— Peterboro', B; Waltham, A.C; Bomsey, B. (Nuns); Exeter, S.C; Llandaff (Wales), S.C; Kelso (Scot land), B. Southwell and Waltham, although served by canons, were not Sees, but of the monastery churches of the Benedictine Order named above the following were also cathedrals in Norman times : Canterbury, Durham, Winchester, Ely, Norwich, Eochester, and Worcester ; CarUsle became a cathedral soon after the church was built ; and to the list should be added Old Sarum, the original site of Salisbury.1 England and Germany seem to have been the only two countries in Eomanesque times in which a bishop's chaU was placed in a Benedictine monastic church. In such cases the bishop was little more than a nominal head, aU administrative power being in the hands of a Prior and Chapter. So many bishops in early days were "half priests, half warriors," were engaged in political intrigue, or in public duties for the good of the country, that they were often absentees from their churches for long periods. The fact of a church being merely a monastic one, or a cathedral ruled either by secular canons or lay monks, seems to have made no difference whatsoever in the plan. Peter boro' was only an abbey church, Ely an abbey church that was also a cathedral, whilst Chichester and Hereford had no monasteries attached to them ; and yet the only difference that can be noted is that the two last have slightly shorter naves than the others. Even this is no real distinction, as the nave of Eochester — a cathedral served by Benedictine monks — is even shorter. Gloucester, a monks' church, pure and simple, has some features which are absent from the others (see p. 145); but it stands almost alone in this respect. The Nor- The above Ust is no mean record of fifty years' buUding, and bffityand ** 1S far fr°m complete. To it must be added the hundreds of priest- parish churches which from Sussex to Northumberland, from Cornwall to Kent, were built either on new sites or in place of 1 Other Benedictine monastic churohes, such as Chester, Gloucester, Peterboro', etc., did not become cathedrals until the Beformation. ROMANESQUE IN ENGLAND AND NORMANDY. 245 those which were deemed unworthy. The Norman Conquest did something more than substitute one race of rulers for another. It introduced throughout the country new reUgious ideals, a new intellectual standard. The strict monasticism of the Benedictine Order had, it is true, been introduced into England a few years before ; but it had only found favour in London and one or two other centres. Now it was carried all over the country. Many of the Norman nobles had spent some time in the University of Paris, then the chief centre of intellectual activity in Western Europe ; Thomas a Becket was educated there later. The Saxons, both priests and people, were far behind the priests and people of France. The Normans may have been rough and rude, as judged by modern standards, but they were polished as compared with the greater number of the EngUsh thanes and nobles. The majority of French cathedrals were built after plan English development had practically ceased, and consequently, even when eSqUe altered later, each conveys the impression of having been begun churches and finished with very Uttle break. Not so with English unaltered. cathedrals, especially with the ones given in the previous list. Owing to various causes, there are few amongst those mentioned in which the whole development of EngUsh mediaeval archi tecture from Eomanesque to Tudor cannot be traced. The necessity for a longer choir led to the destruction of the apses of Canterbury, S. Albans, Ely, S. Etienne, Caen, etc. In no cathedral in England does the Norman east end remain exactly as it was when buUt ; even Norwich, otherwise complete in plan, has lost the central chapel of its chevet. Peterboro' retains the apse to its choir, but eastwards is an addition of later date. The Durham apse had to make way for the Chapel of the Nine Altars. The alterations to western arms and transepts have not been so extensive as to eastern. The craze for vaulting, however, has changed the appearance of many a nave, notwithstanding that the arcades, triforia, and clerestories remain much as they were built. In the naves of Ely and Peterboro', and in the transepts of Winchester, wooden ceilings remain; elsewhere, at Gloucester, Tewkesbury, Norwich, etc., they have been replaced by stone vaults. Structural defects sometimes entailed wholesale re- buildings, as at Ely Cathedral, where the central tower fell, bringing down with it the neighbouring bays. Fire, which played such havoc all through the Middle Ages, also did as much damage in England as in France ; although why it should have 246 A HISTORY OF ARCHITECTURAL DEVELOPMENT. been so frequent and fatal in vaulted churches is difficult to understand. At Chichester Cathedral, after the fire in 1186, the three-quarter attached shafts in the angles of the piers were replaced by more slender detached ones ; the cushion capitals disappeared, Ughter ones being substituted, and mouldings were worked on the wall face of the arches on the nave side in order to tone down their severity. The disgust which the heavy pro portions of Eomanesque evidently aroused in the later builders, and their desire to have their churches up-to-date, led to much fine work being entirely remodelled. At Gloucester, the monks, unable to bear the cost of pulling down and rebuilding entirely their choir, faced the whole with delicate pierced stone panelling. From their stalls all they could see was modern ; from the aisles and triforia alone were the solid piers and sturdy old arches visible. At Winchester, Bishop William of Wykeham, either more blessed with this world's goods or else a bolder spirit, carved the Eomanesque piers on the south side of the nave into the forms that pleased him better ; and on the north side, finding the other method too slow or too expensive, removed the facing stones of the piers, leaving only the cores which he recased. He swept away the triforium, raised and altered the shape of the arches of the main arcades, inserted new windows in the clerestory, and over the whole of the central area built an elaborate ribbed vault (see Figs. 105 and 219). As he paced up and down the church, provided he did not allow his eyes to glance to right or left as he passed the transepts — where the old work remained in aU its native vigour — he might congratulate himself that he was in a brand new church. At Exeter much the same thing had been done some fifty years or so before, but in a stUl more drastic fashion; and at S. Albans, on two separate occasions, with an interval between them of nearly a hundred years, determined attacks were made on the old work. Length of One of the most marked pecuUarities of English Eomanesque churches is the long nave. The nave of Norwich Cathedral has four teen bays ; the Cathedrals of Ely and S. Albans have each thirteen; Peterboro' has eleven, and Winchester, which now has twelve, origin aUy had the same number as Norwich. In the west of England, at Gloucester, Hereford, and Tewkesbury, and also at Eochester and Chichester, the naves are shorter. On the Continent, with the exception of the great Basilican churches in Eome, S. Peter's, S. Paul's, etc., there are very few churches which can compare naves. ROMANESQUE IN ENGLAND AND NORMANDY. 247 in length of nave with English examples. S. Sernin, Toulouse, is one. It has fourteen bays, including the two between the towers at the west end. The church at Vezelay has only ten bays, and even at Cluny there were only eleven ; excluding in each case the narthex, which formed practically a separate church. At Caen, the Abbaye-aux-Hommes has nine bays, the same as Gloucester Cathedral, and the Nuns' church has one more. The nave of Durham Cathedral is comparatively short, only eight bays, but that is partly accounted for by the rapid fall in the ground at the west end, which rendered further extension in that direction impossible (see Figs. 177 and 182). The eastern arms are short in comparison with the western, Eastern but they are longer than in contemporary churches abroad. From arm> the first the tendency to lengthen this part of the church, so marked in later EngUsh examples, is evident. Norwich Cathedral has four bays between the crossing and the apse, and that was the original number at S. Albans and Ely. Gloucester and Chichester only had three bays ; but, then, their naves are shorter. Durham has four bays, a number that makes the choir more proportionate to the nave. S. Etienne, Caen, had only two bays to the east, apart from the apse ; and that is the number at S. Georges de Boscherville, S. Sernin, Toulouse, etc., notwithstanding the great length of nave of the latter. In any comparison of the propor tionate number of bays in the eastern and western arms, it should be remembered that in Eomanesque churches the seats for the choir were placed partiaUy or entirely in the nave. In Norwich and S. Albans Cathedrals and Westminster Abbey they still occupy three bays of it. East of the crossing there was little except the high altar, and the eastern arm could consequently be short. At Norwich, Gloucester, and Worcester it finished with a chevet; and this was also apparently the plan at Durham. Canterbury Cathedral, as built by Lanfranc, ended in one apse ; Peterboro' and, probably, Chester in three apses. The Nuns' church at Eomsey has a square east end; and this traditional Anglo-Saxon ending was also foUowed at Hereford and Eochester. At S. Etienne, Caen, the original choir ended with an apse, but the aisles finished square. AU English Eomanesque cathedrals are cruciform in plan; Transepts the transepts in most of them having the considerable projection which afterwards became so marked a feature of English work. In Normandy, as a rule, they are equally strongly marked ; the 248 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Internal divisions. arms north, east, and south of the crossing being equal, as at S. Georges de Boscherville. The transepts of Ely and Winchester Cathedrals have aisles on both west and east sides, and the aisle passage is returned along the north and south ends of the latter, and the same was probably also done originaUy in the former. The transepts of Durham Cathedral have no western aisles, whilst those of Norwich and Gloucester Cathedrals have no aisles at all. In the two last, however, a large apse opens directly out of each transept on its east side. This was the customary plan in most Eomanesque churches, each apse forming a chapel with its own altar. The two churches at Caen have, or had, similar apses; and the same may be said of most of the contemporary churches in both Northern and Southern France. In all large churches built by the Normans, whether in England or Normandy, the division of the nave wall into arcade, triforium, and clerestory is universal. But there is by no means equal agreement regarding the height to be allotted to each. In the Cathedrals of Norwich and Winchester (as originally buUt) all three divisions are approximately equal. Ely is much the same, except that its clerestory is somewhat low. In Durham Cathedral and in the naves of Tewkesbury and Gloucester, the arcade storey is by far the highest of the three. In the choir of Gloucester, on the other hand, the arcade is low, the triforium lofty, the proportions being much the same as those of Norwich. The following table gives the approximate heights of the internal divisions in four typical examples, and their total heights : — Norwich Cathedral .... Peterboro' Cathedral . . . Gloucester „ (nave) Durham „ ... Floor to string below triforium. 25 0 32 0 39 6 40 11 Triforium to string below clerestory. ' I" 24 0 23 9 9 6 17 10 Clerestory to apex of vault or wall plate.1 24 02 20 0 18 62 12 11 2 Total height. 73 0 75 9 67 (5 71 8 Norwich and Durham were commenced in the same decade, were both Benedictine cathedrals, are about equal in height; but their heights are very differently divided. These two churches show that there was absolutely no rule of relative 1 In churches now vaulted the apex of the vault is about on a line with the original wall plates. 2 Vaulted. Photo: Author. Pig. 174. — S. Etienne, Caen: Tbifoeium. Photo : Author. Fig. 175. — Ouistbeham Chubch. [To face p. 249. ROMANESQUE IN ENGLAND AND NORMANDY. 249 proportions for the internal divisions of Eomanesque churches. Of the two principal churches at Caen, S. Etienne resembles Peterboro' in its proportions ; whUst La Trinite is more like the nave of Gloucester, the height of the triforium being even less than in the English church. DVRHAn CATHEDRAL. OriE BAY OF" riAVE. .11 II l|l II II II.IJ II II II ¦ II II II H | II II Durham has the finest pro portions of aU. Its lowest storey considerably exceeds in height the two upper ones together, and yet does not swamp them entirely, as happens at Gloucester. Its total height is not great, but it is sufficient ; the verticality of the great piers counteract ing their massiveness and also the bulk of the intermediate columns, and allowing it to tell. The design of the triforium is as varied as its height. In Norwich Cathedral, Waltham Abbey Church, S. Etienne, Caen, etc., each bay consists of a single arched opening, about the same width as the opening below. The most usual design, however, is a pair of openings enclosed under a single arch, as at Winchester (originaUy), Ely, Peterboro', Chichester (Fig. 79), etc. The stonework of the tympana under the en closing arches has often chevrons, triangles, and other patterns roughly chiseUed or axed on its face, as at Christ Church, etc. Sometimes there are three or more openings, as at Jumieges (Normandy) and Malmesbury Abbeys. At La Trinite, Caen, the triforium is not pierced at all. In this church, and in all those with open triforia as well, a passage- SCAL.E OF* FEET. Fig. 176. Triforiumdesign. 250 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Naves of English cathedrals. way round the church is provided in the thickness of the upper walls, immediately below the clerestory windows. The design of the clerestory is nearly always the same. On the inside face is a lofty arched opening (central with the window on the outside face) which is flanked by two small openings, each divided from the middle one by a shaft. The effect is far more dignified than that produced by an upper triforium, which cuts off the clerestory from the main triforium and produces another division, as in the somewhat later French churches of Laon and Noyon. In the naves of Ely, Peterboro', and Durham, EngUsh Eomanesque architecture can be seen in perfection. The naves of Tewkesbury Abbey and Gloucester Cathedral are also fine, but spoilt by later vaults. Very different from the two last DVRMA/A CATHEDRAL SCALE OF FEET. Fig. 177. Durham Cathedralvaults. is their neighbour, Hereford Cathedral. At Tewkesbury and Gloucester the columns are possibly too high, and the capitals and arches exceedingly plain. At Hereford the columns are certainly too low, the capitals very elaborate, and the arches unusually rich. The naves of Ely and Peterboro' are surpris ingly alike. In both churches the shafts that support the ends of the tie beams of the roof start from the floor, and are incorporated in the piers. At Ely, all the piers are similar in bulk, but are different in section; and the alternation of the two forms is one of the reasons — and not the least — why the nave of this church ranks amongst the most striking in England. Durham differs from all other English Eomanesque cathedrals, in that its nave vault is approximately the same date as the rest of the church. The church was commenced in 1093 by Photo: Salmon. Fig. 178. — Winohestee Cathedeal : Teansept. Fig. 179. — Dubham Cathedbal. [To face p. 250. ROMANESQUE IN ENGLAND AND NORMANDY. 251 Bishop CarUeph, who, according to Mr. C C. Hodges,1 " had been an exile in Normandy for three years, and there can be no doubt that his design was inspired by what he had seen in progress in that country." If this is so, the inspiration was merely an incentive to better things, for Durham Cathedral far surpasses all Eomanesque work in Normandy. The great piers that alternate with the cylindrical columns suggest S. Ambrogio, MUan, rather than any church in Northern France. When Bishop Flamard was appointed in 1099, the eastern arm was finished, and considerable progress had been made with the western. That the aisles were akeady vaulted with ribbed vaults seems certain ; although it is doubtful if any other portion of the church was intended to be covered in like manner. In the tran septs, some shafts, which start from the triforium level and run up to the apex of the present vault, suggest that there was no intention of vaulting this part at aU. The alternation of large and small supports is not positive proof that vaulting was to be carried throughout. In S. Zeno, Verona, and in many early churches in Germany, there is similar alternation, but no suggestion of any vaults to the naves. Whether Carileph wished to vault his nave or not, his successor, Bishop Flamard, according to Mr. Hodges, had no such desire. He proposed a flat ceiUng, broken by strong transverse arches over the great piers, as in S. Miniato, Florence, S. Zeno, Verona, etc. The existing transverse arches, wliich are pointed, may possibly be his.2 If so, the reason why they are nearly the width of the three shafts below is plain,3 and so are the pecuUarities in the existing vaulting to be next described. At Flamard's death, in 1128, the present nave vault was imme diately commenced, and to the credit of his successor must be placed the buUding of one of the earliest ribbed vaults of wide span with pointed arches in Europe. Mr BUson has proved conclusively that this dates from 1128-1133. The most curious point about the Durham nave vault is the awkward way in which the diagonal ribs start from corbel heads, and not from the side vaulting shafts, as might have been expected. If the transverse arches were built before the vaults were commenced, this is 1 The Builder, Cathedral Series. 2 The four great arches at the crossing, which support the central tower, and the transverse arches of the transept vaults, are semicircular, and probably the work of Carileph. 5 In Autun Cathedral, Burgundy (see Fig. 161), the transverse arches under the barrel vault of the nave are as wide as the pilasters and side shafts below. 252 A HISTORY OF ARCHITECTURAL DEVELOPMENT. West fronts. accounted for. There is a head each side of the capitals of the main piers, and a pair of heads under one abacus over each column, from which the two diagonal ribs which meet here spring. There is no transverse arch above this point. The clerestory windows are not exactly central with the vault compartments. A fact which gives some colour to the theory that Carileph intended to vault his nave, is that if the vault had been a sexpartite one — a form commonly adopted in Normandy in early work — and if the diagonals had started from above the side shafts of the main piers, and had been bisected by an intermediate transverse arch above each column, the windows would have centred exactly. Against this, it may be pointed out that each bay of vaulting between the great piers (especiaUy the second and third bays west of the crossing) are much longer than their width, and the span of the diagonal ribs would consequently have been excessive. The vault of the choir is undoubtedly of later date than the substructure. The fifth bay, which occupies the space of the old apse, was built when the chapel of the nine altars was added in the middle of the thirteenth cen tury, and its vault is of this date. The vaults of the other four bays are a trifle earlier, but they probably also belong to the same century. The west fronts of Eoman esque cathedrals in England have suffered as severely as their eastern arms. Durham Cathedral retains its two towers, but the Galilee porch, added towards the end of the twelfth century, has oblite rated its original entrances. The addition is no drawback ; on the contrary, its lowness gives scale to the front and to WE5T E/1D TEWKE5BVRY ABBEY CH. EXISTING WALLS BLACKED. SCALE OF Fig. 180. the towers which rise behind it, and its foundations carried down the slope of the hUl help to Tewkesbury Abbey has the finest Eomanesque portal give height Photo: Bedford. Fig. 181. — Tewkesbuey Abbey Chuech. photo : J. Valentine & Co. Fig. 182.— Dubham Cathedeal. [To face p. 252. ROMANESQUE IN ENGLAND AND NORMANDY. 253 in England. On each side are six attached shafts, and the remains of a seventh. These are carried up to support a semicircular arch of many orders which rises higher even than the transverse arches inside at the crossing. Below the arch is now a wall, in which is inserted a window of late date,1 but it is by no means improbable that originally the wall was farther back, and that the archway was but the entrance to a recessed porch. The responds inside at the end of the nave arcades are unusually deep, so there would have been room for an outer porch. Lincoln Cathedral has three western porches, and although the central one has been raised the original design is clear. The porches at Lincoln, and the Tewkes bury one, if it may be admitted as such, are the forerunners of the fine later example at Peterborough. 1 The window is dated 1686, but, except that its detail is curious, it is a very fair copy of fifteenth-century work. CHAPTEE XVI. THE CATHEDRALS OF NORTHERN FRANCE, 1 120-1500. Intro- The revival in Church buUding in Central and Northern France commenced about the end of the first quarter of the twelfth century, at the moment when the great outburst in England initiated by the Norman kings was on the wane. The country immediately surrounding Paris and to the North and East of it — the only part which then admitted cUrectly the French king's rule — was at that time slowly emerging from a period of great depression, which was not finaUy dissipated until the accession of Louis le Jeune in 1137. The commencement of this king's reign marks a new era in France — political, social and architectural. The power of the French king rapidly increased ; the strength of the semi-independent rulers of Flanders, Aquitaine, Anjou, etc., slowly but sensibly diminished. The land became peaceful, and peace brought prosperity. The farmers could tUl their fields without a constant fear, almost a certainty, that bands of marauders would render their labour vain. The trade of the dwellers in towns advanced rapidly, and these men soon became strong enough to assert their rights, which had been in abeyance for so long, and to claim privileges of charter, which they had not previously enjoyed. Moreover, the people both in town and country began to interest themselves in matters of religion to a far greater extent and in a more thoughtful manner than they had done before. The authority of the bishop, the people's representative, became more widely acknowledged, in fact the bishops regained to a great extent, if not entirely, the power they had possessed in the days before Charlemagne and before the ascendancy of the Monasteries. The monastic orders were stUl strong, in fact were multiplying rapidly in number, but that the Benedictine order was no longer sole and supreme, was in itseff a source of weakness. It is not too much to say that the revolt of S. Bernard at the end of the eleventh century marks the THE CATHEDRALS OF NORTHERN FRANCE. 255 commencement of the gradual decline of monkish rule. This, however, was not so much the cause of that decline as the growing intelligence of the people ; although dissension and secession are weakening to any body. UntU then, the attitude of most laymen towards the Church was either one of absolute dependence on monkish guidance, or else of opposition to that guidance. Before the middle of the tweffth century was reached both feelings had undergone considerable modification. On the one hand, religion was becoming more and more a matter understanded of the people ; and on the other, the sagacity of the monks had shown them that the best way to disarm opposition was to work in harmony with others for the advancement and good of the Church. Thus, in the building of the Abbey Church of S. Denis, near Paris, in 1140, the people banded themselves together to raise the money required, acting in concert with the Abbe Suger, the man who, amongst his contemporaries, did more than any other to advance church architecture in France in the twelfth century. The fervour with which aU men threw themselves into the task of buUding the great cathedrals of Northern France is a proof both of the prosperity of the country and of the better under standing which existed between aU classes of the community. A large percentage of the revenues of sees were devoted to the work ; the canons helped with their share ; the nobles and rich burgesses gave large sums ; and the people contributed what they could, sometimes their chattels, sometimes their money, or else their time and strength to transport the materials required. M. Enlart says that the columns for the church of S. Denis were dragged there from the quarries of Pontoise (some fourteen miles distant) by the faithful themselves harnessed to the carts.1 At Laon oxen were necessary to drag the stones up the steep hUl to the site, and the services rendered by these animals are commemorated in their stone effigies which to this day look down from the niches in the towers of the cathedral. V Large sums were also realized by the display of relics, which'were carried about the country to stimulate the generosity of the faithful. The requests for funds were, in some cases, by no means confined solely to the diocese in which the cathedral was being built. M. Anthyme Saint Paul 2 recounts how in the rebuilding of Laon Cathedral in the twelfth century, begging expeditions were made one year throughout the whole of 1 " Manuel d'Archeologie Francaise." 2 " Histoire monumentale de la France." 256 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Northern France, and in the next that they penetrated even as far as England. Archi- Qne fact jn particular illustrates the altered conditions under tects. r which the great French cathedrals of the twelfth and thirteenth centuries were built. The artists responsible for their design were in most cases laymen. UntU the middle of the tweUth century the monks themselves had been the architects. They had designed, superintended the erection of, and in most cases found the money for the large churches. The monasteries were at that time the centre of intellectual, artistic life. In the schools attached to the larger and more important ones were trained priests and laymen alike. The men thus trained either worked on the church and buildings of the monastery in which they had learned their craft, or, i£ found worthy, were sent by the monks to superintend building operations for them in other countries, sometimes in far distant lands. Thus ViUard de Honnecourt went to Hungary for this purpose. Gradually the teaching of apprentices in purely technical matters passed from the monks to the lay master masons, as it was natural that when the bishops, chapters, and people took the buUding of cathedrals into their own hands they should prefer laymen as their head workers. Many of these had received in the monastic schools the best education the times could provide, and were as weU read as the majority of clerics and nobles. They probably took good care that the general education of their apprentices was as sound as their own. Some of these master masons were attached to cathedrals ; some to towns ; whilst others took service with high nobles or with the king.1 That they were skUled craftsmen is undoubted, capable of executing themselves all work required ; they had served their apprenticeship, and that in the middle ages meant that they had mastered the technicaUties of their trade. But it does not follow that when they reached the proud position of master mason they continued to work with their hands as they had done when younger. They had higher work to do. They 1 M. Enlart, in his Manuel, gives amongst his list of Gothic churches in Prance, the names of many of the men who in turn superintended the work in different cathedrals. Some are commemorated in the French fashion, which is so excellent, in the names of streets. Thus at Amiens, facing the south transept of the cathedral, is the Bue Bobert de Luzarches, and at Beims, on the north side of the cathedral, the Bue Bobert de Coucy, who, however, can hardly have been the first master mason of the cathedral, as the rebuilding commenced about 1211, and he died exactly a hundred years later. THE CATHEDRALS OF NORTHERN FRANCE. 257 dictated the plan and general ordinance — the first essentials without which no buUding can be carried out satisfactorUy — drawing them on parchment or paper sufficiently well to make their meaning clear to the workmen under them. The modern architectural draughtsman may smile a superior smile at the execution of some of these drawings, but they served their purpose. That so few of them have been preserved is owing to the fact that they were regarded merely as a means to an end, as all architectural drawings should be, and when the end was accompUshed, their utility ceased.1 The full size detaUs of mouldings were drawn on the spot, probably on boards, — much in the same way as the full-size heads of traceried windows, for instance, are set out now — either entirely by the master himseU or else were merely corrected by him. Whether the men who did work of this description should be termed architects, or masons, or masters of the work, is an academic question which hardly requires discussion, except that the term "mason" is confusing, inasmuch as it conveys a totaUy wrong idea of the position these men occupied. The laymen who, in France in the thirteenth century, took the place of the monks of the previous centuries were men of substance, held in high repute by their patrons and townsfolk, and artists in the true sense of the word. De Honne- court's sketch-book shows that he was interested in more than the plan and fabric of a buUding, and that in his travels he sketched and studied architectural accessories as weU. The post of master mason to a town or to a cathedral was a high and responsible one, and descended in some cases from father to son, as at Amiens Cathedral, where, after B. de Luzarches, Benaut de Cormont succeeded his father, Thomas ; or, as at Strassburg, where Erwin von Steinbach's two sons continued the work he had commenced. The spirit of the age was such that there was no fear of the workers in different crafts being out of harmony. The mason had confidence that the painter would not want to apply his colour and gilding on the wrong mouldings, and that the glass stainer would not try and ignore the design of the window. By the beginning of the thirteenth century there was absolute sympathy between all branches ; and to this sympathy is largely owing the beauty and completeness of the mediaeval cathedral. 1 ViUard de Honnecourt's sketch-book has already been mentioned. M. Enlart mentions, amongst others, twenty-two drawings at Strassburg. In Spain are preserved many others, although most of these are later in date. VOL. II. S 258 A HISTORY OF ARCHITECTURAL DEVELOPMENT. To the infusion, so to speak, of secular blood in the twelfth century, is due in a great measure the enormous strides made in France in architectural construction and design between 1150 and 1220. The monk-designer was by no means a recluse, Uving a life of seclusion within the cloister walls; but his training had saturated him with traditional methods which he found difficult to discard. This, as already stated, is especially noticeable in Burgundy, the stronghold of the monastic orders in Western Europe. The lay-designer, on the other hand, although he might have served his time in the same school, was outside monastic life, and mixed freely with aU sorts and conditions of men. The result was inevitable. GraduaUy old ideals gave place to new. The traditional methods of ornamentation — methods based on old Classic or Byzantine design — were sup planted by fresh ones. Tradition was treated reverently — as it must be if good art is to result — but nature and not tradition became the governing factor for ornament and figure sculpture. Owing to this the differences between the carvings executed at the beginning of the twelfth century and those of fifty years later, are even greater than the structural changes which took place during the same period. There was a slack tide towards the middle of the century when ebb and flow conflicted, but this did not last long. By 1160 throughout the greater part of France proper and Normandy the transition was complete, and for the next 300 years or more there was no looking back. Priority There can be no doubt, and it is best to admit this quite frankly, Fran e ^at by the end of the tweUth century French architectural art was considerably in advance of our own, as well as of that of all other countries in Europe. The work in England and Normandy had been done a hundred years before, and with us the necessity for new cathedrals no longer existed. When the awakening came in Central France, the magnificent monuments of Eomanesque art we possessed helped at first more to retard than to advance architectural progress in our country. The fact that the workers in the Isle de France had lagged behind in the eleventh century was no drawback to the builders there of the tweUth. They had nearly everything before them and little behind them. Their existing traditions were hardly their own, but belonged rightly to the surrounding districts of Burgundy, Normandy, and the South. In consequence they clung lightly, and were easily Photo: Author. Fig. 183. — S. Piebee, Lisieux: Nave. Photo: Author. Pig. 184. — Noyon Cathedbal : Nave. [To face p. 258. THE CATHEDRALS OF NORTHERN FRANCE. 259 thrown off. The result was a progress unparalleled for rapidity in architectural history. Most French writers claim that the birth of Gothic — and by The rise Gothic is understood pointed arches, ribbed vaulting, and flying °^?thio buttresses, combined with a new system of mouldings and a fresh teoturc. feeling in carvings — took place in the Isle de France, and from there spread outside. To some extent they are right, but not entirely. Properly speaking, Gothic art had no birth. What is called the birth of Gothic was but the coming of age of Eomanesque; and that the celebration of this majority took place solely in one particular part of France is .open to question. There is so Uttle difference in date between the early examples scattered about the country that it is permissible to conclude that the movement was a far wider one than the above contention allows. Otherwise it could not have risen to maturity so soon over so large an area. It is unnecessary to do more than mention by name the Early churches buUt between 1120 and 1140 in which the germ of tuildlligs. Gothic is apparent ; such as S. Etienne, Beauvais ; the narthex of Vezelay Abbey Church (Burgundy) ; the Cathedral of Evreux (Normandy) rebuUt, according to M. Enlart, immediately after a fire in 1119; and not much need be said about those buUt in great part between 1140 and 1160 in which development is more advanced, in some cases almost complete. The best known amongst these latter are S. Denis, near Paris (1140-1144), which, though it retains the semicircular arch over its lower windows, and traces of Eomanesque feeUng in its carving, is a Gothic church; and the nave of Lisieux Cathedral (Normandy), Gothic of pure early type. 1141-1182 are the dates given to the latter by M. Enlart. If the first is the correct one, then priority for Normandy over the Isle de France could easUy be claimed, as the work is far more advanced than any elsewhere of similar date. The probabiUty, however, is that little was done before about 1160, but even taking 1160-1180 as the date it proves that architectural design was quite as far advanced in Normandy at the beginning of the second half of the twelfth century as in any other part of France. Other early Gothic churches are Angers Cathedral (1150-1156, see Fig. 47), although it is not the Gothic of the North, inasmuch as it has no aisles, and consequently no triforia or flying buttresses ; Sens Cathedral, Burgundy, commenced 1140, doubly interesting to Englishmen, 260 A HISTORY OF ARCHITECTURAL DEVELOPMENT. as William, the master mason of Canterbury, came from there; and Noyon Cathedral (c. 1150), with its apsidal ends to the transepts, the first church in France to show evidence of a strong desire for Ughtness in all parts. The plan of Noyon was no doubt inspired to some extent by the fine church at Tournai, Belgium (see p. 202), the transepts of which are similar. The piers are alternately large and smaU, the arches of the arcades and triforia are pointed, but elsewhere pointed and semicircular arches are used indiscriminately. The triforia are large open gaUeries, as at Laon and Paris, with arcading in the wall above, as in the former church. Pairs of semicircular headed lights to each bay form the clerestory. The church is now vaulted with quadripartite vaults added after a fire in 1293, but the original vaulting was sexpartite. On the whole the design inside is too small in scale, has not the breadth of later examples, and the desire for lightness is carried too far.1 The south transept of Soissons Cathedral (c. 1170) also ends with an apse, and is built on the same scale as extends through out the whole of Noyon Cathedral. It is, however, far superior in plan to the transepts of Noyon, as the apse is surrounded by an ambulatory aisle — the Noyon transepts have no aisles — out of the east side of which opens an unusually large, chapel. The rest of Soissons Cathedral was luckily carried out on a far bigger scale than the transept, and presents one of the finest examples in France of simple, robust buUding, spoilt now, alas ! by staring mortar joints. Nave, Le The nave of Le Mans Cathedral is a striking example of how ^hedral* a Romanesque unvaulted nave of ten bays, with arcades of semi circular arches supported by columns, was converted about 1153 into a Gothic nave vaulted throughout with rib-vaulting, without taking down the nave arches. The arches were originally of two orders, unmoulded, the outer order having a projection of only a few inches.2 The lower order of each bay was removed, and 1 The most interesting portion of the church to my mind is the choir, with its fittings, which date from Louis XIV.'s time. The plan of these is very dignified. The high altar stands under the crossing, not quite in the middle, but slightly in front of the eastern piers. The stalls of the same peried start one bay back, and extend along two bays each side and in front of the apse. An iron grille of eighteenth-century workmanship surrounds the choir, and there are also many fine screens of the same date in front of the chapels lining the nave. - The easternmost bay on both sides remains unaltered, as far as the arches are concerned, so the original design is easily seen, Photo: Author. Fig. 185.— Soissons Cathedbal: South Teansept. Photo: Author. Fig. 186 — Le Mans Cathedeal : Nave. [To face p. 260. THE CATHEDRALS OF NORTHERN FRANCE. 261 underneath the upper one was built a pointed arch (one half of which is concentric with the original arch), which, being narrower than the other, allowed room for the great clustered piers, the outer members of which are carried up to the vaulting to support the massive transverse arches and the diagonal and wall ribs of the vault. As an example of adaptation it is a monument to the clever ingenuity of the mediaeval builders (for plan see Fig. 98). The above-mentioned churches do not, with the exception of The great S. Pierre, Lisieux, display fully those characteristics of early g^onea French Gothic which are so marked in the group next to be considered. The alternation of large and small piers at Le Mans, and Noyon, for instance, belongs to an earUer school. Between 1160 and 1200 the plain cyUndrical column — sturdy, but never heavy, as in our Eomanesque work — was general; and this was often employed even when it had to support as many as five independent vaulting-shafts, as at Laon,1 Lisieux, and Paris, in addition to the arches of the arcade. This afterwards gave way to the column with attached shafts, already referred to (see Fig. 19), as at Chartres, Eeims, etc., but the proportions and appearance of the columns were very little changed by the addition of the shafts. A point worth remembering in connection with French Gothic is that very few changes of any moment were introduced between 1200 and 1350, except in window design. The waUs, it is true, were made thinner; windows were divided into a greater number of Ughts — a window of two lights is the general rule in early work ; the main vaulting-shafts were carried down to the floor, as in Eomanesque churches, and once again formed part of the piers; but very Uttle alteration was made in the mouldings untU towards the end of the fourteenth century, and the vaulting, as already stated in a previous chapter, retained its simple quadripartite form for a couple of centuries or more. There are differences in the carving of capitals, etc., but they are slight as compared with those which took place in EngUsh work during the corresponding period. In resting content with what he had done the Frenchman showed not only sound common sense, but good taste as well. He knew when he had got a good thing and, in the main, he stuck to it. He modified here, lightened there — after 1250 in his deUght in technical skill he often went too far in this respect — but until France went to 1 Some of the columns at Laon have detached shafts in front of them, which help to carry the vaulting-shafts, but not all. 262 A HISTORY OF ARCHITECTURAL DEVELOPMENT. pieces after the battle of Crecy, in 1347, he kept his art free from trivial novelties. Owing to this in a great measure there is a completeness and uniformity about French cathedrals which seldom exist in those of other countries. The French builder of course had not to contend with the disadvantages his EngUsh brother laboured under, at least not to the same extent, of having to enlarge churches built one or more centuries before; but apart from this, a wholesome conservatism prevented the glaring differences which are sometimes so marked in England.1 When fire played havoc with a new church, as it frequently did, its ravages were repaired without introducing marked changes in design. Thus, at Amiens Cathedral, there is little difference between the original work of c. 1220, and the reconstruction after the fire in 1257. Window tracery is practically the only detail in which changes of any importance were introduced. Even in this, when complete development was once reached, which may be said to date from the buUding of La Sainte Chapelle, Paris (c. 1240), few modifications followed untU after the end of the fourteenth century, when the Hundred Years' War had sapped the vigour of French art. Best The reigns of Philip Augustus (1180-1223) and Louis IX., French.0 known as Saint Louis (1226-1270), form the glorious period of art- church building in Northern France. During these two reigns most of the famous cathedrals were in course of erection. To Philip's reign belong mainly Laon Cathedral, commenced 1160 ; Notre Dame, Paris, commenced 1163 : the work in both churches continuing to the middle of the thirteenth century; Chartres Cathedral, the major rebuilding of which dates from 1194 ; Eouen Cathedral, chiefly 1202-1220 ; Troyes, with its fine choir of the early thirteenth century; Mantes Cathedral, not far from Paris, and evidently inspired by the capital ; and Eeims Cathedral, rebuilt after a fire in 1211. The great achievement of Louis' reign is Amiens Cathedral. Commenced in 1220, and completed in 1257, the greater part was burnt in the following year, the rebuilding lasting until 1288. Other contemporary famous cathedrals are Bourges, which belongs chiefly to the middle of the thirteenth century, although some 1 Take S. Albans Cathedral, for instance. It was no doubt impossible to harmonize the later work on the south side of the nave with the earlier work on the north; but the thirteenth and fourteenth-century portions need not have been so different. THE CATHEDRALS OF NORTHERN FRANCE. 263 portions are earlier and others later; Tours, to which the same remark appUes ; the choir of Le Mans, 1217-1254, the chevet of which is externally the most striking in France, partly from its forest of forked buttresses, flying buttresses, and pinnacles, and partly from its fine position at the end of a large market-square, above which it towers ; and Beauvais, that memento of a soaring ambition which came to grief. The cathedral was commenced soon after Amiens, with the notion of surpassing it in all respects. The choir alone was finished in 1272, but twelve years later the vaults fell in and wrecked the church. The work of reconstruction was not accompUshed untU 1347, the piers being doubled and other portions strengthened, and later the transepts were added.1 The time had gone by for great efforts in church building, and a nave was never attempted. The added piers iu the choir must, on the whole, be counted a gain; as the great width of most of the original bays and the slenderness of the piers must have caused an appearance of weakness, in addition to being weak. If the builders, after the catastrophe, had continued the outer member of the original arch so that it enclosed two narrow arches (as at Boxgrove Priory, Sussex) the effect would have been exceUent. The coUapse, in the hands of an able man, might have proved a blessing, but the builders seem to have lost their heads under the magnitude of the disaster. The new intermediate piers are the same depth, from north to south, as the old ones, but they are considerably narrower. Corresponding with the above cathedrals are others in the Cathe- neighbouring provinces, many of which are little inferior to d.1"als °ut- the great examples mentioned. In Normandy are Coutances Isle de (c. 1251-1274), famous for its side chapels, divided from one Franoe- another by walls pierced above with traceried openings similar to the windows alongside them ; Sees (c. 1270), the distinguishing mark of which is the slender detached shaft which starts from the floor, in front of each cyUndrical column, and reaches to the vault, passing in front of a beautiful Uttle cusped niche. The effect of these shafts is most striking, and very unusual, if not unique. On the boundary of Normandy and Brittany is Dol 1 The transept vaults are dated 1577 and 1578, and 1575 is the date on some of the bays of the choir vaulting. The vaulting throughout the church is sex partite, owing to the added piers. The vaulting in the aisles is especially curious, as each intermediate transverse arch consists really of a ridge rib supported by an arch, the stonework between being pierced. 264 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Cathedral, with a fine choir, 1231-1265. In Burgundy, the work proceeded at Sens Cathedral towards the end of Louis' reign; and at Auxerre the choir was built 1215-1234, the nave belonging to the next century and the west front being left unfinished. At Dijon is one of the most remarkable churches of second rank of this period, N6tre Dame de Dijon, which was built about 1240. It appears earlier, as it has all the characteristics of the work of Philip's reign, but Burgundy was conservative and to some extent cut off from the movement taking place farther north. It is a model of early thirteenth- century French methods and, although the work is somewhat frigid inside, few larger churches can beat its fine proportions. Vezelay choir is much earUer (c. 1180), but contrast with the Eomanesque nave makes it look later than the date given.1 Sites of One fault sometimes found with French cathedrals is that cathe- ^gy seidom, if eVer, stand clear. They are surrounded by streets, hemmed in by houses; and one sometimes longs for the cool green sward and the quiet of an EngUsh cathedral close. At the same time their positions emphasize the connection between the Church and the people. The cathedral was their work; they wished it in their midst, where they could see it without let or hindrance, enter it without encroaching on property not their own ; and so, as at Amiens especially, it towers above all surrounding buildings, and from a distance stands out a great black mass, a land mark to the dwellers in the country round, a beacon to town folk when returning home. Owing to their surroundings it is difficult, as a rule, to obtain a good view of any cathedral except from a distance. Most of the east ends are enclosed, and visible only from private gardens — often those of the bishop or archbishop — and the approaches to the west fronts, or to the fine transept entrances, are often poor in the extreme. The Gothic buUder, unfortunately, did not appreciate what the Eoman architect knew instinctively; that the approaches to a building are quite as important as the design of the buUding itself. In some of the Eomanesque and slightly later monastic churches and cathedrals in France, there was, no doubt, originally something of the nature of an atrium or forecourt to the west. At Chartres, Sens, Paris, Caen (in front of S. Etienne), and elsewhere are large open spaces opposite the main entrances ; but in the majority of cases, either 1 Most of the dates are taken from M. Enlart's " Manuel d'Archeologie Francaise." THE CATHEDRALS OF NORTHERN FRANCE. 265 land was too valuable to be devoted to such a purpose, or else the builders were indifferent to everything except the church itself. The east end of Le Mans Cathedral stands free, as already stated ; and at Bourges a magnificent view is obtainable of the cathedral, with its long, unbroken roof-line, from the formally planned gardens to the south-east. This view is one of the finest archi tectural ones in France. At Auxerre is another fine view of an east end, although a more distant one as it is from the opposite bank of a river. The cathedral, raised on its double crypt, towers above the old bishop's palace, the grounds of which slope down to the river-side. The lower portions of the church are hidden, but the clerestory windows and buttresses stand out well. An analysis of the plans of leading French cathedrals shows proportions very different from those which, as will be shown later, appertain to English ones. French cathedrals, on the whole, are considerably wider ; their height is far greater ; and although there is not such great difference in the actual length of the examples in the two countries, French cathedrals, owing to their width and height, appear much shorter. Their proportionate length to width, including both side and end chapels, is approxi mately as follows, the width being taken as the unit : — Proportions. To width of western To width of eastern arm. Paris 2'8 2'8 Chartres .... 377 2'8 Beims . . 4-55 303 Bourges . . . 2'31 231 Amiens . . . 2'95 2-9 Their superficial area and cubic contents are far in excess of ours. The average width of the churches named in the Appendix table, exclusive of side chapels, is 111 feet, as against 82 in the English cathedrals mentioned. The real differences between the transverse dimensions of the churches of the two countries lie in the fact that so many of the great churches in France have double aisles, sometimes to both nave and choir, sometimes to choir alone, whilst beyond the aisles are frequently chapels. Notre Dame, Paris, has a total width of 149 feet, Bourges Cathedral 16 feet more, in both cases the side chapels being included (see Fig. 106). The best-known churches Double aisles. 266 A HISTORY OF ARCHITECTURAL DEVELOPMENT. with double aisles all round, including the chevet, are the two just mentioned. Troyes has double aisles to nave and choir, but only a single aisle divides the chapels of the chevet from the apsidal termination of the choir. The east ends of Amiens and Eeims are similar to Troyes, but the naves have only single aisles, although at Amiens there are chapels beyond, which is not the case at Eeims. Chartres, Le Mans, and Coutances, have Ukewise only single aisles at the west end, but beyond the transepts they widen out into double aisles which are continued all round the east end REI/AS CATHEDRAL d t:-.'#5-.-;:*;-;'*: ¦"' : ."-**. '\ ' :' >•---.+.- - -yfc--;.^-- O SO IOO ISO H SCALE OF FEET. Fig. 187. with radiating chapels in addition. The chapels in Le Mans are continued down the sides as well, making thirteen in aU (see Fig. 98).1 Internalresults of plan. Interiors. The main results of the French plan of doubUng the aisles and adding chapels are, in the first place, that the effects obtained by looking across the churches from north to south, either straight or diagonally, are, as a rule, far finer than the vistas from end to end; and, in the second, the great spaciousness of the central part surrounding the crossing. The latter is partly due to the fact that the transepts in nearly all the principal examples have an aisle on both sides. The transepts themselves have very slight projection, far less than in England, but this so far from 1 The vaulting of the choir of Le Mans is much higher than that of the nave ; and by hipping back the nave roof, it has been possible to insert a window in the west wall of the orossing above the nave vault, which produces a particularly good effect inside from the east. THE CATHEDRALS OF NORTHERN FRANCE. 267 detracting from the effect mentioned, in reality adds to it. Amiens Cathedral is the most remarkable in this respect. The appearance inside is that of a great square of flooring broken in places by a few columns ; the nave — which is unusually short by comparison with the rest of the buUding — and the chevet appear ing but as adjuncts. The effect of spaciousness is no doubt helped by the great height of the church, but it is mainly due to the plan. The chevet, as before mentioned, forms the ending to most of chevet. the great cathedrals of France, the exceptions being chiefly in the South. Although few wiU deny that externally the chevet sur passes in effect the result obtained by the two characteristic AAVIE7S5 CATHEDRAL, ^+ -.•K ;?. >- --?---?- ii iL-;cXXXiXX\iXJXi >'OXiX:XXi"'c' i X • X' i X ¦ X' i X I X iX&tiX i '¦*' i - i X^JSX! iLJm\l'' \''' ^''' **•/ \ './ \*/ \' / vi^ - -r- * vV.' \ !/ N i/' "^/ M' '• 'V- V-'T^I ' » o 50 100 150 I I I 1 1 I I I SCALE OF FEET. rt Fig. 188. EngUsh methods described on p. 151, it is very doubtful if the gain outside is sufficient compensation for the cramping effect it often produces inside. In order that all supports in the chevet shaU radiate from one centre, the bays round the end of the choir are too numerous (especially in those churches in which double aisles surround the apses) and too narrow; the piers, as a rule, too slender, and the arches, in many cases, disagreeably stUted in order that they shaU be the same height as the other arches of the choir. In most churches with only a single aisle behind the choir apse, such as Eeims and Eouen Cathedrals, there are only five bays. At Amiens it is difficult to say whether the apse should be termed one of five or seven bays, as its centre is on a line with the middle of a bay, and not on a Une with a pier. This modifica tion of the usual plan is foUowed at Chartres and Le Mans, and 268 A HISTORY OF ARCHITECTURAL DEVELOPMENT. was all the more necessary, as the apse of each is surrounded by double aisles. In the Cathedrals of Bourges and Paris there are again only five bays to the apse of each, although both churches have double eastern ambulatories. At Bourges the spaces between the columns dividing the ambulatories from each other are un usually wide ; at Paris a more ingenious plan is followed. The columns between the aisles are doubled in number, there being nine, to four round the apse. Some of the columns of the back row come central (on the radiating lines) with the columns in front, and the others are placed in between. The vaulting of the aisles is naturally affected by this duplication of supports, and the severies are all triangular in plan instead of being quadrangular. In nearly all the large churches built after 1250 — S. Ouen, Eouen, is an important exception — the apses are generally seven-sided ; and as, in consequence, the bays are very narrow, the piers very thin, and the arches much stilted, the east ends of these churches are their least satisfactory part. In some of the latest examples, such as Orleans Cathedral, the stilting is not so strongly apparent because the piers throughout are without capitals, and the spring- ing-line consequently is not emphasized. Apse The piers of apses are often totaUy different in section from piers. ^e g-^g pjers 0f choirs. This is especially the case in the early chevets. In these the supports frequently consist of two columns one behind the other, often attached to each other; with sometimes, as at Bayeux Cathedral, a slender detached shaft nestling between them on each side. The object in changing the plan of the piers was to give as much width as possible to the openings, and at the same time not diminish unduly the strength of the supports (see Fig. 19). Choirs of The success of the choirs of Chartres and Paris Cathedrals and Paris. *s largely due to the fact that in these two churches the faults in apse planning, apparent in others, have been avoided. Chartres, it is true, has more bays at the east end than some churches, but its choir is the widest in Northern France, and so the bays are of fair width (Fig. 194). Moreover, all the arches of the main arcades throughout the church are stilted, and the little extra stUting of the apse arches is hardly apparent. At Paris, the bays of the apse are only a trifle narrower than those at the sides, and all the piers surrounding the choir are of the same size. These facts, coupled with the fine perspective effects east of the apse, due to the number and Pig. 189.— Notbe Dame, Pabis. [To face p. 268. THE CATHEDRALS OF NORTHERN FRANCE. 269 spacing of the columns between the ambulatories, help largely in making the Paris choir the finest in France. The most important exception in Northern France to the Square chevet plan is Laon Cathedral. The original choir consisted L"*nend' of six bays (or three double bays, as the vault is sexpartite) and ended with a chevet. About the year 1200 the chevet was removed and four bays added to the east, making the choir ten bays long and the end square.1 The square ending is more Ukely due to English influence than to the preaching of S. Bernard, as Cistercian choirs are invariably short, and the length of the choir of Laon is characteristically English, and unusually long for France. Other English features in the Cathedral are the considerable projection of the transepts, and the presence of the dog-tooth ornament. The latter feature cannot be claimed as an exclusively EngUsh one, as it is present in Le Mans Cathedral and in a few other French churches outside Normandy, but it was not used as an orna ment in France to anything like the same extent as it was in England. With these exceptions Laon Cathedral is as characteristicaUy French as Salisbury Cathedral is English. In the cathedrals of Paris, Laon, Sens, Bourges, etc., which But- are sexpartite-vaulted, the flying buttresses which transmit ^rpartite the thrusts of the intermediate transverse arches only, are made vaults. as strong as those which transmit the thrusts of the main transverse arches and diagonal ribs as weU. M. Corroyer, in his "Architecture Gothique en France," terms this illogical. To a certain extent it is ; but aesthetically there can be no doubt that the arrangement he objects to is right. Laon, Bourges, and Paris would lose considerably in external effect if half their flying buttresses were omitted or made much lighter. Structurally, there must be buttresses; and considering that the function of a flying buttress is to support the high nave wall, quite as much as to transmit thrusts, M. Corroyer's objection seems un- caUed for. In the sexpartite-vaulted church of La Trinite, Angers, the side buttresses are alternately large and small, but then there are no aisles, and consequently no flying buttresses, and the walls are of no great height (see Fig. 163). Laon and Paris are the only two of the great Cathedrals — and Triforia. 1 The later work is easily distinguished from the earlier by the different carving on the oversailing course under the top parapet of the easternmost four bays outside, and inside by the carving on the capitals. 270 A HISTORY OF ARCHITECTURAL DEVELOPMENT. to them may be added Noyon Cathedral, and the south transept of Soissons — which have the fine open triforia, vaulted like the aisles, which are so frequent in the earUer Eomanesque work of England, Normandy, and Italy. At Paris the upper triforium, Fig. 190 (G. Langshaw). present at Laon and Noyon, is wisely omitted, and the extra amount of plain wall space thus obtained is an undoubted gain.1 1 Notre Dame, Paris, as originally designed, had an upper triforium. The lower one was covered by pointed barrel vaults, at right angles to the nave, like those formerly over the aisles in Fountains Abbey Church. In the transepts and in the easternmost bay of the nave, between the main triforium and the clerestory windows, are pierced circles divided by bars. In the transepts these circles are glazed, and over each is a very wide single light. This was the original design of the church before it was altered at the beginning of the thirteenth century. Photo : Author. Fig. 191. — S. Pieeee, Chartees. [To face p. 271. THE CATHEDRALS OF NORTHERN FRANCE. 271 In most of the large French churches of c. 1180-1250, Chartres, Glazed Eeims, Amiens, etc., the triforia are absolutely closed by a wall behind the arcading in front ; and this remained the usual design until the craze for lightness and for glass opened them out in another sense, and occasioned the glazed triforium. The reason for the waU at the back at Chartres, Eeims, etc., was partly that the builders feared for the stability of their churches, as their vaults, in each succeeding example, rose higher and higher; and partly that they felt the advantage of a bUnd storey, to divide the open arcades from the nearly equally open clerestories. The semi-open triforiuni of England, as at Salisbury, Lincoln, etc., is rare in France. To pierce and glaze the wall behind the arcading of what had hitherto been a blind storey, did not mean much actual weakening of a church, as the ashlar mulUons on the outside face are, as a rule, strong and fairly numerous; but the sense of security given by the wall disappeared. In the choirs of the Cathedrals of Metz, Amiens, Beauvais, etc., in the churches of S. Ouen, Eouen, and S. Pierre, Chartres, etc., the triple division of arcade, triforium, and clerestory is retained, but glass practically reaches from the top of the main arcades to the summit of the vaults. Before the triforia could be glazed it was necessary to get rid of the lean-to roofs behind them over the aisles, which would have blocked aU Ught. Lead flats were sometimes substituted for these, or else high-pitched roofs of double slope with a gutter on each side, that against the nave wall being at about the same level as the floor of the triforium. Often each bay of the aisle has its separate high-pitched roof, hipped on all sides, as at S. Ouen, Eouen. These numerous roofs add a certain picturesque- ness to many a late French church, but it is a question if this is not dearly bought by the loss in dignity which results. In all mediaeval work, and not merely in France, the tendency, as development proceeded, was steadUy in the direction of increasing the height of the main arcade and of the clerestory, and of diminishing that of the space between them. In the choir of Le Mans the triforium is omitted entirely, and the clerestory windows start from the top of the arcade. The following table Ulustrates the development, and at the same time emphasizes the loftiness, at all periods, of the ground storey of French cathedrals. In English ones this storey is considerably lower; even after aUowance is made for their less total height. 272 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Name of cathedral. Paris . . . Chartres . . Amiens . Beauvais . Bourges . . . Le Mans, choir Height from floor to string-course under triforium. Feet. 37'050'0 70'0 740700700 Height between string courses above and below triforium. Feet. 330160 22-020-0 190 Height from string course under windows to apex of vault. 40-0 Feet. 40048'0 48'0600360 Bourges Cathe dral. Westminster Abbey has practically the same proportions as Amiens, namely, one half the height to the arcade, and the remaining half, one-third to the triforium and two-thirds to the clerestory ; but no other French example foUows exactly this division. In a few cathedrals built towards the close of Gothic art in France, Flour, in the south of Auvergne, for instance, and in a great many large contemporary town churches, such as S. Aignan, Orleans, S. Pierre, Auxerre, etc., there is considerable space between the arches and the windows above, where the triforiuni generally comes, but it is neither pierced nor arcaded. The expanse of plain wall surface is a great gain. From the first the French buUders had a fondness for plain walling which they never entirely lost, even in their days of huge windows. They realized its advantage as a foil to surrounding moulded work. The space between the top of the main arcade and the string under the triforium is considerable at Laon, Paris, Chartres, Amiens, etc., and is left quite plain. In the nave of Amiens the plain wall surface is especiaUy valuable, as it lends emphasis to the richly-carved string-course above it. In the four cathedrals mentioned, and also in Eeims Cathedral and in other churches, there are no hood moulds over the arcades, such as are almost universal in England — a wise omission, as they are unnecessary in a church faced with ashlar inside, and they take up space which is better left plain. Bourges Cathedral requires special description, as its section is unique. The church has double aisles. Over the main arcade throughout are triforium and clerestory. Arcade, triforium, and clerestory, all of good size, also come between the inner and outer aisles on both sides. The three combined are the same height as the arches alone of the nave, which measure nearly 70 feet — the height of the nave of many an English cathedral. This design has no parallel in France or in any other S.AIC/NA/S, OR.LEAM5, VOL. II. Fia. 192. 274 A HISTORY OF ARCHITECTURAL DEVELOPMENT. country.1 The choir of Beauvais is the nearest approach to it, but in this church the corresponding triforium and clerestory are mean. In Le Mans and Coutances Cathedrals there is the same triple arrangement between the inner and outer aisles surrounding the choirs, but in neither church has the choir itself a triforium. In the nave of Eouen Cathedral the arches between the aisles and the chapels beyond them are as high as the main arcade and the trifo rium combined, but then the nave arches are low. The interior BOVRGES CATHEDRAL SCALE OF io o 150 FEET. I Fig. 193. of Bourges stands alone ; and although it may not equal in beauty the sturdier work of Paris, Chartres, and Eeims, it produces an effect of great height, and appears far taller than Amiens Cathedral which really exceeds it by nearly 20 feet. The outside of the Cathedral also presents some unusual features. There are no transepts whatsoever, a distinct advantage, as the church is less than 400 feet long and is 165 feet wide, and the roof runs 1 In some of the large double-aisled cathedrals in Italy, such as Milan, there is the same gradation of nave and aisles, but not the same three divisions (see Pig. 242). THE CATHEDRALS OF NORTHERN FRANCE. 275 unbroken from east to west.1 Over the chapels of the chevet are steep-pitched stone roofs, similar to those over the apses of the eleventh-century church of S. Nicolas, Caen. At the west end are five magnificent double doorways, affording direct access to the nave and the four aisles. Of the different cathedrals in France, Beims is probably the Stability best buUt. Street eulogizes the construction as perfect, and says churches. that there is not a crack in the building. This is partly because there has been no tampering with it. The aisle walls have not been cut away to form chapels, as at Laon, Paris, Amiens, etc.,2 and the fine aisle windows are still framed, as they should be, by projecting buttresses. The main columns are sturdy but not heavy ; the shafts of the triforium (12 inches in diameter) are a trifle stouter than in other churches ; the buttresses are substantially weighted on top by well-designed canopied niches, each containing a figure; and throughout there is a feeling of weU-balanced strength which is often absent from churches of more massive proportions. The walls of the nave are carried high above the vaulting — the space between the tie-beam of the timber roof and the extrados of the vault below is about 10 feet — and the added weight gives stability. But the main reason why the church has stood so well is the absence of false bearings. There is very Uttle balancing of upper waUs over next to nothing, as is the case at Amiens and Beauvais. At Beauvais false bearings and excessive height brought destruction. At Eeims the builders played no pranks (see Fig. 62). EXTERIOKS. The results of the proportions already detaUed and of the Towers. peculiarities of the planning are shown as unmistakably in the exteriors as in the interiors. The great height of the majority of French cathedrals rendered a central tower, in most cases, an impossibUity. Its accomplishment would have been a tour-de force from which even the builders of the Isle de France shrank, greatly though they deUghted in overcoming structural difficulties. There is one at Laon, but the dimensions of the cathedral as 1 In Southern France transepts are frequently omitted, and it was probably the influence of the South which suggested the plan adopted at Bourges. Bourges is about 150 miles due south of Paris. 2 The old stone seat against the aisle wall still runs round the greater part of the church. 276 A HISTORY OF ARCHITECTURAL DEVELOPMENT. regards both width and height are more modest than in most other examples. In Normandy central towers are common, and nothing can exceed the beauty of the one at Coutances. Else where in France, in Gothic times, they left them severely alone ; substituting for them, in many cases, a jleche of wood and lead at the intersection of the roofs — a poor substitute for the square tower, or tower and spire, of so many an EngUsh cathedral. At Amiens the top of the fleche is 422 feet from the ground, some 20 feet higher than the top of the stone spire which rises above the well-proportioned and comparatively high central tower of SaUsbury. Another result of the height of French churches is that the western towers are often dwarfed by the roof. At Amiens the ridge actually rises higher than the parapets of the towers. Transept That the Frenchmen had no objection to towers is shown by towersand tQe fact tnat seven were commenced at Laon Cathedral, a central entrances. CHARTRES CATHEDRAL. T. TOWER ABOVE SCALE, or 'EET. Fig. 194. tower, two west-end ones, and two to each transept. At Chartres eight were actually intended, there being no central tower, but two additional ones are placed, one on each side, over the outer aisles of the choir, just before the chevet begins. Eouen, like Laon, has, or was intended to have, seven. Eeims has six, two at the west end and two flanking each transept. Transept towers, on a big scale, occur first in the Eomanesque part of Tournai Cathedral, Belgium (c. 1120), where their lower portions are Fig. 195. — Amiens Cathedbal. [To face p. 276. THE CATHEDRALS OF NORTHERN FRANCE. 277 hidden by the aisles of the semicircular transept apses which stand in front. In France the whole tower, as a rule, shows. The reason for these towers in French cathedrals is the natural objection, which aU architects must feel, to the display in elevation of the ugly sectional outline of a church which has a high central part with lower side aisles. The Frenchmen saw no reason why this outline should appear any more at the sides than at the west end. For it must be remembered that transepts in France have an importance almost equal to a west front itself. Their door ways are often as fine as those of the main entrance. At Chartres the porches in front of the transepts, with their wealth of glorious sculpture, are far more important than the western entrances. At Sens, Beauvais, Eouen, Eeims, Abbeville, etc., almost everywhere in fact, the north and south doorways, with their large windows over, are amongst the richest portions of each church.1 The towers of the transepts of Eeims and Eouen Cathedrals are aUke in design. AU are unfinished, being carried no higher than the springing of the gable. In each tower the upper storey is very open, and is pierced on three sides by two very long unglazed Ughts. At Laon the west towers of the north and south transepts and the central lantern are complete, and, owing especially to the position of the cathedral, form a striking group. In towns, however, in which the church occupies a more cramped position than Laon Cathedral does, it is a question whether the multiplication of towers was of any advantage, except, perhaps, from a distance. It is certainly difficult to understand the reason for the two eastern towers of Chartres. At Coutances Cathedral there are two smaU ones in simUar positions, but then they are Uttle more than large pinnacled buttresses, and the tran septs, moreover, have no flanking towers. The builders of Notre Dame, Paris, were the first to abandon towers at the transepts, and their example was followed at Amiens and in subsequent cathedrals. It was probably felt in most cases that they were not necessities, and in addition there was the unlikelihood that they would ever be finished. So many western towers in France were left incomplete until the fifteenth century — some were not finished until the Eenaissance had obtained a footing in the follow ing one; Tours Cathedral is an instance — that the desire of the 1 At Beims the north transept only has doorways. Over the westernmost one of the three is incorporated some interesting carved work from the original cathedral which was burnt down in 1211. 278 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Easterntransepts. West fronts. builders to concentrate their energies on work the completion of which they might possibly see, may be the sole reason for the disappearance of towers from the transepts.1 Eastern transepts, as at Salisbury, Beverley, Lincoln, etc., were impossible in French cathedrals. They would have curtailed too much the apparent length, which already was too short. For the same reason the builders were careful not to give too much projection to the main transepts, which sometimes, as at Paris, are practically flush with the side chapels of the ground storey, although, of course, owing to the sectional outline of the church, they stand clear above. The earliest west front to which the term Gothic can be applied is that of Chartres Cathedral (c. 1145). Here is found the ordinance which, slightly modified from time to time, appertains to all French cathedral facades. Two western towers may be said to be the universal rule ; 2 portals are generally three in number, Bourges Cathedral being an exception with five ; over the central doorway comes a huge rose window; immediately above it, as at Laon (c. 1180) and Eeims (1211-1311), or with windows in between, as at Chartres. At Paris (1208-1235) the rose window is separated from the doorway below by a band of figures which is continued across the towers the whole width of the front. At Amiens (c. 1260), owing to the great height, more division was necessary. Between the three portals and the central rose window and windows alongside it, is a row of windows (see Fig. 208), and above that a band of figures simUar to the one at Paris. Above the rose window at Chartres is a straight band of arcading, the spaces being filled with figures, which forms a parapet. Behind this, set back some feet, rises the gable ter minating the nave. The crowning horizontal line is a charac teristic of all churches of the thirteenth and fourteenth centuries in France in both western facades and transepts. The gable was never allowed to come to the front until towards the fifteenth century, although a portion of it sometimes shows behind the straight line, as at Chartres. At Laon it is hidden entirely at the west end. At N6tre Dame, Paris, the nave roof and gable 1 On the west side of each main transept of Canterbury Cathedral is a small Bomanesque tower, but the pair form the only attempt made in England in this direction, and that not a very important one. 2 In some abbey churches of late date there are no western towers, as at Vondome, or only one, as at S. Biquier, near Abbeville. , V^y -ii'rTrTr;r;ig it % 9 ¦£ a ^ a ¦r-<«™i,».HH#'*|^i- m mm #^'-rI, e ^ g s THE CATHEDRALS OF NORTHERN FRANCE. 279 are set back on a line with the eastern face of the towers, and are practically invisible, allowing the arcading above the band of figures to be pierced, and at the same time giving an importance to the side towers which they would not otherwise possess. This arcading is continued straight across the towers, and projects a trifle in front of them, thus producing a lighter effect than if it had been set back against the waUs, and preventing too strong a contrast between the open centre and semi-closed sides. At Laon the arcading in the centre is a trifle higher than at the sides. At Eeims the crowning arcade becomes a series of niches filled with figures, and over each niche is a canopy with a pediment on top. The result is spikey, and far inferior to the simplicity of the earlier work ; the general effect being by no means enhanced by the struggles, so to speak, of the gable behind to get to the front. At Amiens the designers wisely placed their figures below the rose window and not above it, as the upper part of the centre of the front cannot be seen unless one goes on the housetops. This consists of a gallery, with some pierced panelUng above, behind which rises the gable. Even in so late an example as Orleans Cathedral the straight line is retained, the gable not showing at all. In Tours Cathedral it is more in evidence, but this was not finished before the end of the fifteenth century. From the above it wUl be seen that horizontal lines are far more marked in western facades than in any other part of a church. Vertical Unes are stiU strong — the churches could hardly be Gothic if they were not — but there is not the same insistence on verticality to the exclusion of everything else that is so noticeable in the interiors. At Amiens, it is true, the carved string-course under the triforium marks a stronger horizontal line than is generally found inside Gothic churches, either in England or France, E. de Luzarches and the De Cormonts probably feeling that the great height was sufficient to stand it. In the church of Notre Dame de Dijon, the horizontal lines on the western facade are especially prominent. The entire front is carried up above the roof, and finishes square. The horizontaUty of the design, however, is not so much due to this as to three carved bands of boldly projecting animals, which stand out from the wall, framing in, and dividing from each other, two bands of arcading. These animals are in many respects even more remarkable than the animals and devils that look down from the parapet of Notre Dame, Paris. The great portals of France far exceed in magnificence the Doorways. 280 A HISTORY OF ARCHITECTURAL DEVELOPMENT. doorways of English cathedrals". Great simUarity runs through them all. The detaU varies, and the later examples are richer than the earlier ones, but in the main there are few differences. The doorways themselves are generally double, and are always spanned by lintels. Over the lintels is a tympanum framed in by an arch of many orders, which forms the head of the portal. The jambs of the portal, or porch as it may be called if the doorways are much recessed, the support between the doorways, and each order of the arch, are fiUed with figures under canopies which form a bewUdering array of sculpture of the highest order. The tympanum also has figure sculpture, generally arranged in bands.1 In the sculptured tympana and carved concentric orders of the porches the Gothic builders were only developing the idea introduced in Eomanesque times. There is no difference in principle between early and late designs in this respect. By their multipUcation of the arch orders, however, and by their greater skill as sculptors, the French craftsmen of the thirteenth century advanced far beyond their predecessors. At Chartres their superiority is especiaUy marked in the sculpture and carvings literally covering the projecting porches of the north and south transepts. No other country can boast such Gothic doorways as France, although in Spain an early example at Santiago de Compostella is probably finer than any French Eomanesque one.2 In Spain also and in Germany there are some late doorways of the fifteenth and sixteenth centuries which, if excess of richness were a desideratum, might compare with French ones. Wide west The west fronts of Bourges and Eouen Cathedrals deserve a special word because of their width. In nearly all other churches the towers come at the end of the aisles. Even in Notre Dame, Paris, a church with double aisles, the space between the towers is only the width of the nave. Each of the side towers is as wide as the two side aisles, which, under the tower, have no dividing column and become a single square. At Bourges, however, the towers start from beyond the inner aisles, and project considerably beyond the outer ones ; even beyond the side chapels. 1 At Beims the sculpture of the three tympana was removed, probably in the fifteenth century, and windows inserted. From the inside the effect is not so bad ; outside it is deplorable. About the same time the simplicity of the gables over the three porches was destroyed by the addition of groups of figures, standing on clouds ; and in the case of the central one, surmounted by gimcrack tabernacle work. 2 A full-size cast of this doorway is in the Victoria and Albert Museum, South Kensington. fronts. THE CATHEDRALS OF NORTHERN FRANCE. 281 Hence the possibility of the five doorways, each a double one, already referred to. In Eouen Cathedral, which has only single aisles, although there are chapels beyond, the towers start outside the aisles altogether. A west front of unusual width is the result, which, before it was altered in the fifteenth century and " restored " in the last century, must have been very striking. The main characteristics of the great churches of France have Sum- been outlined in this chapter ; a similar treatment is accorded to rflary- their smaller brethren in England later. French cathedrals, owing to their height, double aisles, side chapels, and the resulting com- pUcated arrangement of buttress, flying buttress, and pinnacle — famiUar to all acquainted with French architecture — produce in some a feeling of unrest which displeases them. In the chevet the culmination is reached. " An arch never sleeps," says the Arab proverb ; and throughout a French cathedral, at the east end especiaUy, one has always the sense of movement ; the feeUng that stone is always grinding against stone ; that the whole, although a triumph of beautifully balanced counteracting forces, cannot possess that stability which a lasting monument should have. It is but fancy. The French cathedrals have lasted as long as ours, and are in as good a state of preservation as ours ; in many cases in a better. EngUsh cathedrals are quieter ; more soothing ; less daring; more peaceful. Flying buttresses are often lacking altogether. Pinnacles are few, and dimensions modest. The EngUsh church is suitable to its close; the French church in harmony with the turmoU that often surrounds it. Each is best in its own place, and no one can truthfully say that the churches of one country are right, and that those of the other are wrong. With the building of La Sainte Chapelle, Paris (1240-1248), Later French Gothic architecture enters on its second stage. But work' there is no radical change to chronicle, save in the tracery of the window heads, and no specific difference that can be stated in terms between the churches of the succeeding hundred years and those of the previous century. But there are differ ences nevertheless. The viriUty of Chartres and Eeims was passing away. The work was already somewhat wirey and hard. The buUders, fortified by experience, were becoming technically bolder. Plain waU surfaces, both outside and inside, were dis appearing as though by magic. The enlarged windows left no room for them, and the glazed triforium had destroyed what before 282 A HISTORY OF ARCHITECTURAL DEVELOPMENT. S. Ouen, Eouen. English influence on four teenth- century churches. had been a strengthening band, both in appearance and in actuaUty, between arcade and clerestory. The piercing and panelling of flying buttresses, which commenced about this time, destroyed further the vigour of the churches outside, although adding to their richness. All structural supports and adjuncts grew smaller. Soon little was left inside except the piers, arches, and vaults ; and outside, the buttresses, flying buttresses, and window-frames. The result was well summed up by the late G. F. Bodley, in a lecture at the Eoyal Academy, in the phrase, " late French cathedrals are all vigour and glass." The most striking church of the middle period of French art is S. Ouen, Eouen, commenced 1318, by l'Abbe Marc d' Argent. The choir and transepts were finished at his death in 1339, but certain portions, the central lantern, for instance, were not completed until the beginning of the sixteenth century. Its internal height is practicaUy the same as that of Westminster Abbey, although the width of its nave is a few feet less. In total width the Eouen church is the larger, as its aisles are wide, as in many other late examples. The detail of Marc d' Argent's portion of the church is very pure, and not unlike the sUghtly earlier work at York Cathedral. The nave is more flamboyant, but shows more restraint than is customary in late work. Many of the fourteenth-century churches in certain parts of France show traces of English influence. This is especially marked in the Southern and Western districts, which at that time were either under the direct rule of the English kings, or in close communication with England. At Figeac, for instance, the clerestory windows in the Church of S. Sauveur are strangely like contemporary ones in this country. The feeling is the same in many of the Brittany churches — at Guingamp, Quimper, Treguier, etc. The size and shape of the windows, the lines of their tracery, and the sections of the mouldings are all more in accordance with EngUsh dimensions, proportions, and forms than with French ones. England had learnt from France towards the end of the twelfth century ; a hundred and fifty years later, when France was in a parlous state, her kings weak, and her lands devastated, England was able to teach in her turn. Her influence did not extend throughout the whole of the country. To the east and north of Paris there is little trace of it ; the churches there are unmistakably descended from the great French cathedrals of the thirteenth century ; but in the west, in the towns mentioned THE CATHEDRALS OF NORTHERN FRANCE. 283 above, and in others, are several examples wliich owe many of their features to English architecture. M:D deBOM SLCOVRS GVI/NGAAAP. rrf Fig. 198. French archaeologists use the term " flamboyant " for their archi- Flam- tecture of the fifteenth century, and the word expresses it well ; oyan ' not merely because the tracery of the windows forms flowing Unes, but because throughout aU the buildings of the period there is a light-heartedness which no other term would convey the effect of 284 A HISTORY OF ARCHITECTURAL DEVELOPMENT. so well. Contemporary work in England is equally well ex pressed by the term " perpendicular," not only because the heads of windows are filled with vertical bars, and the wall surfaces panelled with upright panels, but because the architecture, in its main lines, is stiff and angular. The work in France expresses the joie de vivre, inherent in the nation, which found expression after the triumph of the French in freeing the country from the English; and English work shows evidence of the gloom which settled over the country after the Black Death and War of the Eoses, which was not dissipated until the Tudors came to the throne. The first sign of flamboyant feeling, according to M. Enlart, is seen in one of the chapels of Amiens Cathedral, built 1373, but it was not untU some fifty or sixty years later that it appeared in fuU force. S. Maclou, Eouen (1437-1450), is the most delightful existing example of it. The west front expresses in overwhelming fashion the joie de vivre already referred to. It is to earlier work what the reel is to the minuet ; a burst of jollity which the purist in Architecture may shake his finger at, but none the less a tangible and vivid expression of the feeUng of reUef and delight which the deUvery of the country from war and anarchy had produced. At Alencon is a somewhat similar porch at the west end of the cathedral, but it lacks the abandon which is so marked in S. Maclou. Throughout Normandy are many other fine churches of the fifteenth century, such as S. Vulfran, Abbeville, unfinished, but nevertheless most impressive ; the churches at Caudebec, Dieppe, and elsewhere. The two churches which display the richness of this period more extravagantly perhaps than any others are Notre Dame de l'Epine, near Chalons-sur-Marne, and the Abbey Church of Vendome, near the Loire. The west fronts of both are striking ; that of the former especially so, with its two towers crowned by spires which recall somewhat the north-west spire of Chartres Cathedral. The last A considerable amount of church building was done after the sixteenth century had commenced — far more than in England where Protestant ideas took root earlier — and much of this is a mixture of Gothic and Eenaissance detaU which is often far from unpleasant. In the doorways of many churches, Beauvais Cathedral and at Abbeville, for instance, rich fifteenth-century Gothic masonry frames in Eenaissance woodwork, which is by no word. uo •A o £ THE CATHEDRALS OF NORTHERN FRANCE. 285 means out of keeping with its earUer surroundings. In travelling through France one is frequently astonished at the dates given to churches in which Gothic feeling is still paramount, although the detaU shows that its reign was nearly over. S. Pierre, Auxerre, is a striking church, and yet it is stated to have been rebuUt entirely in the seventeenth century. S. Nicolas, Coutances, the nave of which belongs to the sixteenth century, the choir and transepts to the foUowing one, is quite a remarkable example of imitation thirteenth-century work. The detaU is by no means bad, although it betrays its date in places. The traditional methods of mediaeval buUding are followed reUgiously, and the church is vaulted with ribbed vaulting, whilst under the tower the corbelUng has been copied from that over the crossing in the cathedral. S. Pierre, Coutances, is earlier (c. 1500). Of the same date is S. Jacques, Lisieux, a far finer church, with the simple quadripartite vault which the French clung to until the end, enriched with most effective painting dated 1552. These late churches, and others which might be mentioned, show how traditions were handed on unadulterated across the barrier of the century, and that many years after Gothic was counted as dead it survived in the churches of France. CHAPTER XVII. GOTHIC ARCHITECTURE IN SOUTHERN FRANCE. The Eomanesque monuments of Southern France are so full of originality and beauty, that no wonder they influenced to a considerable extent the churches of the thirteenth and fourteenth centuries of the district. All are founded on old Eoman traditions, which were still powerful when the latter were built. Wherever the Eoman settled he left behind him a legacy of grand scale. His love for large parts and for as few supports as possible, his preference for open, unencumbered floor spaces, and above aU his dislike for active demonstration of the means by which his buildings stood, are as evident in the mediaeval churches of Southern France as they are in the buildings of Imperial Eome. The result is that the main characteristics of the Southern church plan are wide naves, square bays (with some exceptions), internal buttresses, and a complete absence of flying buttresses. No cathedral in Northern France has the span of either Angers Cathedral (56 feet) or Albi Cathedral (60 feet). Not even in sexpartite-vaulted churches of the North do any double bays equal in size the single, square quadripartite-vaulted bays at Angers.1 In detail the Southern churches may not differ much from the Northern; in plan and general ordinance they belong to a totally different school. Types of Most Southern churches can be grouped under three heads — churches1 0-) Churches without either aisles or side chapels, such as Angers Cathedral (see Fig. 162). (2) Churches without aisles, but with chapels at the sides between internal buttresses of considerable projection, as Albi Cathedral. (3) Churches with aisles, the aisles being nearly as high as the central nave. The section was a common one in Eomanesque 1 The sexpartite bays of Notre Dame, Paris, are each about 45 feet by 35 feet, GOTHIC ARCHITECTURE IN SOUTHERN FRANCE. 287 churches of the South, such as S. Nazaire, Carcassonne, and the later Gothic churches of this type are their direct descendants. Poitiers Cathedral (c. 1160) is certainly the most remarkable £°jj™rs of those under the third head, and it is probably also the earliest dral POITIERS CATH : F.M.s.i88t Fig. 200. of the Gothic examples. In it, and in other similar churches, there are neither triforia nor clerestories. The lighting is entirely from the aisles and from the west and east ends. Owing to the great height of the aisles, and the absence of side windows in the nave walls, the aisle windows are considerably larger and of greater importance than in the Northern type of church. Poitiers 288 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Cathedral has no chevet ; its east end is square, although in the thickness of the wall are three apses, as in the early churches of Syria, Egypt, and Eome. The transepts are mere chapels, and of Uttle importance. The church is narrower at the east end than at the west, both in total width and in the width of its nave. The vault diminishes in height towards the east end, not so much owing to the decrease in width, as because its springing-line over the three bays forming the choir is the same as the springing-line of the side arches, whereas in the nave it starts from a consider ably higher level. In the church at Pontaubert, Burgundy, the chancel arch and vault to the east are dropped in a similar way, the springing-line of the vault of the apse at the end of the chancel being lower still. In the church of Montreal, also in Burgundy, the gradation in height is obtained in a stiU more subtle fashion. The transverse arches of the nave are stilted; those of the chancel are not, the result being that the vault of the latter is a few feet lower that the nave vault. In all three churches the differences in height are not noticeable; the only place in each where they are apparent at aU is above the chancel arch, as there a small portion of waU shows below the nave vault. The above-mentioned devices, and others which are generally met with in countries where old classic traditions stUl lingered, are not only legitimate, but are deserving of all praise. They are evidence of the care and thought which the builders devoted to their work ; and when unobtrusive, as in the examples described, are well worth the attention of modern architects. La Chaise A connecting Unk between Poitiers and Albi Cathedrals is ieu- the Church of La Chaise Dieu, Auvergne (c. 1344), although chronologically it is later than either. It is one of the best examples of a monastic church in France. The nave only con sists of three bays, whereas the choir has seven. The stalls, one hundred and forty- six in number, return at the west end and separate the monks' part from the people's. The nave and choir are about 50 feet wide. At the sides are aisles, which continue round the church, except at the east end, where they stop against the five radiating chapels which open directly out of the choir. These chapels are unapproachable from either the nave or aisles ; they belong exclusively to the choir. The church is like Poitiers Cathedral in having neither triforia nor clerestories: it differs from it mainly in the narrowness of its aisles, which, although lofty, are little more than ambulatories. Pig. 201. — Albi Cathedbal. [To face p. 289. GOTHIC ARCHITECTURE IN SOUTHERN FRANCE. 289 The cathedral at Albi (c. 1282) and the numerous churches Albi Ca in Toulouse form a group in which can be studied the essential 6 ra ' differences between Northern and Southern ideals. Almost all. are built of red brick. Even the jambs of windows and the buttress set-offs are generaUy brick, although the mullions and tracery are stone. Nearly all are aisleless, but the naves and choirs of most are surrounded by chapels, buUt between the strong internal but tresses which carry the thrusts of the vaults. At the west end of Albi Cathedral a square tower, with great circular buttresses at the angles, takes up the whole width of the front, and there is no western doorway. This alone marks a difference between ALBI CATHEDRAL 10 o jo h I I I I I I SCALE OF Fig. 202. Northern and Southern work. A Northern buUder would no more have dispensed with his western doorway than he would with his vault. There are no transepts. The only entrances are lateral, the main one on the south side consisting of a fine fifteenth- century stone porch, near the middle of the church, which is approached by a flight of steps, at the bottom of which is a gate way nearly as elaborate as the porch itself. The buttresses out side are segmental and have very slight projection. They, like the walls, are buUt with the long, thin bricks of the South, with wide mortar joints. The appearance inside is somewhat dis appointing after the towering height of the outside, helped as it is by the lines of the buttresses and the unusually lofty proportions of the windows. A hundred feet is no mean height for a vault ; vol. 11. u 2oo A HISTORY OF ARCHITECTURAL DEVELOPMENT. but when that vault has a span of quite 60 feet, its springing- line is no higher from the floor than the nave is wide. The effect would probably have been better if the length had in some way been broken. Nothing of consequence marks the division between nave and choir. The choir is merely a portion of the east end screened off by a stone screen, so as to allow of free passage to the chapels at the sides and end. Over the chapels which surround the church are galleries, the vaults of which are as high as the main vault. The gaUeries, like the aisles below, are divided into distinct bays by walls, and one misses, in consequence, the per spective effects which are CROSS SECTION or ALBI CATHEDRAL 5CALE OT FEET given by the lofty side open ings of the Church of La Chaise Dieu. SimUar open ings would undoubtedly have added greatly to the apparent height of the church, but would have necessitated more pronounced external buttress ing than the architects of the South desired. Notwithstand ing these defects, if defects they are, the church is the most interesting and inspiring in Southern France. Its plan for congregational purposes is far superior to the triple divi sion plan almost universal for large churches in Northern Gothic. The Cathedral of Perpignan, commenced about the same time as Albi, has a simUar plan and like dimensions, although its height is a trifle less. The main differences between the two churches are that there are no galleries over the chapels at Perpignan, and that over the arched openings to the chapels, which are much higher than at Albi, are uncusped circular windows, one to each bay. Clerestory windows of this simple type are common in Italy and Spam, and probably found their way to Perpignan over the Pyrenees. Toulouse. The Church of the Jacobins, Toulouse (c. 1300), is one of the most striking in a town full of churches of unusual plan and noble proportions. Most were built by either the Dominicans or the Franciscans, who exerted themselves so unenviably in the Fig. 203. GOTHIC ARCHITECTURE IN SOUTHERN FRANCE. 291 thirteenth century in the extermination of the religious sect known as the Albigenses. The Church of the Jacobins differs from the others in that, as is frequently the case in churches built CHVRCH OF THE JACOBINS, TOVLOVSE. SCALE OF FEET FOR PLA/1 CROSS SECTION. 10 o i-ii,i,ii.i v 50 J 1 1 LO/1G. SECTION. ELEVATlO/1. IOO ISO SCALE OF FEET FOR SECTIONS A/ID ELEVATlO/1'. Fig. 204. by the Dominican order, it has double naves of equal height divided from each other by lofty cylindrical columns which reach from the floor to the springing of the vault, and help to make the 292 A HISTORY OF ARCHITECTURAL DEVELOPMENT. church appear far more lofty than the single-span Cathedral of Albi, although in reaUty it is nearly 10 feet lower. Along the sides are low chapels without galleries over them. Outside, the buttresses, which are internal for the height of the side chapels, stand out boldly above the lean-to roofs. They have few and unimportant set-offs, and are connected by arches thrown across from one buttress to another above the windows. Over each arch is a pierced circle. This type of design was a favourite in Toulouse. It still exists in the partially-ruined Church of the Augustines, and it also occurred in the Church of the CordeUers, now destroyed, which in many ways must have been the finest in Toulouse, the Eomanesque S. Sernin alone excepted. The destroyed church CH.ofthe CORDELIERS, TCA/LOVSE. v1VVlvi\/?\ SCALE or'?, ? i » I too __l FEET. Fig. 205. had a nave nearly as wide as that of Albi Cathedral, with low chapels opening out of it on either side. Over the chapels surrounding the apsidal ending were other chapels, carried up to the full height of the church, so that round the east end the design followed still more closely that of Albi. Of the Church of the Augustines only three of the original bays remain. These are 21 feet from centre to centre, the width from north to south being 60 feet. Altogether there are few towns in France which can compare with Toulouse in churches the plans of which are so suitable for large congregations. They are proof that the Gothic of the North, with its nave and aisles, its flying buttresses and its inside divisions of nave walls, is not the only Gothic worthy of study. The fame of Chartres, Eeims, Paris, and Amiens has eclipsed the GOTHIC ARCHITECTURE IN SOUTHERN FRANCE. 293 Southern work ; but when one comes to practical considerations, to deaUng with the plan and ordinance best suited to modern requirements, it is a question whether it is not far better to shut one's eyes resolutely to the glamour of the Northern churches, and to open them freely to the many good points which the Southern ones possess. CHAPTER XVIII. GOTHIC ARCHITECTURE IN ENGLAND. Intro- The middle of the twelfth century marks the commencement of duction. a new era -n j5ng]ign church building. From it dates the rise of a national feeling which in less than half a century led to the freeing of English art from continental tradition, and to its development on Unes different from those which were being pursued elsewhere. Nearly a century had passed since England had come under foreign rule. During that time great changes had naturally taken place in the relations between the two races, between the conquered and the conquerors. The old conditions, under which a hard-and-fast line was drawn between Normans and English, had to a great extent changed, especially amongst the dwellers in towns. The Norman traders who followed the Conqueror to England had at first Uved apart in their own quarter, alongside that occupied by the original inhabitants but separate from it. Gradually the barriers were broken down. The descendants of William's followers had no longer need for them. England was the country of theu- birth ; and by the intermarriage of their forebears many were half of English blood. A more powerful reason for fusion was the identity of interests between all burghers, no matter what their descent. Union between the two races was necessary in order to obtain freedom for their cities, security for themselves, and to repel more easily the aggressions of the barons, the extortions of the king. The marriage of Henry I. with MatUda, an EngUsh Princess, helped, to a great extent, to reconcile the EngUsh to foreign rule. It bore fruit when Henry's elder brother Eobert, Duke of Normandy, claimed the EngUsh throne. The sympathies of the Norman barons in England were largely with Eobert, and the army that followed Henry to Normandy and won the battle of Tenchebray was composed mainly of English yeomen. With the accession of Henry IL, the grandson of Matilda, through the marriage of Geoffrey of Anjou with the daughter of the first GOTHIC ARCHITECTURE IN ENGLAND. 295 Henry, one more step was taken towards breaking down the barriers of race. Besides being King of England, Henry II. was the master of Anjou, by descent, of Normandy, which his father had conquered, and of Aquitaine, through his marriage. He therefore owned practically haU France as well as this country. His accession meant that the crown passed to one whose Norman strain was equaUed by his English, and who, in addition, was half Angevine. No wonder the accession of Henry II. brought together Angevine the dwellers in England. The king was as much a foreigner to mfluenoe- the Normans as to the English, and just as much a feUow-country- man of both. At the same time it was long before there were any changes in the Court and in the Church. The barons and prelates remained Norman, with the addition of a few Angevines. No bishop or abbot was EngUsh. Wuffstan of Worcester was the only surviving bishop of English blood when WiUiam the Conqueror died, and no EngUshman was appointed to a see untU the thirteenth century. But a break had been made in the Norman succession, and the break brought about a change. It came graduaUy. The Benedictine monks still controlled all the great monastic churches in England and stiU looked to Normandy, or to Burgundy the cradle of their order, for guidance in architectural matters. UntU the end of the twelfth century, as conservative here as their brethren were abroad, they continued to buUd on the old lines, as at Ely and Peterboro'. It might have been thought that the union of England, Anjou, Aquitaine, etc., under one ruler would have led to the importation of Southern methods of building into this country. But such was not the case. Although the inhabitants of the above countries were in a sense feUow-countrymen, Angevine characteristics — the aisleless plan for large churches, domical vaults and heavy transverse arches — found no echo in English work. To the activity of the Cistercians, foreigners as they were, Cistercian must be attributed in part the break with Norman traditions, activity. In the first quarter of the twelfth century they flocked to England, bringing with them the maxims of their order and their ideals of church buUding. Their first settlement was at Waverley, near Farnham. That the square east end was an innovation of theirs has already been denied (see pp. 161 and 237), although their advocation of it probably helped to make it more general. Neither can they be said with any certainty to have introduced the pointed 296 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Transi tional work. French influence. arch into England, although it is quite possible they did so. No church of theirs with pointed arches is so early as the nave vaults of Durham Cathedral, and as this was a Benedictine church it is unlikely that they had any say in its building. What is commonly called Transitional work in England covers approximately the latter haU of the twelfth century. In it there is an increasing Ughtness of all parts, a mingling of pointed and semicircular arches — the former being used for main arcades and for vaulting, the latter for heads of windows and other smaU openings — and a change in mouldings and ornamentation. The work is transitional in a double sense ; it not only marks the transition from Eomanesque to Gothic, but it expresses also the change that was taking place in England in the relations of the people towards the barons, the barons towards the king, on which the seal was set by the Magna Charta of 1215. The period was by no means one of stagnation ; but advance was only partial. There was no longer the same building activity as prevailed during the previous century. WhUst contemporary French art was advancing by leaps and bounds, EngUsh art was but carefully feeling its way. For this the quarrels of Henry with the church and barons, the restless ambition of Eichard, which emptied the exchequer, and the double dealing of John with both the Pope and his own people, are mainly responsible. Although French Gothic art of the latter half of the twelfth century was considerably in advance of contemporary work in England, there are very few examples in this country which can be classed as direct evidence of French influence. Indirectly there may be many, as the English builders cannot have been ignorant of the great strides their art was making across the channel. The choir of Canterbury (c. 1175-1178) is universally admitted to be the work of William of Sens, a Burgundian, and to be continental in design. The mouldings of the arches, the carving of the capitals, the shape of their abaci, the starting of the vaulting shafts from above the capitals, and the sexpartite character of the vault recall Laon, Paris, and Lisieux. The coupled columns of the eastern apse are simUar to those in Sens Cathedral, and in the apses of some of the large Normandy churches. In the choir of Eipon Cathedral (c. 1170) the vaulting shafts start as at Canterbury, and the quadripartite vaulting of the aisles is more domical than is usual in English work. The retro-choir of Chichester Cathedral (c. 1186) is, in some of its GOTHIC ARCHITECTURE IN ENGLAND. 297 detaUs, similar to Canterbury, although built after the death of WUliam of Sens. The foreign influence traceable in Chichester may be due to its proximity to the coast of Normandy, it being the nearest English cathedral to that country. But these examples are few on which to base the assumption that English Gothic was derived from the French. It developed mainly out of its own Eomanesque. There are other contemporary churches in England which show little trace of French Gothic influence and some which show none at all, proving that the Englishman, slowly but steadily, was finding a way for himself. Amongst the earliest examples in England in which the pointed arch occurs in arcades are the abbey churches of Malmes- FOWITAIfiS ABBEY CHVRCH SHEWING AISLE VAVLT RESTORED Fig. 206. bury (c. 1145), Buildwas (c. 1148), and Fountains (c. 1150). The first was Benedictine, the other two Cistercian. At Buildwas the design is frankly Eomanesque ; the substitution of the pointed arch for the semicircular being the only modification. The same Transi tionalexamples. 298 A HISTORY OF ARCHITECTURAL DEVELOPMENT. may be said of Malmesbury nave, except that the pointed form is also used in the vaulting of the aisles. At Fountains the aisles are vaulted by pointed barrel vaults running from north to south, each bay being carried on semicircular transverse arches in similar fashion to the nave vaults of S. Philibert, Tournus. The two bays at the west end of the nave of Worcester Cathedral (c. 1160-1180) show a lighter and more advanced treatment, although the work is still impregnated with Eomanesque feeling. The piers have detached shafts in the angles instead of the attached ones which had hitherto been customary. In the abbey church of Much Wenlock, Shropshire (see Figs. 25 and 190), ascribed by Mr. Prior to c. 1180, a great advance is noticeable; in fact, if the above date is correct little improvement was made for the next twenty or thirty years. Connect- ^ne connecting links between the Transitional work of the ing links, second half of the twelfth century and the fully developed Gothic of the following one are to be seen best in the two cathedrals of Wells and Lincoln. In these can be traced the steps by which the later work emerged from the earUer by a natural development, unassisted, or only assisted indirectly, by contemporary work on the continent. Mr. Prior gives the date of the nave of WeUs Cathedral as 1170. It may have been commenced then, but it seems unlikely that much was done beyond foundations before the end of the century. The Uttle ruined chapel of Glastonbury Abbey close by was begun after a fire in 1184. Notwithstanding its delicacy and refinement it is stiU essentially Eomanesque in its ornamentation, without a trace of the later feeling conspicuous at WeUs, excepting the pointed arches of its vault. Allowing for possible jealousies between the monks of Glastonbury and the canons of Wells, it is inconceivable that the work of the cathe dral, if it had been at aU advanced, should have exercised no influence on a building only a few mUes off; a buUding, more over, which for richness of detaU has no contemporary equal. The piers of Wells are square with added attached shafts. The main difference between them and Eomanesque ones is that the shafts are clusters of three instead of single shafts. It is quite possible that in 1170 the earUer form may have been contemplated, but as the work proceeded the section was changed.1 The Wells 1 In the nave of Chichester Cathedral, the semicircular shaft on the face of each pier was carvod into a cluster of three when the rebuilding took place after the fire in 1186. GOTHIC ARCHITECTURE IN ENGLAND. 299 piers are certainly an advance on earlier English forms, and they are, moreover, quite different from continental ones. The square abaci of the capitals show the Ungering of Eomanesque traditions, and are not suggested by the square abaci of France. The nave of WeUs is thoroughly English ; and the only question is whether it should be assigned to 1170 or to some thirty years later. The triforium storey consists of very narrow pointed headed openings, which form a continuous band of arcading, there being no vaulting shafts above the piers to divide them into bays. The existing stumpy vaulting shafts start above the arches of the triforium, and appear as after-thoughts. It is quite possible that at first there was no intention of vaulting the nave at all. The work throughout is very simple, and is as sturdy as that of Eeims A.WELLS CATHEDRAL PIER E>. A ROM AM- -ESQVE PIER Fig. 207. Cathedral. Detached shafts are conspicuous by their absence, and although one misses the differences in colour which are so striking in Salisbury Cathedral and Worcester retro-choir, it is question if the sense of dignity and repose which the nave of Wells possesses is not ample compensation. The eastern transepts and the choir of four bays of Lincoln Cathedral, the work of Bishop Hugh, were commenced in 1192, and early in the next century a start was made with the great transepts (Fig. 100). The designer was De Noyer (or De Noyers), but whether he was an Englishman or some one brought by S. Hugh from his native town, Avallon, in Burgundy, is a matter of dispute. The introduction of eastern transepts suggests the great church at Cluny ; and the double apses to the east of each may also be regarded as a somewhat foreign plan. These, however, are quite as likely due to the bishop as to his architect. 300 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Fully de velopedGothic. Normandy French. The great transepts and the aisles of the smaller transepts have sexpartite vaulting, and the vaulting shafts of the former start French fashion from the top of the capitals of the piers. Other wise, outside inspiration is lacking. The vaulting of the choir shows an irregularity in the plan of the ribs which no Frenchman would have dreamt of attempting. It can hardly be considered satisfactory, but may be regarded as an early attempt to reduce the size of the compartments of each bay ; a desire which animated all English builders throughout the following centuries, and led to the developments in vaulting pecuUar to England which have already been described (see Fig. 52). A feature in S. Hugh's work is the double arcading which runs under the windows round the aisle walls. TrefoU arches supported on shafts stand in front of pointed arches against the wall, the arches crossing one another. The idea was doubtless taken from the intersecting semicircular arcades so common in aU Eomanesque work, but only in England was it developed in this manner.1 In the GalUee porch of Ely is a particularly good example of somewhat simUar design. By the beginning of the thirteenth century EngUsh Gothic had fairly found its feet, and was capable of walking alone. With the one exception of Westminster Abbey, and that only as regards its plan and proportions, there is no church built subsequently which can truthfully be said to owe its plan, its ordinance, or its detail to French influence. Window tracery was certainly more developed in the middle of the century in France than it was in England; La Sainte Chapelle, Paris, is earUer than either Westminster Abbey or the Presbytery of Lincoln Cathedral. But the designs of English window heads executed towards the end of the century are very different from contemporary French ones, and early in the following century our masons in their turn gave a lead to the French by the substitution of flowing lines for purely geometric forms. The beginning of the thirteenth century saw the severance of England and Normandy. Chateau Gaillard, which had been buUt by Eichard I., a few years before, surrendered to PhUip Augustus in 1204, and the whole duchy passed to the French crown. Intercourse between the people of the two countries 1 Lincoln Cathedral ranges over the whole gamut of Gothic, from the Romanesque of the entrance doorways to the fifteenth-century chapels flanking the south porch of the presbytery, which give such scale to the great mass of the eathedral rising behind them. Pig. 208.— Amiens Cathedral. [To face p. 300. GOTHIC ARCHITECTURE IN ENGLAND. 301 ceased, except on occasions when it was otherwise than friendly. Notwithstanding this, it is interesting to note that for many years architectural development proceeded on much the same lines in some respects in both countries. Constantly in Normandy, especiaUy in its western half, one comes across detail of the thirteenth century, as at Bayeux (c. 1230), Coutances (c. 1220), etc., almost exactly the same as is to be found in England. In the dismantled church of the Abbaye d' Ardennes, near Caen, this similarity is especiaUy marked. Detached shafts, circular capitals and mouldings of alternating deep hoUows and bold round members are almost as general there as here. The masons of the two countries had been trained in the same school, and they continued to work on much the same Unes and with similar stone. In plan, Normandy churches of the thirteenth century are very different from our churches. The Normans employed the chevet at Lisieux before the separation came, and this plan they did not abandon untU late in the fifteenth century, and then only partiaUy. In other respects — side chapels to nave aisles, for instance — they also followed the Isle de France. As time went on and generation suoceeded generation the masons' work changed, and the later mouldings in Normandy bear Uttle resemblance to our fifteenth-century ones. 1200-1250. Comparison has so frequently been made between Salisbury Salisbury and Amiens that the present choice of these two cathedrals for Amiens. this purpose may suggest a lack of imagination. But no others can so weU and forcibly bring home the fact that English and French methods of design were proceeding on totaUy different Unes when these two churches were building.1 They are brothers, but as unlike each other as brothers often are who come of a common stock. Salisbury was commenced by Bishop Poore in 1220 and was finished some forty years later, with the exception of the cloisters (1263-1284), which show the hold tracery had then obtained over the masons, and the central tower and spire, the 1 For plans of these two churches see Figs. 100 and 188. In one respect com parison is unfortunate, as Amiens is the largest of French cathedrals, and the highest except Beauvais, while Salisbury is exceeded in height by York Cathedral and Westminster Abbey, in length by no fewer than nine cathedrals (including old S. Paul's) — Winchester being nearly one hundred feet longer — and in width by six ; but the comparison is not so much made as regards size, but as regards design. 302 A HISTORY OF ARCHITECTURAL DEVELOPMENT. greater part of which belongs to the fourteenth century. Amiens was begun in the same year, and although finished earlier had to be partially rebuilt, owing to a fire, as already stated. In the rebuilding, alterations were made in the original design, especially in the clerestory windows, and it is not, therefore, so complete an example of consecutive building as the English church. The choice of Salisbury Cathedral as the chief example of work of the first half of the century possesses this great advantage, that the church was built on a new site, and the builders consequently were not hampered by having to consider old foundations or existing remains. Both plan and general ordinance may there fore be taken as representing the ideals of the EngUsh builders of the day. The most noteworthy points about the plan are its absolute symmetry and its rectangularity (see Fig. 100). So much is said about the irregularity of Gothic work, that it is as weU to emphasize the fact that when not tied down and restricted by existing surroundings, the mediaeval builders adopted plans as symmetrical as any to be found in ancient Eome. The only features which have not their counterparts are the north porch, and the cloisters and chapter house to the south. Salisbury has a nave of ten bays west of the crossing, and a choir and presbytery of seven bays to the east of it, beyond which is a Lady Chapel of four bays, aU ending square. The great transepts project three bays beyond the aisles, the smaller transepts two bays. On the east side of both pairs of transepts is an aisle divided into chapels. There are single aisles to nave and choir and no side chapels. Amiens has a nave of seven bays and a choir of four, or five including the apse. Beyond is the semicircular chevet with its radiating chapels. The eastern arm has double aisles, the nave single aisles, beyond which are chapels, and the transepts have an aisle on both sides. The transepts project only one bay beyond the outer choir aisles, and that bay on each side is much narrower than any other in the church (see Fig. 188). There are no eastern transepts. The internal width of Salisbury is 78 feet, exactly half of which is the width of the nave from centre to centre of the piers, the aisles dividing the remaining half. The western arm of Amiens is 103 feet wide, exclusive of chapels, and is divided in the same way. The eastern one is about 155 feet wide, nearly double that of SaUsbury. On the other hand, the measurement of Salisbury across the transepts from north to south is 204 feet, which is some 5 feet wider than in the great French church. In total length Photo : Frith & Co. Fig. 209. — Salisbury Cathedral. [To face p. 302. GOTHIC ARCHITECTURE IN ENGLAND. 303 Salisbury also has a slight advantage. From the west wall to the arcade at the back of the high altar — the high portion of the church — is 377 feet, the total length being 450 feet. Amiens, measured from a similar position, is 373 feet to the apse termi nating the choir, and 449 feet to the east waU of the Lady Chapel. The total eastern arm in each church is a trifle longer than the western, but as the full height is not maintained for the whole length in either, externally the western arm appears to be longer in both. The differences in proportion become still CROSS SECTION 10° OF A/1 ENGLISH A/ND A FRE/iCH CATHEDRAL. SALISBVRY. AAMEttS, Fig. 210. more marked when the heights are considered. Amiens, the second highest church in France, is 140 feet high from floor to apex of vault; Salisbury, the third highest in England, is only 84 feet. The above is sufficient to show how, whilst breadth and height were the aims of the French builders, the EngUsh, modest in these respects, reUed more on length. The latter could, in consequence, afford to give their main transepts considerable projection and to add eastern transepts. They could carry a stone tower and spire above the crossing, a far more effective feature than the wood and lead fleche of Amiens, although the latter does rise some 304 A HISTORY OF ARCHITECTURAL DEVELOPMENT. 20 feet higher than the former. A tower and spUe to the latter church, even if possible, would have been a mistake.1 It would have detracted from rather than added dignity to the outside. No one studying the plan, appearance, and main ordinance of these two churches can doubt for a minute that the ideals advanced on one side of the channel were not those striven for on the other. If it be asked which are right and which wrong, there is only one answer. In art there are no rules, no laws, no absolute right or wrong. Each building is a law to itself and to itself alone. Each should be studied separately ; its defects and beauties noted ; its plan, features, and ordinance analyzed. There is a reason for every difference. Eaise the vault of Salisbury, and its tower would require to be reduced in height, its spire omitted; its eastern transepts would have to go, or else its Lady Chapel be raised to the same height as the choir. At Amiens the cathedral rears its whole body above the houses which hem it in on all sides ; the tower and spire alone of Salisbury Cathedral form the landmark. But each building tells the same story, although in a different way, the story of the majesty of the Church, of which it is the visual monument. We need feel no envy of the 140 feet that the vault of Amiens rises above the pavement, of the greater bulk of the buUding, of its forest of flying buttresses. Salisbury inside looks its height, and it is doubtful if that can be said of its great rival. Bobert de Luzarches and the De Cormonts were bolder constructors than the men who served Bishop Poore and his successors so faithfuUy, but it does not follow that they were bigger artists, or that they have left behind them a more beautiful monument. West The west fronts of EngUsh Cathedrals are weak, and Salisbury is no exception to the general rule. In fact, it is one of the least satisfactory, especiaUy when seen sideways. There are no western towers, and a horizontal parapet connects the gable with angle turrets which are poor substitutes for them. At Lincoln, horizontality is stiU more marked, as on each side of the central gable is a screen which extends far beyond the western towers, which rise behind it, and finishes with turrets simUar to those at SaUsbury. The two most famous thirteenth-century fronts are Peterboro' with its triple porches, the central one, curiously enough, narrower than the side ones, and Wells (c. 1220). More has 1 At Beauvais an attempt was made in the sixteenth century at a crowning feature over the crossing, more a dome than a spire, but with dire results. fronts. Photo : Frith S: Co. Pig. 211. — Wells Cathedral : West Front. Photo : Frith i- Co. Fig. 212. — Salisbury Cathedral, erom North-east. [To face p. 304. GOTHIC ARCHITECTURE IN ENGLAND. 305 probably been written about the west front of Wells Cathedral than about any part of any other building, mainly because of the beauty of its sculpture, which extends round the towers at the angles as weU as along the entire front. The west front is unusuaUy wide for an English Cathedral, because the towers stand entirely beyond the aisle walls, and do not terminate the aisles, the usual custom both in England and abroad. Nothing can be finer in design than the lower part, notwithstanding the insignificance of the doorways. For the smallness of the door ways leading to the aisles there is every excuse ; in fact, they are better small, because they give scale — as the small doorway does in the north side of the nave of Westminster Abbey ; but the main portal might with advantage haves been twice the size, and yet have been none too big. The difference in relative size and importance of French and EngUsh entrances is an interesting point of difference in the work of the two countries. CUmate can hardly account for it ; and WeUs is a proof that it was not want of skUl on the part of EngUsh sculptors which led them to dispense with the many figures in the jambs and the multitude of canopied figures in the arched heads which are so characteristic of French work. Other countries copied the doorways of France, England did not. The east end of Salisbury spreads itself out over much ground East ends. with buUdings lower than the bulk of the cathedral behind. Not so Ely Cathedral, which is, perhaps, the finest example of work of the first half of the thirteenth century in England. There the choir and presbytery (c. 1235) are continued their full height to the east waU. In these two cathedrals, therefore, the two methods of eastern ending can be studied (see also pp. 151 and 152). The large window in the gable of Ely Cathedral lights only the roof above the vaulting, and consequently does not show at all inside. At the east end of Lincoln the window which occupies a simUar position is even larger. In many of the Yorkshire churches, Selby Abbey, for instance, the top windows are also of considerable size. They add greatly to the external effect, but otherwise are not of much use, as much smaller windows would have given all the Ught necessary. Additions which were made towards the end of the first half of the century at Durham Cathedral and Fountains Abbey Church show that neither the Benedictine monks of the former nor the Cistercian monks of the latter were desirous of imitating the chevet plan which was general on the continent. At vol. 11. x 306 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Other examples. Early English. Durham the Chapel of the Nine Altars (1242-1280) (see Fig. 177), and at Fountains a similar chapel for seven altars, form transepts at the extreme east end, stretching beyond the aisles on either side. In both cases the resemblance in plan to the eastern transepts of some of the large basilican churches in Eome is marked, except that there are no apsidal projections to the eastward in either church similar to those in the earlier basilicas. Cistercian influence is probably responsible, to some extent, for the form these additions took, notwithstanding that Durham Cathedral belonged to the monks of the rival order. The Durham chapel forms a magnificent vaulted hall about 35 feet wide, 127 feet long, and 80 feet high. Its floor is dropped some 6 feet below the choir floor, which accounts for its height being greater than that of the rest of the church. At Fountains the addition is even larger, about 36 feet wide by 132 feet long, but its length is broken by two piers, continuous with those of the choir, which rise to a great height and support arches immediately under the vault. Its effect as a great hall is therefore somewhat impaired, but on the whole the introduction of piers is an improvement. The south transept of York Cathedral (c. 1230) is another contemporary example, and equally fine is the north transept (c. 1250), in which the " Five Sisters " window shows the determination of the English builder to adhere to the taU lancet lights as long as possible, and bis disincUnation to spoil what he regarded as right proportions by the insertion of traceried heads. Besides its metropolitan church, Yorkshire suppUes other interest ing work of the period in the ruins of its abbey churches — some Cistercian, as Jervaulx, Eivaulx, etc., commenced towards the end of the twelfth century, and others Benedictine, as Whitby, built early in the following one. The greater portion of Hexham Abbey Church, Northumberland, is also contemporary. Between 1180 and 1250 is comprised what is commonly known as Early English. It is the Doric of Gothic architecture. At no other period do the designs show such strength, freshness, sim plicity and refinement. There is a stateliness and quiet dignity about everything built between these years which had not been reached before and was never equalled afterwards. The work may not have the vivacity and little fascinations — architectural chiffons — of that of fifty years later, but it possesses other and sounder qualities. Moreover, it is thoroughly English. The steps by which it was evolved are so unmistakable, and can be traced so Photo : W. A Pig. 213. — Ely Cathedral : East End. [To face p. 306. GOTHIC ARCHITECTURE IN ENGLAND. 307 easily, that any student of architecture who knows it, and also preceding work in England and contemporary work in France, can see that its growth was a natural one from local examples, and that it owed little to outside influence. There are no breaks or sudden changes such as would have been perceptible if it had been an importation from outside. Its counterpart, it is true, is some times to be met with in Normandy ; but in other parts of France the churches show other quaUties, not less beautiful and often more advanced, but none the less quite different. 1250-1300. Two examples, Westminster Abbey and the Angel choir of Lincoln Cathedral, wiU suffice to demonstrate the characteristics of the architecture of the second half of the century, and wherein it differs from the work of the first half. Henry IIL, who is responsible for the rebuilding of Edward the West- Confessor's Abbey Church at Westminster (c. 1245-1256), took a 5£?ster keen interest in architecture, but he was so much under the thumb of foreigners that his church at Westminster naturaUy possesses some of the characteristics found mainly in the art of France. That he invariably spoke French is not of much moment, because that tongue was for another century at least the language of the court.1 The chevet plan of the eastern arm and the main proportions throughout the church point to the employment of a French architect to make the design. The size of the windows (in the clerestory (c. 1253) the Ughts are each 4 feet wide), the number of flying buttresses (Westminster is the only church in England which has double tiers), and the relative width of wall to column (see comparison with SaUsbuiy, Fig. 17) are also evidence of continental influence. But if the designer was a Frenchman, the masons employed were English. The use of Purbeck marble, the detached shafts round the columns and the moulded bands, the rounded abacus and the detail of capitals, bases, arches, etc., the design of the triforium and wall arcading, the absence of any stilting of the window heads in order to get larger circles above 1 Until the Great Plague of 1349 all teaching in the better class schools and all pleadings in the law courts were in French. John Kichard Green, in his " Short History of the English People," records that in 1362 an order was issued that English was to be used in law cases because " the French tongue is much un known," and he adds that after 1385 French was abandoned entirely in English Grammar Schools. 308 A HISTORY OF ARCHITECTURAL DEVELOPMENT. A = WESTMI/NSTER, CHOIR. B = LI/SCOL/S, A/SGEL CHOIR. OF FEET Fig. 214. GOTHIC ARCHITECTURE IN ENGLAND. 309 the Ughts, and the diaper-work on the walls, all belong to this country. By the middle of the thirteenth century the English masons had established a tradition, and although willing to adopt features which they felt would be improvements, they were by no means willing to sacrifice others which were to their liking. Westmmster Abbey undoubtedly exercised some influence on work immediately following it, although not to any great extent. The clerestory windows of Hereford Cathedral (c. 1260) strongly resemble those in the outer wall of the triforium at Westminster, and the latter may also have suggested the spherical triangles, fiUed with cusped circles, which form the clerestory windows in Lichfield nave (c. 1275). The Angel choir or presbytery of Lincoln (c. 1256) foUows Angel close on Westminster. Tracery had been tentatively introduced °^°1T< J J Lincoln. in the triforium arcades of both the transepts (c. 1210) and the nave (c. 1230), but the design of the windows of the presbytery shows a marked improvement on these, and also on the windows at Westminster. If the builders of the extension learnt anything from the latter, they mastered their lesson so thoroughly that they were able to advance far beyond it. They probably knew little or nothing about the other church. Certainly two contemporary examples more unUke it would be difficult to find. At Lincoln the problem was how to improve on the design of the rest of the church and yet make the new work in harmony with the old. The bays of the Angel choir are perhaps the most beautifully proportioned of any in England, and they are as unlike the bays of Westminster as could weU be imagined. Each bay in the latter church is 17 feet 6 inches wide from centre to centre of the columns, and the height from the floor to the apex of the arch is about 42 feet. At Lincoln the corresponding dimensions are 23 feet 2 inches by 33 feet 6 inches. The triforium in both examples consists of four openings, grouped in pairs, but the greater width of the bays in the cathedral makes the proportions altogether different. The Westminster triforium is lighted by windows on the outside waU, but Lincoln triforium is dark behind. Triforia in English cathedrals of the thirteenth century and first half of the fourteenth, as a rule, are more important than in con temporary churches abroad. They are not vaulted, as in early French churches, but are covered by lean-to timber roofs, which finish under the clerestory windows. In some cases these roofs are of flat pitch, and the walls of the aisles are raised high above 310 A HISTORY OF ARCHITECTURAL DEVELOPMENT. their vaults so as to allow of windows above the aisle windows. In the western arm of the Cathedral of Worcester (which is late fourteenth-century work) pairs of windows to each bay light the triforium. The plan is in many respects superior to that in which the roofs are of steeper pitch, and the aisle walls in consequence no higher than the aisles themselves. The extra height given to the side walls is a great gain to the outside; the roofs over the galleries are entirely concealed by parapets ; the triforia are well lighted, and a certain amount of subdued Ught finds its way into the church through its middle storey. At Westminster, the clerestory windows are each of two lights, uncusped, with a foUated circle above ; at Lincoln they are each four Ughts, with three circles above. Moreover, in the cathedral a pierced screen of simUar but richer design is repeated on the inside face of the waU behind each window. The vaulting at Lincoln starts unusually low, some feet below the clerestory string-course ; that at West minster starts many feet above it, an exceptional position for England, but common in France. The general effect would probably have been better at Lincoln if the vault had been raised, but its apex was fixed by the earUer vault alongside it. Churches of the Fourteenth and Fifteenth Centuries. Intro- With the thirteenth century departed much of the virility duction. which distinguishes the first phases of Gothic art in England. Much good work continued to be done, work far more elaborate than anything which had been attempted before, but lacking the rare charm and simple directness of that which preceded it. For this, social changes and national calamities are largely responsible, although in all art movements a period of advance is invariably followed by a gradual decline, just as daylight breaks, expands, and then wanes. The preaching of Wyclif and the consequent agitation of the Lollards caused dissension and weakness in the Church. A series of years of famine was followed by the Black Death in 1349, a scourge which, it is estimated, destroyed more than one-half of the population. It affected all classes. The farmers were unable to tUl the land and gather the crops for want of labour ; prices rose considerably because of the amount of land thrown out of cultivation, and all workers, whilst having to pay a far higher price for the necessities of life, were forbidden by law to demand a higher wage than that which had been customary GOTHIC ARCHITECTURE IN ENGLAND. 311 before the pestilence came. The effects of the Black Death would not have been so disastrous, and recovery would have been more rapid but for the continuous wars abroad, and the constant demands on the people's purse which their prosecution entailed. The unrest throughout the land was exemplified by the peasant revolt of 1381. Early in the fifteenth century the victory of Henry V. at Agincourt brought glory to the king, but little profit to the people. The latter part of the century witnessed the War of the Boses, but this had not the paralyzing effect on commerce and trade that civU war generally entails, as it was a Barons' struggle in which the people took little part. No century, hardly any decade, in the Middle Ages passed without insurrections or wars, and those of the fourteenth and fifteenth centuries damaged architecture far less than the Black Death. The extent of cathedral and abbey church buUding was less than in the previous centuries, because early activity had left little to be done in that direction. But there was by no means stagnation in ecclesiastical centres ; only their energy took another form. The number and size of parish churches increased to an extraordinary extent. No king interested himseU strongly in church architecture after Henry IIL, but many bishops, such as William of Edington and William Wykeham, both of Winchester, were as keen on having their churches beautiful and up to date as any of their predecessors. The clergy, outside those of high position, were poor, and the monastic orders indifferent, but another race of donors was springing up. Increase in trade had brought great wealth to the merchants and others of the middle class, and many were willing to spend it royally. The rich burgesses had no doubt subscribed freely in the past, but not to the same extent as their descendants did in the fifteenth century. At Lavenham, Suffolk, the greater part of the cost of the church was defrayed by a wealthy clothier named Spring, and evidence of secular generosity, as well as anxiety for the welfare of their souls, is shown in the chantry chapels erected in many of our churches and cathedrals at the expense of laymen. The fourteenth century is perhaps more noteworthy for its castel lated architecture than for its ecclesiastical. No cathedral was commenced after the thirteenth century, and none illustrates the characteristics of either fourteenth or fifteenth century Gothic so thoroughly as Salisbury and Lincoln do those of the first and second halves respectively of the previous one. But although none was begun, such extensive remodelling was carried out at 312 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Absence of con servatism. Exeter Cathe dral. Exeter in the fourteenth, and at York and Winchester in the same and the following century, as almost to amount to entire rebuildings. Up to the end of the thirteenth century a certain amount of respect for earlier efforts is noticeable in additions, and a desire to harmonize new work with old ; but in much subsequent build ing a total disregard of what existed becomes apparent, and a determination to obliterate it at all costs. Thus, at Gloucester (c. 1340-1350), the monks, unable to afford to take down bodUy and rebuild the sturdy walls and piers of their predecessors, never theless went to great expense to conceal them with a veneer. They did take down the eastern apse, and beyond it built a magnificent east window, cleverly planned with sUghtly canted sides to rest on the walls of the old ambulatory. A more conservative spirit was shown in the naves of Worcester Cathedral and Westminster Abbey. In the former a most creditable attempt was made to foUow the very beautiful earlier work of the retro-choir. The bays on the north side were built first c. 1320), and are better than those on the south side, which followed some fifty years later. On both sides the capitals have circular abaci — an unusual form at this period — and, like those of the presbytery, are of Purbeck marble. The design of the triforium of the eastern arm is also repeated in the western, with some modifications, the most important being the omission of the Purbeck marble shafts which are such a feature in the earlier work. In the nave of Westminster Abbey, commenced 1350, but not finished until some seventy years later, absolute harmony with the choir was obtained, although differences in detail are apparent to those who look for them. The columns are surrounded by attached shafts, not detached ones as in the eastern arm, and the capitals and bases are octagonal in plan and not circular. The bands which bind the shafts to the columns in the earUer work are retained, and their presence may be forgiven, as, although no longer needed, they help to produce a uniformity of design throughout, which few large churches in England built at different times present. Exeter Cathedral may be taken as a fairly complete and representative example of fourteenth-century work, although its foundations and portions of its superstructure belong to the earlier church which was built at the same time as the Norman towers which still stand over the transepts. Owing to the lateral position of these, the piers at the crossing are no larger than those of the GOTHIC ARCHITECTURE IN ENGLAND. 313 nave and choir, and consequently there is no break in the arcade. It is this, coupled with the continuous vault, that makes the church appear too low, and not so much its actual want of height (see table, p. 376). Its lowness would be far more marked if it were not for the screen and organ which luckily stand at the entrance to the choir — by far the most dignified position for an organ — thus giving scale, and at the same time providing a break in the length where one is most needed. The proportions notice able inside are still more remarkable outside, especially at the west end, and at the east where the SaUsbury plan of a low retro-choir and Lady Chapel is followed. The design of the west front shows a determination to sacrifice height to breadth which is never met with in Gothic churches abroad. The battlemented parapet of the CATHEDRAL. SCALE OF SO I 1 1 1 250 FEET. Fig. 215. main roof is continued horizontally across the front, the gable standing back some feet behind it. It is also returned on the rake down the sides, masking the aisles, 'with not a very happy result. The front would doubtless have looked better if it had not been for the porch, added in the fifteenth century, which extends across the full width and forms another horizontal division. The builders of Exeter were consistent throughout. In all the windows the lights are unusually short for the traceried heads above them. In the west window the great decorated circle reaches from the springing of the window arch almost to the apex, whilst the tracery of the Ughts starts well below the springing-line. Contemporary with Exeter Cathedral are the nave of York Omission (c. 1300-1338), the choir, etc., of Lichfield (c. 1330), and the pres- °^»- bytery and Lady Chapel of Wells (c. 1326-1363). All three show 3 14 A HISTORY OF ARCHITECTURAL DEVELOPMENT. a disposition towards a dual rather than a triple division of the side waUs, but only in Lichfield choir is the triforium absolutely abandoned.1 At Wells there is no string-course above the arcade, a network of panelling filling the spandrils between the arches and reaching to the siUs of the clerestory windows, The effect, although rich, is hardly satisfactory. At York the triforium is recessed, and the muUions of the clerestory windows, which are near the inside face of the waU, are continued down to the string above the main arcade to form its divisions. In Lichfield choir the design is somewhat similar, but the effect more marked, as the window is set near the outside face. The passage way above the arcade is retained, and is, moreover, pro tected by a pierced parapet, but the panelled waU behind it is flush with the windows above, the jambs of which run down to the top of the parapet. The result is two divisions, and not the three hitherto customary in large churches. The omission of the triforium undoubtedly adds height, and verticality is further emphasized in this church by the vaulting shafts, which, following the custom in Eomanesque days, start from the floor and form part of the piers.2 In the choir of York Cathedral, which is later than the nave, the triple division survives, and the triforium has a string-course above and below it. The clerestory windows are flush with the inner face of the waU, and on the outside is a pierced screen. SimUar screens are by no means uncommon in fifteenth-century Gothic.3 The earUer English custom was to have a screen on the inside, as in Lincoln Angel Choir, the windows being placed near the outside face, and this was also sometimes done in late work, as in Melrose Abbey Church, Scotland, and in the great east window of York. The design of the triforium of York choir is much the same as in French cathedrals with glazed triforia, except that the wall behind, although pierced, is not glazed. 1 In parish churches and in many abbey churches of fair size, such as Boxgrove, Sussex (c. 1235), Pershore, Worcestershire (c. 1230), there is no triforium gallery. The change recorded in cathedral ordinance is, therefore, only an adoption of what had hitherto been general in smaller churches. - At York the vaulting shafts also start from the floor, but alas I save for the tas-de-charge, which is stone, the vault is only of wood. Judging by the size of the flying buttresses outside, which are broken away, no other material would be possible. They are absurdly small as compared with those of Amiens, for instance, which has a vault of the same width. 3 In the choir of the Cathedral of Sees, Normandy, the clerestory windows have screens on the outside (c. 1330), but the design was not a favourite one with French architects as a rule. Photo : J. Valentine & Co. Fig. 216. — York Cathedral. [To face p. 314. Photo : Dr. A. O. C. Omjer. Fig. 217. — Ely Cathedral: Interior of Lantern. [To face p. 315. GOTHIC ARCHITECTURE IN ENGLAND. 315 Glazed triforia on the French model, with glass on the outside face and pierced arcading inside, were never popular in England.1 The unique feature of the fourteenth century is the lantern of Ely Ely Cathedral. In 1322 the old Norman tower at the crossing lantern- fell, bringing down with it the neighbouring bays of the choir, which at that time extended west of it into the nave. To have rebuilt the tower would have necessitated new foundations, and Alan of Walsingham, the sacrist of the cathedral, was probably afraid that the fate that befell the old one might be in store for a new. He, or some one employed by him, therefore hit on the happy idea of covering the ruined area by an octagonal lantern the full width of the church. This necessitated strengthening eight of the existing piers, a less difficult feat than building four new ones. The lantern of Ely measures about 68 feet from north to south, and about 65 feet from east to west. The canted sides are shorter than the cardinal ones, and each is pierced by a large traceried window. Light, in addition, comes from the lantern proper above the wood vault. This lantern is also of wood, covered with lead, and although stated by some to be a temporary make shift it seems difficult to believe that any more solid superstructure can ever have been intended, or that stone was ever contemplated for the vaulting below. Timber vaults in imitation of stone ones were not uncommon in EngUsh churches after the middle of the thirteenth century. At York the octagonal chapter house (c. 1300), which unlike other chapter houses in England has no central column, is vaulted in wood. Its width is only 45 feet, and if the York builders hesitated to build a vault of stone over it, how much greater must have been the disinclination of the Ely ones to use that material at a far greater height over a much wider span. The choir of York (c. 1380-1405) and the naves of Canterbury Fifteenth (c. 1379-1400) and Winchester (c. 1371-1460) form the connect- century. ing links between the fourteenth and fifteenth centuries. The fashion started at Gloucester in the middle of the fourteenth century, of perpendicular panel design for both windows and walls, was adopted generaUy, although in some cases more thoroughly than in others. At Canterbury the window tracery has vertical lines, but there is very little panelling on the walls. At Winchester, on the other hand, the west front is covered with it. This method of ornamentation reached its climax in Henry VII.'s 1 I cannot remember a single instance in which they form part of the original design, although in some churches the triforia have been glazed later. Fig. 218. — Winchester Cathedral, as originally designed. Fig. 219. — Winchester Cathedral, as altered by William of Wykeham. GOTHIC ARCHITECTURE IN ENGLAND. 317 Chapel, Westminster, where no plain surfaces are left anywhere, either inside or out. In the remodelling of the nave of Winchester by WiUiam of Wykeham (bishop 1367-1404), a great improvement was effected in the internal proportions. In the old Norman church the three divisions of arcade, triforium, and clerestory were approximately equal in height. The bishop raised his piers to the level of the floor of the triforium, abolished the triforium itself, whUst retaining a passage gallery higher up, and paneUed the wall at the sides of and below the clerestory windows. The arches of the main arcade he made four-centred, but they are not so flat as later ones of simUar shape, and conse quently do not look weak. Over the arches runs a boldly carved and moulded string-course which forms a strong division between the upper and lower storeys. The latter haff of the fifteenth century saw the beginning of the end of Gothic art in England. But what an end, what a blaze of glory for a finish ! EngUsh Gothic did not flicker and linger and drag out an inglorious life ; but still full of strength and vigour it laid itself down with majesty and pomp, with all the trappings of carving, sculpture, and colour, and surrounded by a richness and wealth of ornament such as never had been known before or has been equaUed since. Between 1450 and 1520 some of the most remarkable of English ecclesiastical buildings were buUt. Little was done to cathedrals, except to finish a tower, as at Canterbury, or to add chantry chapels, outside, as at Lincoln Cathedral, inside, between the columns, as at Winchester. Bath Abbey Church (c. 1500-1539) is the only large monastic church commenced after the middle of the fifteenth century, but extensive alterations were made to Malvern Priory Church (1450-1486) and to the Abbey Church of Sherborne (1436-1504), where the nave piers and arches are richly paneUed, and the vault a fan one. But in place of cathedral and abbey church are a number of Chapels. chapels, royal and collegiate, which if inferior in size to the earlier churches, surpass them in richness. The most famous are King's CoUege Chapel, Cambridge (1446-1461), the fan vault not being finished until early in the sixteenth century. S. George's, Windsor (1481-1537), Eton CoUege Chapel (contemporary with King's, Cambridge), and Henry VII.'s Chapel, Westminster, finished by his successor a few years after his death. The metal screen surrounding his tomb is almost the last word in really fine English mediaeval craftsmanship. The Italian sculptors were 318 A HISTORY OF ARCHITECTURAL DEVELOPMENT, knocking at the door, and the tomb itself is the work of Torregiano, the best known of the first batch of foreigners who came to our shores. Henry VII.'s chapel, with its side chapels, is almost a church BATH ABBEY CHVRCH. ELEV. OF BAY. SCALE OF io SECTlO/1 THROVGH /^AVE . IO 20 30 40 SO 60 70 FEET. The end. Fig. 220. in itself, and the same may be said of the Eoyal Chapel at Windsor. King's CoUege Chapel is a chapel pure and simple, as although it has low aisles, which form side chapels, these are cut off from the main body. The latter forms a magnificent haU, 45 feet wide, 284 feet long, and 81 feet high, vaulted with one of the most successful fan vaults in England, because it is one of the simplest. Unlike most vaulted buildings, the windows do not reach to the apex of the vault, thek heads being little higher than its springing, and the spaces over them are fiUed with sunk panel ling. The chapel might have been barrel vaulted, with but little alteration in external design, and if the treatment of the vault had been similar to that over the little chapel of Abbotsbury, Dorset, an effect perhaps finer than that now produced might have resulted. Little did the workmen who built and gloried in these GOTHIC ARCHITECTURE IN ENGLAND. 319 sumptuous chapels dream that in a decade or two an art revolu tion would take place in England, and that the methods and ideals they beUeved in and followed would be counted as barbarous. The revolution had aUeady taken place in Italy, but it is doubtful if Brunelleschi's triumph at Florence had reached their ears, and even if it had, it was probably merely regarded as " just foreign nonsense." The Pope at Eome was planning S. Peter's when the vault of King's College was nearing completion, and although the wood staUs below the latter show that a few years later the old order was giving place to new, there is no trace of foreign innova tions in the masonry of the chapel. Mediaeval art was dead and buried before the sixteenth century was a quarter over, but its spirit still walked abroad. As late as the middle of the following KiriGS COLLEGE CHAPEL CAMBRIDGE SCALvE OF 10 o 200 FEET. Fig. 221. century, the papist leanings of Archbishop Laud caused the church of S. John the Baptist, Leeds, to be built in a manner which, exter naUy, at least, recaUs the work of two hundred years before. Many seventeenth-century dates on porches, over windows, etc., show that when alterations and additions were needed, an attempt was made to harmonize the new work with the old. Small buildings, such as the chapel of the Whitgift Almshouses at Croydon, the end window of which is dated 1597, stUl had Gothic detaU, and not such bad Gothic either. At conservative Oxford, the old traditions held the field longer than in any other town — the coUege of Christ Church being a notable instance — and in outlying country districts the people continued to build their cottages, farmhouses, and small manor houses in much the same way as their forefathers had done, and with but little difference in detail. In fact, the Gothic spirit lingered in England until, strange irony of fate, it Fig. 222. — King's College Chapel, Cambridge. GOTHIC ARCHITECTURE IN SCOTLAND. 321 was destroyed by the movement in favour of a Gothic revival. An insistence on the letter of the style then cut the threads of a tradition which had descended from father to son for generations, and the Gothic and Greek revivals between them destroyed the last vestiges of vernacular art. Scotland. Brief mention only can be made of mediaeval church work in Scotland, although it possesses many examples, some intact and more in ruins, of great interest. The Irish church, which was so powerful there in the sixth century and in the one following, had sunk almost into obUvion before the eleventh was reached, and consequently exercised no influence on later work. The Eoman esque of Normandy did not reach the country until after the end of the eleventh century, and it continued later than in England, so Scotland can hardly be said to have assisted architectural develop ment to any extent at either the Eomanesque or the early Gothic period. There was considerable building activity during the thirteenth century — activity often occasioned by English mis deeds — but in the fourteenth there was a lull. During these two centuries Scottish architecture proceeded on much the same lines as in England, although often little mannerisms show the inde pendence of the country and its workmen. In the fifteenth century, the close friendship between France and Scotland, a friendship which extended for another hundred years or more, led to the introduction of a few French traits apparent in some of the late work. The paucity of examples remaining intact in the southern part of the country is mainly due to the numerous forays of the EngUsh kings. Near the border hardly a single abbey was not sacked at least once, some more often, and many were never rebuilt. The four principal abbey churches of the Lowlands — Kelso, Jedburgh, Dry burgh, and Melrose — are ruins. But they are none the less interesting because of that. In fact, their present state enhances their value to a student of architecture, because they afford him a peep behind the scenes and an opportunity of studying construction which, in a buUding roofed in and intact, is generally hidden. The four churches mentioned are excellent examples of different periods, and in them architectural pro gression can be as well studied as in any group in England VOL. 11. Y 322 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Jedburghand Dry- burgh. equally close together, except, perhaps, the famous one in the west which includes the churches of Tewkesbury, Gloucester, Hereford, Malvern, Wells, and Bath. Kelso. Kelso Abbey is the earUest. It was founded in 1128. AU the arches are semicircular, except those supporting the remains of the tower, which are later in date and pointed. The chief pecu- Uarity of the church is the exceeding shortness of the nave. This, like each of the transepts, consists merely of a single square. The monks evidently saw no reason why they should provide for the reUgious wants of the people, but were by no means niggardly in providing for their own. They made amends for the shortness of the nave by the length of the choir. Jedburgh Abbey Church is architecturally the most interesting on the Scottish border, although it has not the glamour which has been cast over Melrose. The two westernmost bays of the choir and the west wall of the nave are the oldest portions (c. 1150). The design of these choir bays is simUar to that in some of the bays of Eomsey Abbey Church. A lofty cyUndrical column rises from the ground to the springing of the arches above the tri forium, the arches of the main arcade cutting into its side. The nave of the church is an exceedingly fine example of simple sturdy work of the last decade or two of the twelfth century, rendered all the more attractive by being in ruins. The remains of Dryburgh Abbey are of different dates, but the most interesting portions belong to the latter half of the thirteenth century. The builders were evidently of two minds regarding the triforium of the church. They did not wish to follow the Cistercian example and omit it altogether, and yet they were anxious not to cramp the dimensions of their top and bottom storeys by giving it too much space. The only openings in it are foUated circles, enclosed by arched heads and moulded jambs, one circle to each bay. Melrose. Melrose suffered more perhaps than any other abbey from English incursions, and the last, in the sixteenth century, proved fatal to it. The greater part of what is now standing belongs to a rebuilding after 1385, and to additions made in the fifteenth century. The plan of the church is somewhat curious for so late a date, but is probably accounted for by the fact that the monks had neither time nor money to spend in sinking new foundations, and so in their rebuildings did not extend beyond the old. The arm east of the crossing consists merely of two Pig. 223. — Jedburgh Abbey Church: Choir. Photo: 6. W. Wilson. Fig. 224.— Jedburgh Abbey Church; Nave. [To face p. 322. GOTHIC ARCHITECTURE IN SCOTLAND. 323 aisled bays, and one square additional bay without aisles. The choir aisles really project only one bay beyond the transepts as each of the latter has an aisle on its east side. The result of the shortness of the eastern part of the church is that the choir extended three bays into the nave, as in Eomanesque churches in England and Normandy. The whole of the south aisle of the nave is Uned by chapels. This French trait in planning, which is never met with in English churches, is evidence of the strong sympathy existing between France and Scotland. There is also a suspicion of French influence in the tracery of some of the wmdows, but most of these foUow the perpendicular treatment of contemporary EngUsh ones. The web of the vaulting in the side chapels is buUt with parallel courses, and it speaks well for the skiU of the masons, and is also an argument against the necessity for vaulting ribs, that whilst many of the ribs have fallen the web remains standing. Glasgow Cathedral is the most important mediaeval church in Glasgow Scotland. It is especially famous for its crypt, or under church, ^al6' which extends over the whole of the eastern arm of the church proper, and has no equal in England, probably none in Europe. The rapid fall of the ground towards the east afforded the oppor tunity for this unusual feature. The high portion of the choir above stops short of the extreme end, and behind the high altar are two aisles, side by side, divided from one another by columns. These are low, being little higher than the side aisles of the choir, and although they stand up well outside, owing to being raised above the crypt, they have not inside the dignity and distinction which the lofty eastern transepts of Durham Cathedral and Fountains Abbey Church possess. CHAPTER XIX. GOTHIC ARCHITECTURE IN GERMANY, BELGIUM AND HOLLAND. Intro- So long as the Hohenstaufen dynasty ruled over Germany, the duction. buUders clung to the traditions of their forefathers, and, heedless of the changes taking place in the architecture of other countries, continued to build churches, as at Bamberg, Bacharach, Naum- burg, etc., practically identical with those of a hundred years before. In the Cathedral of Limburg-on-the-Lahn (c. 1210-1240) the influence of the French Cathedrals of Laon and Noyon is seen in the internal proportions, double triforia, etc., but externally the design shows Uttle trace of the new movement. When the above- mentioned dynasty came to an end in 1254, the crown was made elective, and direct heredity became more an obstacle to succession than a help. Constant wars followed between rival claimants and their supporters, and the country, for a long period, was spUt up into what were virtually a number of separate kingdoms, the larger cities also claiming an independence which they had not possessed before. The Hohenstaufen rulers had spent so much time in their Italian domains, that they had individuaUy done little to advance architecture in their native country, but under them their dominion was at least united, and one style of archi tecture was general, although it was a somewhat obsolete one. Such was not the case during the long period that followed. In France and England, Eomanesque architecture passed into Gothic by a natural and gradual process of evolution. It was different in Germany, especiaUy in the centre and the south. The problem of transforming the heavy Eomanesque art into something lighter and more graceful had been solved in France before the Teuton builders dreamt of attempting it. Original effort on their part was therefore unnecessary. In addition, the confusion in govern ment which reigned in Germany during the thirteenth century and in the following one, made a national development an GOTHIC ARCHITECTURE IN GERMANY, BELGIUM, ETC. 325 impossibility. No wonder, therefore, when the need for change became apparent on the Ehine, that the builders turned for help to the country whose art occupied the most commanding position in the civilized world. It is a thousand pities that they did so. Their forefathers had carried Eomanesque architecture so far, and had produced such exceedingly fine work, that if the old instinct had not been dead, a school of Gothic, more robust and less adventurous than that of France, might have arisen in the old empire worthy to rank with that more to the west. Even if the Germans had been content to take the Cathedral of Chartres, or of Eeims, as their model, instead of Amiens, the sudden transition would not have been so great. But Amiens loomed taU on the horizon, in more senses than one ; its fame had spread aU over Western Christendom ; it was the last word that had been spoken ; and so the good men of Cologne threw to the winds their old traditions, and German architecture started afresh on entirely different Unes. Cologne Cathedral was not quite the first church to be buUt in the French fashion, but, owing to its size and impor tance, it exercised an influence on subsequent work which no number of smaUer examples could have done. One of the earUest and at the same time most perfect of fuUy- Lieb- developed Gothic churches in Germany is the Liebfrauenkirche at kirche" Trier (c. 1240). BuUt on the foundations of a tircular buUding, Trier. by a clever disposition of side chapels, it appears cruciform externaUy, and yet preserves approximately the plan of the older building. The plan of the chapels at both ends is the same as at the east end of the Church of S. Yved, Braisne, near Soissons (France), consecrated 1216, which has an oblong nave. The church at Xanten, Germany (c. 1210) has a similarly planned east end. It is not a chevet, inasmuch as there is no ambulatory, but the chapels radiate, and the perspective effects are exceedingly good. The Liebfrauenkirche has no triforium, the columns are lofty (35 feet high), and the internal proportions throughout are excellent. This church is merely one of many which prove the wisdom of the German buUders in so frequently omitting the middle storey. The main arms are 30 feet wide and the vault over 80 feet high, very respectable dimensions. The foundation stone of Cologne Cathedral was laid in 1248, Cologne the only part of the church commenced being the eastern arm. dral.e" Even in that the work progressed but slowly, and the choir was not ready for consecration until 1322. The transepts and nave 326 A HISTORY OF ARCHITECTURAL DEVELOPMENT. were then proceeded with, but funds faUed, and, when half finished, they were covered over for service and left. The unfinished portions soon fell into decay, and it was not until the last century that the church was completed. The eastern arm is therefore the only part that belongs properly to the Middle Ages. The original design for the rest existed and was followed, but the medissval builders cannot be blamed for its present unsatisfactory appearance. The S.YVED, BRA IS/IE FRA/1CE. LIEBFRAVE/IKlRCflE, TRIER. SCALE OF IO O ISO FEET. Pig. 225. eastern arm is almost an exact replica of Amiens Cathedral. The dimensions are the same, and the design of the aisle windows of the chevet is practically identical. The plainness of the lower storey acts as a pleasant foil to the later and richer work above. Apart from the florid character of the latter, Uttle fault can be found with the east end. It is reminiscent of Amiens and Eeims, and its great width and height, its broken outline, and elaborate panelled and crocketted buttresses, produce an extremely imposing GOTHIC ARCHITECTURE IN GERMANY, BELGIUM, ETC. 327 effect. The western arm is practicaUy the same length as at Amiens ; but it has double aisles instead of single, the gigantic western towers occupy two of its bays on each side, their buttresses projecting in all directions, and each transept has one more bay than in the other church. The result outside of these alterations to the French plan is deplorable. The nave, from most points of view, is swaUowed up; only three whole bays stand clear between the aisles of the transepts and the towers (as the end bays east and west are cut into by buttresses), and the towers are so large, so n-r-t i: :|i-r-r- Fig. 226.— Cologne Cathedral. close together, and spread out so much that, except when im mediately facing them, they appear to touch. At N6tre Dame, Paris, the western towers also occupy two bays, but the fuU width of the nave is retained between them, and their buttresses are not allowed to straggle as at Cologne.1 The interior is more satisfactory than the outside, because, notwithstanding that the vault rises 150 feet from the floor, and that the double aisles of the nave detract somewhat from its apparent length, the tower-bays count in the length, the break at the crossing is not marked, and 1 Cologne shows that the Frenchmen, in keeping the towers of Amiens Cathedral comparatively low, were not so wrong after all. It is better that the nave should dwarf the towers than that the towers should completely destroy the nave. 328 A HISTORY OF ARCHITECTURAL DEVELOPMENT. the church as a whole appears long enough. The clerestory windows are of great size, and together with the triforium, which is glazed, take up more than haU of the height. The outer aisles are the same height as the inner ones (63 feet), which is un necessarily high for the outer aisles and not sufficient for the inner. If the gradation of Bourges Cathedral had been followed, the extra 25 feet in height of Cologne might have produced still finer proportions than in the French church. Other Strassburg and Metz are, after Cologne, the most important of plans. German churches designed on French lines. In the former, the extremely fine Eomanesque work at the east end had no sobering influence on the later builders, certainly none on E. von Steinbach, to whom is attributed the design of the west front (c. 1275). The nave of Metz Cathedral was built in the latter part of the fourteenth century, whUst the choir was not finished untU the sixteenth. It is amongst the most airy of continental examples. The triforium is glazed, and the huge four-Ught clerestory windows occupy the whole of the available upper part of the waUs. Both Strassburg and Metz were on the French frontier, and the western arm of the former and the whole of the latter architecturaUy belong to France, although the work in both was executed after the finest period of French art was past. Typical u js a pleasure to turn from these to two types of church in (jrtirmfiTi plans. which French influence, if apparent at aU, is in entire subjection, and the work thoroughly German. In both types most of the churches are of brick. Even the piers, mullions and tracery-bars of windows, and string-courses are of this material, which accounts to some extent for the differences between them and the German churches with French prototypes, just described. The churches of the first type follow the customary ordinance of a high nave and lower side-aisles, whilst those of the second have their nave and aisles equal, or approximately equal in height. The majority of the examples of both classes are in Northern Germany, where the power of the Hanseatic League in the fourteenth century gave an importance to towns and an independence to their inhabitants far greater than existed in either France or England. First The earliest and most important of churches of the first type is ype- the Marienkirche, Lubeck (c. 1275). There is no dependence on French models here. The sturdy brick flying buttresses are very different from the more ornate and Ughter contemporary stone ones elsewhere, and are even more effective. The church was GOTHIC ARCHITECTURE IN GERMANY, BELGIUM, ETC. 329 buUt by the townsfolk for themselves, and the plan is as simple as the general ordinance. There is no triforium, and S.A\AKY. LVDLCK V > \t 1 ./ H-.-ifl ' 1/ . ..--' !V. 11 / » ! / ' 1 1 * 1 * v 1 / \ ' / x ¦ 1 \ v ; 1 ' N * ' \ " -A * • O 50 tOO 150 SCALE OF 1 1 I 1 1 1 I I 1 1 1 FEET Fig. 227. the lofty clerestory windows are carried down to the string-course S.A\ARY, LVBECK. CROSS SECTION 5CA LE OF FEET Fig. 228. above the arcade, the aisle roof each side being a double-pitched one to allow of this. The church is worthy to rank in size with 330 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Hallen- kirchen. S.ELIZABETH MARBVRG the first-class cathedrals of Northern France, except that it is short. Still, as there are no transepts, and no break in the vault, the length is sufficient. The nave is 41 feet wide and 125 feet high, the total width between the aisle walls being about 112 feet. The aisles are 67 feet high, the same as the inner aisles of Bourges Cathedral. The plan of S. Mary's was found suitable for congre gational purposes, and in the fourteenth century was followed in people's churches at Luneberg and many other Northern towns. The favourite type of church, however, in the fourteenth and fifteenth centuries was one with nave and aisles of approximately equal height. The greater number of this class are in the North, but there are also many Southern examples. From their spaciousness and the large accommodation they afford they are termed " Hal- lenkirchen." Poitiers Cathe dral, and other vaulted churches of less importance in Southern France, had earUer been buUt on somewhat similar Unes, but nowhere was the adoption of this plan so general for large churches as in Germany. One of the earliest and best-known ex amples is S. EUzabeth, Mar burg (c. 1235-1283). The nave and aisles are exactly equal in height, about 68 feet, the central vault is covered by a high-pitched roof, whUst each bay of the aisles has its own roof of stiU steeper pitch, hipped in front, which cuts into the main roof at right angles. In its method of roofing, however, it differs from most other examples of the Hallenkirchen type, in which a single roof spans both nave and aisles. In some churches, such as S. Severus, Erfurt, and S. James, Lubeck, there are double aisles, and yet one roof suffices to cover the whole. Vienna Cathedral, S. Peter's, Lubeck, etc., also differ from the Marburg church in that in each the aisles are somewhat lower than the nave, the vault of the latter starting about on a level with the apex of the side vaults. All these churches are perforce lighted from the aisles, although small SCALE OF Fig. 229. GOTHIC ARCHITECTURE IN GERMANY, BELGIUM, ETC. 331 clerestories might have been managed at Vienna and Lubeck, if it had not been for the single roof. In fourteenth and fifteenth-century churches which have clerestories and lean-to roofs over the aisles, there are seldom triforia, notwithstanding that the space between the arcades inside and the windows above is considerable. This is especially marked at Ulm Cathedral (c. 1377). This church has an exceedingly lofty nave, nearly 140 feet high, and as the aisles are 50 feet Absence of triforia. 10 o I I ' 10 I 20 SCALE OF 5. PETER LVDECK. CR05 5 SECTION Fig. 230. wide,1 and the lean-to roofs over them unusually steep, there is a space of about 40 feet each side, between the top of the arcade and the window-sUls, which is unbroken, except by the vaulting shafts which spUt it up into bays. The German builders were ever averse to the middle storey, and much may be said for their disUke of it — a dislike which the later builders of other countries evidently shared. They omitted it generally in their Eomanesque churches, in their later churches of S. Mary, Lubeck, type, and therefore may be said to have been consistent throughout. But 40 feet of plain wall is sufficient to crush in appearance any 1 At the beginning of the sixteenth century the aisles were each divided into two by columns, so the church has now double aisles. 332 A HISTORY OF ARCHITECTURAL DEVELOPMENT. arcade below it, and this has been the effect at Ulm. In smaller churches, such as the Lorenzkirche, Niirnberg, the treatment is not open to the same objection, as the space is so much less in height. Faults. One of the chief faults of many German churches is the careless manner in which the nave and choir roofs are united. In the Lorenzkirche, Niirnberg, one roof at the east end covers both choir and aisles, and rises far higher than that at the west, which is over the nave only. At S. Sebald's, in the same town, a simUar mistake is made. In Vienna Cathedral the nave roof is the higher, and the considerable drop from it to the lower roof over the choir is exceedingly unpleasant. That the halves of the churches men tioned are of different date is no excuse. The want of length no doubt prevented transepts or a central tower, either of which would have concealed the differences in height, but some other means might surely have been found to prevent the ugly junctions which are often so marked between eastern and western arms. Choirs. The want of length referred to is owing to the shortness of the choUs. The churches were not monastic ; they were built by and VLn CATHEDRAL SCALE OF Fig. 231. for the people ; and whUst the naves are large and spacious, the choirs are short and often without aisles. The total length of Ulm Cathedral is only 416 feet, notwithstanding that its width is over 160 feet, and the height of its vault nearly 140 feet. The spirit of anti-monasticism and Protestantism was early abroad in the country ; and the long choUs of England, or the many chapelled east ends of France, were unnecessary. With the exception of the GOTHIC ARCHITECTURE IN GERMANY, BELGIUM, ETC. 333 churches avowedly buUt on French models, such as Cologne Cathe dral, the fully developed chevet was rarely adopted. S. EUzabeth, Marburg, has the primitive ending of three apses. Most later choUs terminate with a semi-octagonal apse surrounded by an ambulatory, but without radiating chapels. In Ulm Cathedral the ambulatory is omitted as weU. The choir of Erfurt Cathedral has no aisles, and their omission provided an opportunity for enor mously tall windows, which the buUders were not slow to avail themselves of. The windows are four-light traceried ones, over 50 feet high, and as the choir is raised on a lofty sub- structure their appearance is very striking. In the ruined church of S. Werner, Bacharach, on the Ehine, are others nearly as tall, and in the Cathedral of Aix-la-ChapeUe, the choir windows are actuaUy 80 feet in height. If the masons of Germany were inferior in artistic feeUng to Manual their brethren of France, in manual dexterity they certainly sur- dexterity- passed them. They wrought in stone results as wonderful as those produced by the Chinese in ivory. The interpenetration of mould ings, if not their invention, was certainly practised by them to a far greater extent than it was elsewhere. The pierced spiral staircases on the outside of the towers of Ulm and Strassburg Cathedrals are marvels of technical skUl. Their canopy work, for mere ornateness, has no equal. They delighted in crockets, finials, pierced para pets and buttresses ; but their great tour-de-force was the pierced spire. Openings in spires are, to some extent, necessary for Ught, and when treated simply as windows, with a gable over each, as in England, produce excellent results. Most telling, also, are the simple quartrefoUs and trefoils, pierced at intervals, through many of the early spires in France, as at S. Pierre, Caen, etc., but the effect of these is altogether different from that produced in many late German spires, which form simply a network of pierced tracery. Freiburg Cathedral has one of the earliest of this type (c. 1300-1350). Germany stood almost alone amongst continental countries in single its preference for a single tower at the west end of a cathedral, westem instead of the pair customary in France and England. Examples of it are at Ulm, Freiburg, Frankfurt, etc. Cologne and Strassburg Cathedrals have two western towers. In the FrauenkUche, Ingol- stadt, Bavaria, the pair are placed diagonally at the end of the aisles, as was done originaUy at S. Ouen, Eouen. A single tower possesses this advantage, there is a probability of finishing it. The French, 334 A HISTORY OF ARCHITECTURAL DEVELOPMENT. in the thirteenth century, discovered that a church could have too many towers ; and the Germans, in the fifteenth century, showed their common sense in restricting theU ambition to one. Even then most of the spires crowning towers remained unfinished for a century or more, and some have only been built in modern times. For a church to have one tower which can rank amongst the highest in the world is a distinction: to have two is an extrava gance. If the Ulm plan had been adopted at Cologne, the gain to the cathedral would have been enormous. In Vienna Cathedral the towers come above the transepts, and in this example alone one feels that a pair of towers of such massive proportions are possible, because they are separated from each other by the entire width of the church. The southern one, which alone is finished, ranks amongst the finest efforts of the German builders. The VIE/1/1A CATHEDRAL. so o ' I I ' ' I SCALE OF FEET. T.T.TOWER5 Fig. 232. junction between tower and spire is not marked at all, and a mag nificent pyramidal effect is produced, the complicated arrangement of buttresses at the angles helping the Ulusion that the spire springs direct from the ground. From the beginning of the thirteenth century to the end of the fifteenth, Belgium was one of the most flourishing countries in Europe. Bruges was chosen by the representatives of the towns GOTHIC ARCHITECTURE IN GERMANY, BELGIUM, ETC. 335 composing the great Hanseatic League as their northern depot, and Ghent, Antwerp, Ypres, etc., were also important commercial centres. With the exception of the nave and transepts of Tournai Cathedral (see Fig. 140), Belgium has no Eomanesque work of especial merit, but it possesses many churches, chiefly of the four teenth and fifteenth centuries, well worthy of attention. Most are principaUy of brick — even the vaults are often of that material — but several are in stone. Ypres Cathedral is one of the most important of the latter, the stone used being mainly a local blue- grey one. Interesting though the churches are, they possess no distinctive characteristics which call for detaUed mention. As a rule they are fine in scale, simple in design, although some are disfigured by later rococo figures and other additions of the sixteenth century. The majority are based on French examples, whilst a few show the influence of Germany. Ypres Cathedral (c. 1225) — its west tower is later — and the church of Notre Dame de Pameleh, Audenarde (c. 1240), are amongst the earliest in which the Gothic feeUng of Northern France is apparent. Work of the fourteenth and fifteenth centuries is well represented by the churches of Bruges, Ghent, Brussels, Antwerp, etc., the last an overgrown monster with its original plan spoUt by additions. The secular buildings of Belgium are more interesting than Seoular the churches, but cannot be considered in this volume. Mention, m however, may be made of one, the Cloth HaU at Ypres, because it is the finest example of secular Gothic in Europe. For noble simpUcity it has no equal. Inside, its upper floor, which is the principal one, forms one huge chamber covered by a simple open timber roof. The outside expresses well the internal plan. It consists of a long, low, unbroken facade, 440 feet in length, perfectly simple in design, but fairly rich in detaU, crowned by a steep-pitched roof. In the centre rises a grand tower without buttresses, but with battered sides, and at the corners of the top storey are the angle turrets which the Belgians and Germans delighted in. SimUar turrets rise above the parapets at the two ends of the buUding. Altogether, this hall has the purity of a Greek temple, and no classic architect ever designed anything more absolutely symmetrical. The chief lesson the churches of Holland teach is that colour Holland. is not essential in Gothic art. Nearly all are whitewashed through out inside with surprisingly fine results. Without in any way depreciating the fine colour schemes of the mediaeval artist, which 336 A HISTORY OF ARCHITECTURAL DEVELOPMENT. THE CLOTH MALL YPRES. BELGIVA\ Fig. 233, GOTHIC ARCHITECTURE IN GERMANY, BELGIUM, ETC. 337 often included everything in a church, it is a question whether the best effects are not obtained by using colour very sparingly, and massing it in a few special places. Too Uttle stained glass remains in HoUand to enable one to judge for a certainty what the effect would be of a church whitewashed throughout, with the windows a blaze of stained glass, and no colour visible elsewhere except perhaps on the reredos over the altar ; but it would probably be an exceedingly fine one. One has the glass in Chartres Cathedral, but not the whitewash ; in Holland one has the whitewash, but rarely the glass. White is such a marvellous foil and frame to colour, that although the Protestants of Holland and other countries certainly did not use it for aesthetic reasons, it is a mistake to take for granted that their application of whitewash necessarily had the damaging effect which is commonly attributed to it. VOL. 11. CHAPTER XX. GOTHIC ARCHITECTURE IN ITALY. Intro- That Gothic feeUng never penetrated deeply into Italy is not duction. £0 kg WOndered at when the earlier monuments of the country are remembered. The remains of the marvellous buildings of the Eomans stood in all its cities, even more numerous in the Middle Ages than they are at present. In Eome the basilican plan and ordinance had for so long been accepted as the only one for churches, that it naturally tinctured some of the subsequent work. At Pisa and the neighbouring cities, a school of buUding, based partly on basilican and partly on Byzantine models, had been perfected, and this survived weU into the thirteenth century. In the Church of S. Caterina, Pisa (c. 1250), the facade is simUar to that of the cathedral — even to the extent of the niches not centring over one another — the only difference being that the arches of the upper arcading are pointed. The crowning feature of Byzantine architecture, the dome, appears in nearly all Eomanesque Italian churches, and is not entirely absent from those of later date. At Siena, the dome over the hexagonal crossing was finished about 1265 ; at Florence, in the fourteenth century, the east end of the cathedral was planned to be covered by a dome, although the present one was not buUt until the fifteenth ; and at S. Fetronio, Bologna (c. 1390), a dome equal in size to the Florentine one was contemplated, but never commenced. Early in the fifteenth century the Eenaissance began at Florence, one hundred years before it obtained a sure footing in northern countries. The result of the early curtail ment of the Gothic movement in Italy, coupled with its late introduction, is that Gothic art in that country is practically compressed into two centuries only. Even during that time it never flourished unadulterated. Classic art was not dead ; it only slept. The semicircular arch was never abandoned. In the middle of the fourteenth century the nave of the Cathedral of GOTHIC ARCHITECTURE IN ITALY. 339 Lucca (S. Martino) was rebuilt with arches of that form. The Loggie del Bigallo (c. 1360), and dei Lanzi, Florence (c. 1375), have also semicircular arches ; and so strongly is classic feeling manifest in the detaUs and other features of these buildings that one hardly knows whether to regard them as evidence of lingering traditions, or as forerunners of the new movement. No country but Italy could have produced the Pisani, father and son, and Giotto. Niccolo Pisano, who Uved in the middle of the thirteenth century, ignored to a great extent the work of his immediate pre decessors, and took antique forms as his models; Giovanni, his son, relied more directly on nature ; but there is little U any trace in the designs of either of Gothic inspiration. Of Giotto's work as a mural painter it is not necessary to speak here. That was his metier, and in that he excelled. As an architect he was less successful. His Campanile at Florence has beautiful detail and richness of colour, but hardly satisfactory proportions. The top- heaviness of the upper part may not be his fault, as he died before the tower was far advanced and it was finished by others. The churches in Italy wliich bear the closest resemblance to Early the Gothic of the North were most of them built in the first half Gotnioexamples. of the thirteenth century. In some, foreign masons were employed, as at S. Andrea, VerceUi (c. 1220), the designer of which is said to have been an EngUshman. The clustered piers and subordinated arches are not ItaUan, but on the other hand, the tall octagonal cupola over the crossing shows that the designer, whoever he was, was not aUowed to depart entirely from local customs. The double church of the Franciscans at Assisi also has indications of outside influence. The lower church was commenced in 1228, and the upper one is said to have been finished in 1253, but the differences in design between the two suggest a greater interval of time. The semicircular buttresses on the outside waUs of the nave are reminiscent of Albi Cathedral, but if the dates given above are correct, the ItaUan church preceded the French one. S. Maria sopra Minerva, Eome (c. 1280), is the only Gothic church in the capital. In Venice, the church of the Frari (c. 1260), is superior to the contemporary SS. Giovanni e Paolo, but neither is reaUy satisfactory from a Gothic standpoint. In most of the above churches the arches are of two or more Piers and orders and the piers are clustered, but the Italians never took af°hes' kindly to the system of subordination which is one of the most important principles of northern art. In S. Anastasia, Verona, 340 A HISTORY OF ARCHITECTURAL DEVELOPMENT. (c. 1280), SS. Giovanni e Paolo, Venice (c. 1260), S. Croce, Florence, etc., and in some later churches, the columns are cylindrical, and the arches unmoulded and have flush soffits. When piers are substituted for columns, as in Florence Cathedral (c. 1300), they are generaUy merely squares, with flat pUasters of sUght projection at the sides. The typical fourteenth-century Italian capital is perhaps the ugUest ever devised. Founded on that of the Corinthian order, it consists of three tiers of acanthus leaves Internalordinance. S.AA1ASTASlA,VERO/1A. TRA/1SVER5E ArtD LCWIGITVDl/IAL SECTIONS. SCALE OF IO 10 20 _J 30 40 _l !_ 30 I IOOFEET. I Fig. 234. placed stiffly one above the other, surmounted by a moulded abacus which over columns is generaUy octagonal in plan. Un satisfactory when there is a column below, it is particularly disagreeable when crowning a pilastered pier as at Florence. Little fault can be found with the carving itself— the ItaUans were always good sculptors— but the design is flat and un interesting. Few ItaUan churches have the division customary elsewhere of arcade, triforium and clerestory. The nave of S. Martino, Lucca (c. 1340), has a triforium of considerable size, with a string-course GOTHIC ARCHITECTURE IN ITALY. 34i above and below it, and each bay consists of many openings as in Northern work, but this was a re-building. In S. Anastasia, Verona, there is merely a series of circular openings to mark the middle storey, and in Como Cathedral are pairs of openings which, however, together with their surrounding frames, are later in date and Eenaissance in character. In the Cathedrals of Florence, Orvieto, and Siena, and in S. Croce, Florence, corbelled out gaUeries run round the naves immediately above the main arcades. These galleries, Uke the entablatures over the arches in basilican churches from which they were derived, form strong horizontal FLORENCE CATHEDRAL. TRA/1SVERSE SCALE OF 10 o ArtD LO/IGITVDiriAL, SECTIONS. 50 lOO 150 200 FEET, Fig. 235. Unes, which are strangely out of character in the two Florentine churches, in which the arches below are sharply pointed. At Orvieto the arches are semicircular. In this church, and also in Siena Cathedral, the waUs and piers inside are faced 'with alter nating courses of black and white marble, which in themselves form horizontal Unes, so that the more strongly marked lines of the gaUeries are not too pronounced. Notwithstanding that triforia are either omitted altogether or are smaU and insignificant, Uttle attempt was made to give prominence to the clerestory windows. Siena Cathedral is an exception. The windows there are large and have traceried heads, as in northern examples. In 342 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Orvieto Cathedral, and in S. Croce, Florence, each bay has a lancet light which, although tall, is very narrow. The most usual opening is merely a circle, sometimes quite plain, as in Florence Cathedral, sometimes cusped, as in S. Petronio, Bologna, and S. Anastasia, Verona. It cannot be said that the result in either case is satis factory, but in Italy the strong light rendered large windows un necessary, and there was not the keen desire for stained glass which existed in the north. The effect inside would be better in many cases if the walls were painted (as doubtless they were originaUy intended to be), as then the centre of interest would be the mural decoration, and not the architecture which is often uninteresting. The extent to which paint helps ItaUan churches, with their few parts and simple Unes, is well seen at S. Anastasia, Verona, the church that excels aU contemporary ItaUan examples in the beauty of its interior. The painted scroU decoration, on a background of white, on the vaults, soffits of arches, etc., gives scale to its fine proportions. This work (c. 1437) is later than the fabric, but there can be no two opinions about its value and intrinsic beauty. Scale of The Italian of the late Middle Ages was often careless about construction, and preferred, as in Eomanesque days, to hold his arches in position by means of tie-rods rather than by proper abutment. But in many churches of the fourteenth century con siderable constructive skiU is shown, and the spans of arches and vaults are far greater than in the churches of either Northern France or England. The naves of S. Croce and Orvieto Cathedral have ordinary timber roofs, simUar to those in the basilicas of Eome. S. Fermo Maggiore, Verona, S. Stefano, Venice, and one or two other examples, have curious trefoU, or cinquefoU, wood paneUed ceilings, Uke that which, in the fourteenth century, was placed over the Eomanesque church of S. Zeno, Verona. With the exception of these and a few others, the large churches are vaulted throughout. The vaults are simple quadripartite ones, and the only peculiarity about them is their size. In the basilican churches of Eome the nave columns are numerous and close together. The fourteenth-century ItaUan architects went to the other extreme. Notwithstanding the great width of naves, the side bays are also exceedingly wide and the supports in conse quence few in number. The old Eoman plan of approximately square bays of vaulting was the one generally adopted ; the oblong form favoured by the Gothic builders of the north being the Fig. 236. — S. Asastasia, Vebona. [To face p. 342. GOTHIC ARCHITECTURE IN ITALY. 343 exception and not the rule, save in the aisles.1 (See Fig. 51.) Few supports, no doubt, possess the advantages of not interfering unduly with congregational worship, but greater reduplication gives better scale. Partly because of this, and partly owing to FLORENCE CATHEDRAL. choirs. Pig. 237. the absence of triforia and the simpUcity of the upper windows, Italian churches rarely convey an impression of then real size. In the eastern arms, the influence of the early basiUcan Plan of church plan is very marked. The transepts are at the extreme east end, and there is nothing beyond them save chapels. In S. Anastasia, Verona, and SS. Giovanni e Paolo, Venice, there are five chapels with apsidal ends, the centre one in each church being wider and longer than the side ones, being, in fact, the width of the nave. In S. Maria Novella, Florence, there are also five chapels simUarly placed, the only difference being that aU end square. S. Croce, Florence, has eleven chapels, the centre one ending with a semi-octagon, the others being rect angular. In this church the central chapel and the two flanking it are together equal to the width of the nave, thus giving an opportunity for effective grouping of a lofty chancel arch with flanking smaller side arches, over which are windows. The fully- developed French chevet is unknown in the Gothic churches of 1 Milan Cathedral is the most important exception. 344 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Exteriors. Italy. S. Petronio, Bologna, would have had the nearest approach to it if the choir had ever been buUt.1 In Milan Cathedral the German plan of an octagonal apse surrounded by an ambulatory, but without chapels, is followed; and the transepts occupy a similar position to those in French cathedrals, and project, as they mostly do, one bay each side beyond the aisles. In the designs for their west fronts, the ItaUans paid no more attention to the sectional outline behind in their late churches S.AAlASTASIA,VEROMA. ¦p iM-t %^ SCALE OF 10 o Pig. 238. than they had done in their earUer ones. They disliked it in elevation as thoroughly as their northern brethren did, but they rarely adopted the only reaUy satisfactory way of masking it, viz. a paU of towers. They preferred either a wide single gable, following the example set at S. Ambrogio, MUan (where, however, the roof behind is also in one span), or else three gables, the side ones being considerably lower and smaller than the central one. The latter is the design at Siena and Orvieto. These gables seldom bear any relation to the roofs behind them. At Siena, for instance, the nave roof is of flat pitch, whUst the central gable is very steep and rises far above it, and the side ones are but ornamental adjuncts. The west front is in fact simply a screen, rich In carving and sculpture, but beautiful only because of the marble with which it is faced. Other fronts are nearly as elaborate, 1 If this church had been finished it would have been the largest in the world, with a length of nearly double that of the present nave, and with deeply-projecting aisled transepts. Pig. 239.— Siena Oathedbai : West Fbont. [To face p. 344. GOTHIC ARCHITECTURE IN ITALY. 345 especiaUy in Tuscany. The sides are often plain brick, but as the churches seldom stand free, it matters little what material is used there. In northern Italy the fronts are not so rich as In the centre and south, and marble, when used at aU, is banded with brick. But the churches lose little by this, and gain considerably through more attention having been paid to the design of other parts. There is never the glamour about the outside of Italian churches that there is about French ones, but there is often considerable stateliness due to their size and vast expanse of plain waUing. Even in their vaulted churches the ItaUans were able to dispense with flying buttresses (except in Milan Cathedral), and to some S CROCE, FLORENCE. SCALE OF Fig. 240. extent with ordinary buttresses, because of the smallness and paucity of the openings. The walls are nearly soUd, and conse quently are capable of resisting almost unaided the thrusts of the vaults, notwithstanding their great span. At the same time, few vaulted churches are entirely without external buttresses, and very effective some of these are owing to their absolute simplicity and absence of set-offs. In S. Anastasia, Verona, the brick buttresses of about 3 feet square projection rise sheer from the ground to above the eaves without a break. The conservative instincts of the northern Italian are shown in the fine tower of this church, which is only a little different from the north tower of S. Ambrogio, MUan, built two hundred years before. MUan Cathedral (c. 1386) deserves a special word of descrip- Miian Ca. tion, because, besides being the largest of all Italian cathedrals, thedral. 346 A HISTORY OF ARCHITECTURAL DEVELOPMENT. excepting St. Peter's, it is beyond doubt by far the most remark able. A double-aisled church, it follows Bourges Cathedral in the gradation of its aisles, but is without triforia. All the roof's, which are covered with white marble tUes, are of low pitch. This accounts for the small amount of wall space inside between the tops of the arcades and the clerestories in both nave and inner aisles, which enabled its designers to dispense entirely with the middle storey. The clerestory windows themselves are smaU and insignificant, and although objection has already been taken to the circles which commonly constitute clerestory windows in Italy, Fig. 241. it is a question whether they are not preferable to the little traceried Ughts of Milan. The side windows in the aisle walls are of great size, and from these most of the Ught in the church is obtained. The tracery of their heads, especially in the windows at the east end, is far more fantastic than any to be found in France or England, although in S. Petronio, Bologna, the designs are equally unconstructional. One curious fact about Milan Cathedral is that the top tier of windows does not light the interior at all. The nave vaulting is very domical, and these windows are above the vault, and merely light the space between it and the tiled roof above. What the effect would have been if all the ridge-ribs of the vaulting had been level, thus allowing the clerestory windows to rise considerably higher than they do at - ^»*sES2r~* GOTHIC ARCHITECTURE IN ITALY. 347 present, is difficult to say, but inside the improvement would probably have been immense. The Italians, however, did not want light, and yet they had to raise the side walls above the apex of the transverse arches of the central vault, otherwise the construction of the timber roof over would have presented diffi culties. A Frenchman, obUged to keep his vault domical and his windows below it smaU, would probably have left the upper niLAn CATHEDRAL TRAA15VER5E AAID SCALE OF IO O 50 I 1 i i i ' i LOriGITVDI/IAL 100 150 SECTIO/1S. 200FEET. Fig. 243. part of the wall unpierced, or have pierced it with very smaU openings. From the inside, of course, it makes no difference what is there, and very little from the outside, as these windows hardly show, owing to the height of the aisle walls and the crowd of pinnacles, flying buttresses, and pierced parapets in front of them. The outside of MUan Cathedral must be seen to be understood. Full of faults, with detail coarse in the extreme, its mass, its pinnacles, and the white marble of which it is built throughout, 348 A HISTORY OF ARCHITECTURAL DEVELOPMENT. render it most striking. A rare lesson may be learnt by a com parison of the cathedral and S. Ambrogio in the same city, the one absolutely quiet and simple, the other all shout and glitter and finery. The inside is far finer than the outside. The lofty piers, crowned by font-Uke shaped capitals, enriched with figures over life-size — the capitals are about twenty feet high — the great width and height of the church, the vistas across it, the dazzling rays of sunlight which, at morning and evening especially, stream through the great aisle windows, all combine to make it one of the most marveUous interiors in the world. Much has been said in abuse of the painted vault, but inasmuch as it should deceive no one, it ought to displease no one. Nobody who knows anything about vault construction can suppose for a minute that the lines he sees on the vault are structural ; that they are any thing but painted. It is not a high form of decoration perhaps, but it helps the general effect, and certainly does not detract from it. CHAPTER XXI. GOTHIC ARCHITECTURE IN SPAIN. Spain exercised little influence on the early development of Intro- mediseval art, because the whole of the country, with the excep- duotlon- tion of the mountainous district near the Pyrenees and the north west part along the Bay of Biscay, was in the hands of the Moors until the eleventh century was far advanced. The triumph of the Christians commenced with the recovery of Toledo in 1085, and this was foUowed by other successes, until the whole of the northern haU of the country was in their possession. SeviUe was not recaptured untU 1248, and the Moors' last stronghold, Granada, not until 1492. Street, in his " Gothic Architecture in Spain," mentions two churches in Barcelona which were probably buUt in the tenth century, and some eight or ten which belong to the foUowing one, but with the exception of the great cathedral at Santiago de ComposteUa, the Church of S. Isidoro, Leon, and a few others, most Spanish churches are later. In one respect Spain was fortunate. It was regaining its freedom at the time when mediaeval art was advancing rapidly in other countries, and although the constant fear of Moorish inroads prevented the estabUshment of a national school in Spain itself, the Spaniards were wise enough to turn for assistance to the buUders of the many fine churches to the north of the Pyrenees, which had either just been finished or were in course of erection. The eleventh- century churches are mainly of two types, both First type of which were borrowed from Southern France. S. Pedro, Huesca °,f earJv church. (c. 1100), may be taken as typical of one. In it the internal ordinance is the same as in S. Nazaire, Carcassonne. Barrel vaults start a few feet only above the crown of the side arches, and there is no room for either triforium or clerestory. The plan of S. Pedro is very simple, and consists of an aisled nave of five bays, with three apses at the east end — the central one being a little longer than the side ones — covered by semi-domes. 350 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Four of the western bays are barrel-vaulted, but over the fifth and easternmost, which is wider than the others, is a lantern. This last feature, the idea for which was doubtless taken from S. Ambrogio, Milan, or from one of the many churches in Northern Italy, Germany, or Southern France, in which simUar lanterns occur, is found in nearly all early Spanish churches, and in some it is treated in a distinctly original manner, as wUl be shown later. Churches Churches of the second type follow S. Sernin, Toulouse, and of second many churches in Auvergne, etc., in being without clerestories, and in having their barrel vaults raised weU above the main arcades so as to allow space for roomy triforia. The vaults are semicircular, and are divided into bays by transverse arches as in the French examples. Some churches of this type, such as S. Isidoro, Leon, follow in plan S. Pedro, Huesca, with slight modifi cations, whilst others are markedly cruciform, and have fully developed ambulatories and chevets at the east end instead of three separate apses. The finest amongst the latter is the Cathe dral of Santiago de ComposteUa (c. 1078-1128). In general ordinance and plan it is a repUca of S. Sernin, Toulouse, built some twenty years before (see Fig. 157), except that the nave has only single aisles, instead of double, the transepts, however, projecting one extra bay each side by way of compensation. The examples mentioned are sufficient to show how dependent the Spaniards were at first on French models. The style on Spanish soil was an imported one, but it took root there and continued to flourish long after it had been abandoned elsewhere in favour of fully developed Gothic. The old Cathedral of Salamanca (c. 1120-1178), the Cathedral of Tarragona (c. 1131), and the church at Lerida (c. 1203-1278), differ only from earUer churches in that the principal arches are pointed, and in the substitution of ribbed vaults for barrel ones. In one respect the Spaniards improved on French work. No Eomanesque church in France — or for that matter in Italy or Germany either — has so effective a central lantern as can be seen in the old Cathedral, Salamanca, in Zamora Cathedral, or in the Collegiate church at Toro. S. Sernin, Toulouse, has a taller tower ; the church of Chiaravalle, near Milan, one of more pecuUar shape ; and many German churches have cupolas of wider span, but none is so original as the three Spanish examples mentioned. Over the crossing at Salamanca is a two-storied arcaded drum, carried on pendentives, and pierced with windows on its cardinal sides. Above is a dome, with ribs on its intrados GOTHIC ARCHITECTURE IN SPAIN. 351 which start from shafts between the bays of arcading of the drum. Outside, the effect is even more striking than inside. The drum forms a sixteen-sided tower, on the four principal canted sides of which are large circular turrets which help to resist the thrusts of the dome. The inner dome is covered by an outer one of steeper pitch, which has more the appearance of a strongly entasised low SALA/nAMCA OLD CATHEDRAL. . I/ITERIOR OF LAHTEK/I . EXTERIOR OF LA/1TER/1. Fig. 244. stone spire than of a dome. At Toro the design is very similar, but in place of the outer dome is now a low-pitched roof. Nearly contemporary with the church at Lerida, are the introduc- Cathedrals of Toledo (c. 1227), Burgos (c. 1221), and Leon ^°£ (c. 1250), but they show a very different feeling. The fame of Gothic. the cathedrals of Northern France had reached Spain, and all three were evidently designed, and doubtless partly built by French masons. Here, as in Germany, one has to regret that the people did not try and develop the style they had become accustomed to, instead of merely reproducing what had been done 352 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Churcheswithout aisles. in another country. To such an extent was Northern French Gothic the rage that, at Leon Cathedral, in the latter half of the thirteenth century, the west porch was buUt in exact imitation of the famous porches in front of the transepts at Chartres. Of this cathedral Street says, "The church from beginning to end is thoroughly French. French in its detail, in its plan, and in its general design." Toledo Cathedral, a church with double aisles, is the largest of the three, its width being as much as 178 feet, which is many feet wider than either Cologne or Bourges Cathe dral, and is excelled by some 12 feet only by MUan Cathedral. The most interesting Spanish churches of the fourteenth cen tury are those in which the Spaniards returned to their first love — C5ERO/1A CATHEDRAL. (from street). •lSOjrgET. Fig. 245. the architecture of Southern France- great aisleless churches of Albi and Toulouse and took as their models the At Barcelona the Church of S. Maria del Pino (c. 1329) has a nave 54 feet wide, with chapels on either side, but no aisles. Although the Spaniards took the plan from France, aU French churches of this type are excelled in size by the great cathedral at Gerona. The east end was buUt first (c. 1312-1346), and in itself presents nothing of particular interest, being divided in the ordinary way into central and side aisles with radiating chapels round a semi circular end. The great feature of the church is its aisleless nave of four bays — with side chapels as in other churches of similar plan — which was commenced in 1416. This actually has a span of 73 feet, exclusive of chapels, and a height of 117 feet. No vaulted buUding of such a width had been attempted since GOTHIC ARCHITECTURE IN SPAIN. 353 Constantine built his basilica at Eome.1 The side chapels follow more the Perpignan plan than the Albi one, inasmuch as they are comparatively low. Over the arches in front of them runs a small triforium with big clerestory windows above. The design is thus more commonplace than Albi Cathedral, the lofty side vaults of which, towering above the galleries, produce such splendid effects. The triple division of the earlier eastern half, however, makes a fine ending to the nave, and the piers and arches and windows over them give scale. The fault of Albi is that there is too little in the structure itself to enable one to grasp its size, and if it were not for the screen round the choir, the inside would not look even as large as it does. The idea of adding a nave without aisles to a choir with them was not a new one, as it had been done two or three centuries before at La Trinite, Angers, S. Eadegonde, Poitiers, etc., but these churches, although by no means small ones, are far inferior in size to the Spaniard's masterly tour-de-force. The problem to be solved at Gerona was how to provide space for a large congregation, aU of which could have an uninterrupted view of the high altar, and in no other example has this been more successfuUy accompUshed. At Barcelona (besides the Church of S. Maria del Pino already mentioned) is the still larger S. Maria del Mar (c. 1328), designed with the same aim in view, but in a different manner. The church is an aisled one, and consists of a nave of four great bays, each over 40 feet square, with a narrow eastern bay and an apse beyond surrounded by an ambulatory. There are consequently only four piers each side, and these naturaUy offer Uttle obstruction. In the number and size of its bays the church resembles some of the late ItaUan examples, such as S. Petronio, Bologna, Como Cathedral, etc., and in general ordinance it also foUows these, as it is Ughted mainly from the aisles, the clerestory windows being small and consisting of small traceried circles, one to each bay. Whilst the two Barcelona churches mentioned, and some others in the town, were planned for congregational worship, the cathedral, on the other hand, was especially arranged for private devotion. There are as many as twenty-seven chapels inside the church, arranged round the 1 A few domed Byzantine buildings are wider, but no mediaeval ones. Street gives in an appendix to his " Gothic Architecture in Spain," an extremely interesting account of the fears which this bold project raised amongst the Chapter, and of the Junta of architects it called before it would allow Guillermo Boffiy, the master of the works who proposed it, to proceed with his plan. VOL. II 2 A 354 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Position of choir. chevet and opening out of the aisles of the nave, and, as though these were not sufficient, no less than twenty- two more are grouped round three of the sides of the cloister. One peculiarity in Spanish churches deserves mention. The choir is generally placed west of the crossing in the nave, and is cut off entirely from the altar, except that a passage-way is BARCELONA CATHEDRAL, AflD CLOI5TERS. FROM STREET. Pig. 246. screened off for the priests to pass from one to the other. The plan doubtless originated in the short Eomanesque choirs which did not allow sufficient room for the clergy after their numbers multiplied. In all countries extra provision had to be made for them, but in none, except Spain, was the altar ever divorced from the choir. The reason why it was done in that country is not clear, and Street implies that it is " a comparatively modern innovation," dating in most cases from the early days of the Eenaissance. The people are admitted to the transepts in all the GOTHIC ARCHITECTURE IN SPAIN. 355 large cathedrals, the nave west of the choir being, as a rule, little used. In S. Pedro, Huesca, the choir is actually placed against the west wall of the church, occupying one bay only, and is separated by four bays from the apse containing the altar.1 The great cathedrals of Salamanca (the new church) and Sala- Segovia, are the last words of Gothic in Spain. Salamanca ™l°a Cathedral was not commenced until 1512, and considering the Segovia. date it is wonderfully mediaeval in feeling. Segovia Cathedral is even later. In 1522 the architect of Salamanca Cathedral was appointed to superintend the work, the greater part of which was finished some fifty years later. Both are big churches of con siderable width — over 110 feet exclusive of the side chapels— but no great length, proportions that are general throughout the country, width being invariably regarded as the chief desideratum. SeviUe Cathedral is actually 247 feet wide, if the side chapels are included, whUst its total internal length is only about 430 feet. What strikes one most in these late churches is the enormous Height of height of the piers. In English cathedrals, where piers are pier3, notoriously short — often unpleasantly so — a height of 30 feet is rare, and is barely met with in Westminster Abbey, and in some of the later examples in which there is no triforium, such as Win chester. In Northern France, piers are higher, but not so high as Spanish ones. In Amiens Cathedral they are 48 feet high, being one-third of the total height. Even Milan Cathedral, with its height of 157 feet, has piers only 55 feet high. In Poitiers Cathedral, in the Hallenkirchen of Germany, and in the Church of the Jacobins, Toulouse, with its central arcade as high as its vault, the piers are exceedingly lofty, but it is hardly fair to compare these with the Spanish churches, as they have nothing above the arches, neither triforium nor clerestory. But even taking these for comparison, not many will be found to have taller piers. The most extraordinary instances of this peculiarity are at Sala manca and Segovia. These two churches are very similar in thek dimensions and internal ordinances. The total height of each is about 120 feet to the apex of the vault, nearly one-half of which is pier. Sixty feet is about the full height inside of some English cathedrals, which are therefore little higher than these piers. Moreover, the greater portion of each pier is carried up to the springing of the vault, another 34 feet or so in the case of the 1 The arrangement in Spain is much the same as that now existing in West minster Abbey, except that the distance from the choir to the altar is greater. 356 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Cathedral of Segovia, without any break whatsoever. At Sala manca the capitals at the springing-line of the side arcades are continued round the piers, and consequently in this church the height is broken to a small extent ; at Segovia only the members BARCELONA CATHEDRAL. WEST EA1D OF nAVE . FROM STREET. Fig. 247. of the piers immediately under the arches at this point have capitals, the other members which form the vaulting shafts — and they are exceedingly large and numerous — having none. Late Spanish churches, like Italian ones, rarely have more than the two divisions of arcade and clerestory. The roofs over the GOTHIC ARCHITECTURE IN SPAIN. 357 aisles are, as a rule, absolutely flat, and consequently no wall space inside was required. In Barcelona Cathedral there is a kind of triforium, consisting of a passage-way with a band of arcading in front, which separates the arches from the windows above, but this starts above the tas-de-charge of the vaults and cuts into the waU ribs. There is much to be said for a gallery in this position, both structurally and aesthetically. There is no weakening of the abutments below, no curtailment of the height of piers, and no difficulty in providing a well-lighted passage-way all round the church and easy access to the roofs at any necessary point. CHAPTER XXII. ENGLISH PARISH CHURCHES— TIMBER ROOFS. Of all countries England is the most remarkable for the number, beauty, and variety of its parish churches. They cannot compare in size, as a rule, with those abroad, but they have an individual distinction of their own. The great number of country examples we possess is due to our insular position, which saved our fore fathers from the ever constant fear of forays from neighbouring states that kept the continent in a continual state of unrest. The whole of England was dotted with vUlages, and each had its own church. In some cases the vUlage has disappeared and only its church remains, as at Magdalen Laver, Essex. Often the villages touched one another, like Pevensey and Westham, Sussex, where two old churches are within a stone's throw of each other. In France, Germany, and other countries, villages were few and far apart. The people, for safety, were forced within walled cities. The farmers and agricultural labourers who Uved outside had, in many cases, to trudge far to church; as in Brittany at the present day, where in many parts the church of a village serves a large outlying district. At S. Thegonnec, near Morlaix, and elsewhere, the peasants on Sunday mornings can be seen on the roads for miles round, walking to mass, their coats over their arms U the time be summer. Not only are there more country churches in England, but the number in towns is, or at all events was, far greater. Norwich at one time boasted sixty-one churches, and York nearly as many. The majority were small, but to this is owing much of their interest. Their con struction and general ordinance differ in many respects from those of the larger churches, both EngUsh and foreign, already described. Almost all have open timber roofs. In France, the majority of town parish churches are vaulted, and there is nothing in their design to distinguish them from cathedrals ; in size, even, they are often little inferior. S. Mary Eedcliffe, Bristol, is one of the few ENGLISH PARISH CHURCHES — TIMBER ROOFS. 359 in England of which the same may be said. Its plan, with aisled nave and choir, projecting transepts (each with east and west aisles), and Lady Chapel, its elaborate vault, and the internal divisions of its nave and choir walls, belong to a cathedral more than to a parish church. The number and smallness, as a rule, of EngUsh town churches are early evidence of that independence in reUgious matters which occasioned later the caustic saying of Voltaire, that England was a country with a hundred religions and one sauce. As, until a few years ago, an Englishman was disincUned to dine in pubUc, so in the old days he preferred not to pray in pubUc. In France, the people for private devotions betook themselves to the numerous chapels lining the naves of cathedrals. The plan had not sufficient privacy for the English man. He wanted a church he could call his church, which he shared with a select few ; a priest he could regard as his priest. And so, almost touching one another at times, churches were crowded together in towns, hemmed in by houses, and hardly visible to the casual passer-by, each possessing its own regular congregation. Occasionally a portion of an abbey church was granted to a parish, as at Chester, where the south transept of the present cathedral is far larger than the north, and with its eastern and western aisles forms a church in itself. EngUsh parish churches are of all shapes and sizes. Some Plans. retain thek original plan, others have been altered; aisles have 1FFLEY CHVRCH, OXO/1. SCALE OF FEET. Fig. 248. been pulled down in some, in others they have been added, and not unfrequently chapels as well built by pious founders to contain their tombs. To attempt to enumerate the many pecu Uarities in plan resulting from such alterations would be a tedious, lengthy task, and also an unprofitable one. The simplest form 360 A HISTORY OF ARCHITECTURAL DEVELOPMENT. Chancels. Proportion of nave to chancel. of church consists of merely a nave and chancel, with or without a tower at the west end. To this is often added a porch on the south side, and occasionally one on the north in addition. In the early centuries of mediaeval architecture the entrances in parish churches were nearly always lateral ; it was not untU the fifteenth century that a doorway in the western waU of a tower became customary. With few exceptions the chancels end square. Warwick Church, Cumberland, however, has an apse. The Normans were able to dictate an apsidal ending for their abbey churches and cathedrals, but in the parish churches the people continued the old traditional English plan. Some Eomanesque and later churches without aisles have transepts, like Old Shoreham and Alfriston, both in Sussex ; but the cruciform plan is the exception and not the rule for parish churches. Before the Norman conquest churches had been built with aisles, and the aisled plan was adopted afterwards whenever the accommodation required was greater than a nave alone could contain. Chancels were generaUy buUt without aisles, but amongst early examples with them is S. Mary, Eastbourne, and amongst late ones, Thaxted Church, Essex, Long Melford Church, Suffolk, etc. The chancel is nearly always raised above the nave, and the altar, with very few exceptions, has one or more additional steps surrounding it. In large churches there are often as many as seven or nine steps from the nave to the level on which it stands ; the number, as a rule, being an unequal one. In addition to the step or steps at the entrance to a chancel and round an altar, one step west of the altar rail separates the choir from the space reserved for the clergy at the extreme east end. Besides the altar in the chancel, there was generally an altar at the east end of each aisle, as is conclusively shown by many existing piscinae in the side walls, although the altars themselves have long ago dis appeared. Vestries were not general untU the fifteenth century. As a rule, they are placed on the north side of a chancel, but at S. Mary, Eastbourne, they are at the extreme east end. A smaU doorway is common on the south side to enable a priest to reach his seat without walking the length of the church, and occasionally there is one to the north. No definite rule can be stated regarding the proportion of nave to chancel in parish churches, but, roughly speaking, the later the church the smaller the chancel. In early thirteenth- century work one frequently finds the chancel two-fifths of the ENGLISH PARISH CHURCHES — TIMBER ROOFS. 361 total length, as at Cobham Church, Kent, and sometimes chancel and nave are equal in length, as at Broadwater Church, Sussex. In churches of the fourteenth and fifteenth centuries, a common proportion was three to five, the nave frequently consisting of five bays, and the choU, when aisled, of three. In some examples built towards the end of the latter century the east end is shorter WARAMAGTOA CHVRCM , /NORTtlAAITS . T. TOWER P.P. PORCH SCAL.E OF Fig. 249. stiU, thus paving the way for the time when, after the advent of the Eeformation, the chancel almost disappeared. The proportionate width of aisles to nave varies considerably. Propor- In early work the former are generally narrow; in the fourteenth ais^°to century they frequently approximate to the proportions customary nave. in cathedrals, viz. width of nave double that of each aisle, as in S. Patrick, Patrington, Yorkshire. This church differs from the majority of EngUsh parish churches in having transepts of three bays, each with an aisle on both sides, and a lofty tower over the crossing crowned by a spire. In the foUowing century the aisles are sometimes nearly as wide as the nave, as in the fine church of Walpole S. Peter, Norfolk, where the nave is 19 feet 2 inches wide, each aisle being 17 feet, and in S. Mary Magdalene, Gedney, also in Norfolk, the nave of which is 21 feet 6 inches wide, and each aisle 19 feet 6 inches. In both churches the chancels are without aisles. The aisles of Thaxted Church, Essex, are actually wider than the nave. Occasionally churches followed the plan of the Dominican order, and had but one central arcade, which divided the church in two longitudinaUy. Such was the plan of the old Church of S. Mary, Truro, one half of which stUl remains as an outer aisle to the choir of the present cathedral. The most striking characteristic of fifteenth-century churches is 362 A HISTORY OF ARCHITECTURAL DEVELOPMENT. their spaciousness. They were built for congregational worship, in the days when the power of the monastic orders was on the wane, and often at the sole expense of private donors who were laymen. The central idea that governed their plan was the desire to accommodate as many people as possible. With this object in view, both naves and aisles are as wide as could conveniently be spanned by timber roofs ; the piers are as far apart as possible, and are, moreover, as slender as stability would aUow. The western arms of the majority of churches of this period were Sectional ordinance. S. PATRICK, PATKI/1GTO/1 P. PI5CI/1A SCALE OF Fig. 250 (H. A. Paley). fitted with fixed seats, with ends elaborately carved, and many of these still exist, notably in the western counties. The internal division of nave, triforium, and clerestory, customary in cathedrals, is rare in English parish churches. S. Mary Eedcliffe, Bristol (mostly c. 1375-1440), has already been mentioned as the most conspicuous exception. This church is vaulted throughout, and although the triforium does not, strictly speaking, exist, there is a considerable space between the arches of the arcade and the windows above, which is treated with elaborate panelling similar in design to that of the choir of Wells, finished a few years previously. The majority of churches built between the middle of the eleventh century and the middle of ENGLISH PARISH CHURCHES — TIMBER ROOFS. 363 the fourteenth have no clerestory windows. In some cases the rafters of the nave roofs are continued down to the aisle walls, so that aU three divisions are covered by one roof; in others the aisle roofs are of considerably flatter pitch, and there is a smaU portion of walling between the eaves of the nave roof and the top of the lean-to side roofs. This occasionaUy in the twelfth and thirteenth centuries was pierced with quatrefoUs, tre foils, etc. In churches of the first half of the century foUowing even such smaU clerestories are rare. The typical aisled church of that period is Ughted en tirely from the sides, and from the end waUs. The aisle roofs are sometimes lean-to, sometimes double- pitched like the central one so that each end shows three gables. In many country churches abroad the aisle win dows rise above the eaves, and there is a series of gables along each side, as in one of S.PATRICK, PATR1/1CTO/1 VEST ELEVATIOri. Fig. 251 (H. A. Paley). 364 A HISTORY OF ARCHITECTURAL DEVELOPMENT. the churches at Quimper, Brittany, but this design is rarely met with in England. The foreign examples often have an extremely picturesque appearance, but it is a question if the simpler English plan is not preferable. Owing to the absence of clerestories early fourteenth-century churches are invariably low; their lowness being most noticeable when nave and aisles are approximately equal in width and are roofed in the same way. They were made low deliberately. The builders wanted their churches to be in harmony with the countryside, and not to be blots on the land scape. When they desired a landmark they built a tower, or tower and spire. The west end of Patrington Church is only 46 feet high from the ground to the apex of the cross of the gable, whereas the spire rises to a height of 170 feet. Such dispropor tion — if such a term can be appUed to what is really a beauty — is never found abroad. The love for length and lowness, so marked in EngUsh cathedrals, is still more characteristic of English parish churches of the period mentioned. In the latter half of the century there came a change which was continued all through the following one. Clerestory windows were reintroduced, and became almost universal. There is hardly a fifteenth-century church of any size without them. In many examples they form a continuous band from the west waU of the nave to the waU over the chancel arch, and when the chancel is aisled they are some times continued along it as well. Considering the added light they give it might have been thought that aisle windows would have lost their importance and become smaUer. The contrary was the case; they increased in size. For this the love for stained glass was chiefly responsible. The desire for it was always existent throughout the middle ages, and in the fifteenth century it approximated to a disease. Windows reached from buttress to buttress, from plinth to parapet, and walling almost disappeared. The mason became little more than a frame-maker. Notwithstanding the large proportion of glass to frame the churches were not overlighted, because the glass was coloured. Neither was there the feeling of insufficiency of wall space which strikes one so strongly now that clear glass has taken the place of the other. Void and soUd blended. The question as to which was which did not force an answer. This was especially the case in the Eastern counties, where the outside face of the walls of many of the churches was panelled. But the result was not altogether satisfactory. These churches did not look weak, but ENGLISH PARISH CHURCHES — TIMBER ROOFS. 365 they must have looked flat, as stained glass from the outside has a very opaque appearance, and the difference between the sunk S.PETERS AW1CROFT, NORWICH, 20 FEET. Fig. 252. panelling and the pierced was consequently little noticeable. Much of the charm of twelfth and thirteenth-century work in 366 A HISTORY OF ARCHITECTURAL DEVELOPMENT. both large and small churches lies in the contrast between wall and window. This disappeared to a great extent in later churches, and the absence of it is one of the reasons for their inferiority. Open The open timber roof, treated architecturally, is almost exclu- tunber sively an English feature. The basilican churches of Italy have mostly coffered ceilings, and in the few in which the timbers show, as in S. Miniato, Florence, the design is of the simplest character and the only decoration is paint. Even at Verona, where S. Zeno and S. Fermo Maggiore have more elaborate roofs, the treatment is stUl a panelled one. In France, a church of any size is generally vaulted, and, as mentioned before, most of the parish churches are large. Many small churches there are unvaulted, but the carpentry of their roofs shows Uttle taste or ingenuity in design, and not unfrequently consists of timbers arranged to form a pointed barrel roof which is covered on the inside by plain boarding, with, or without, smaU ribs reaching from waU plate to ridge. The supremacy of the mason in France, and the square mUes of oak forests which England at one time possessed, account to a great extent for the difference between the two countries. No roofs, save a few very plain ones, exist in this country of the twelfth or thUteenth century. Most were probably of ordinary tie-beam construction, with king or queen posts, and were sometimes ceiled, as in S. Albans Cathedral, so that none of the timbers showed. Their pitch was very steep, about 60° or more, and this can often be determined in many churches, the existing roofs of which are later and flatter, by the original stone weather tabling on the east side of a western tower. Some fourteenth-century roofs remain — over a portion of Crick Abbey is a fine one with the ball flower ornament carved on the timbers — and there would be more if it were not that many were removed in the following century, when the side waUs were raised to form the clerestory which the fifteenth-century builders loved. These roofs were seldom replaced, as their steepness, although not quite so great as in roofs of the previous centuries, would have made the churches too high, and their thrusts would probably have been too great for the raised walls. Many fifteenth-century roofs over fifteenth- century churches are of fair pitch ; but roofs of that date over earlier churches to which clerestories have been added are invariably flat. In English roofs, with the exception of ENGLISH PARISH CHURCHES — TIMBER ROOFS. 367 the early ceUed ones aUeady mentioned, and a few later roofs which are paneUed, aU the timbers show from wall plate to ridge. These are of large scantlings, beautifully framed together and pinned — iron straps were unknown — and, except the common rafters, elaborately moulded. Sometimes the rafters are moulded as well, but they are, perhaps, most effective when they are not, as then their sturdiness is most apparent. Six inches by 4 inches is no unfrequent size for them in mediaeval churches, and it must be remembered that in most instances they are laid flat, so that the width of the underside of each is the greater dimension. How different is the 4 inch by 2 inch rafter placed the other way of the buUders' roof of to day ! All timbers were oak. Some churches are stated to have had roofs of chestnut, but there are very few in England in which this wood was used, and possibly none. Mediaeval open timber roofs are divided into two main classes. Types of (1) Single-framed roofs, in which there are no principals. (2) Double-framed roofs, in which the rafters are supported at intervals by principals framed in different ways. The second class can be subdivided into tie-beam roofs, hammer- beam roofs, coUar-beam roofs, and roofs in which the principals consist merely of rafters stiffened by curved braces. In single-framed roofs each pair of rafters is trussed together. Single- The foot of each rafter rests on one end of a plate laid across the i^^ei waU, and the rafter is further supported by a vertical strut tenoned into it, and into the other end of the plate. A triangular truss is thus formed of considerable strength. The transverse plates are notched on to a longitudinal plate, generally in the centre of the wall, which runs the whole length of the church, or the length of the nave or chancel, as the case may be. Each pair of rafters is also united by a collar about one-half or two- thirds above the start of the roof, which is strutted from the rafters. All the different timbers are halved and pinned or tenoned together. In some examples the struts are curved like braces, and a waggon or barrel roof is formed, which is either pointed or semicircular. A moulded cornice (the only part of the roof which is not structural) usually conceals the ends of the transverse plates, and this in many examples is also carved. In several roofs of the above type, both polygonal and waggon shaped, all the timbers are concealed by boarding, which is generaUy divided into panels by ribs. Considerable care was 368 A HISTORY OF ARCHITECTURAL DEVELOPMENT. exercised in the spacing of the timbers so that all the panels should be approximately square and equal in size. At Wim- botsham Church, Norfolk, is a good example of a many-sided HECKI/iGTOAl CHVRCH, unes. SOVTH PORCH. WIMBOTSHAtt CHVRCH, /I AVE. NORFOLK. FROn BRA/1DO/1. Fig. 253. Tie-beam roofs. Hammer-beam roof. panelled roof, and in the West of England, in Cornwall especially, are several panelled waggon roofs, mainly of the fifteenth century. Eoofs with tie-beams were used in all the centuries of mediaeval art, but they were most common in early work and in late. In the former their high pitch aUowed of either king or queen posts, which are strengthened by curved braces. The tie- beams are also supported by braces tenoned into vertical wall pieces which rest on corbels. In the numerous flat-pitched roofs with tie-beams of the fifteenth century there is little room for posts, but the spaces between the beams and principal rafters are filled with pierced or carved panelling. The tie-beams of these are seldom straight, but rise slightly towards the centre, and sometimes, in conjunction with the curved braces underneath, form a series of flat four-centred arches. In all open roofs the ridges are heavy timbers, often elaborately moulded, and frequently supported by curved braces which run longitudinally from prin cipal to principal. In many cases the purUns are similarly strengthened. The most ingenious type of roof is the hammer-beam, or canti lever, and it is also generally the richest. It was the favourite ENGLISH PARISH CHURCHES — TIMBER ROOFS. 369 form in England in the fifteenth century for both wide and narrow spans. The construction of each principal is much the same as at the foot of each rafter in a single-framed roof, except that the hammer- beam, which takes the place of the trans verse plate, does not rest on the wall but projects in front of it, and is supported by a wall piece and brace. The whole forms an exceedingly strong truss. In some roofs there are two or more hammer- beams to each principal, one above the other; whUst in others, instead of the upper hammer-beams, there are collars strengthened by braces. In the roof of Westmmster HaU there is an unusually strong pair of braces to each principal, which start from corbels and unite all the timbers together. The suitabiUty of the hammer-beam roof for wide spans is well illustrated in this building, which is 65 feet wide, and in the halls of Fig. 254. — Construction of Hammer-beam. BACTO/1 CHVRCH, ^§faffldfi^fc-#T—ffejr-5 SVFPOLK. /1AVE.^| ISStil II e^JSfi 311 U^j^E^y^^J — ---lilllHHBa Hire" 1 nllH 1 iis^^^iilili 11 ¦ lllf 111 Hi 12 0 1 2 3 4 5 S 7 ? 9 IO B S FROtt BRA/1DO/1. SCALE OF FEET. Fig. 255. many of the Inns of Court, of the Charterhouse, Eltham Palace, and in several churches of over 30 feet span. A hammer-beam VOL. II. 2 B 370 A HISTORY OF ARCHITECTURAL DEVELOPMENT. roof is preferable to a tie-beam one when the width is greater than a single beam can bridge, as scarfing is avoided, and owing to the science displayed in arranging the timbers so as to transmit all thrusts on to the waU, the absence of a horizontal tie is hardly felt. Fig. 256. — Westminster Hall : Cross section. Purlins,etc. It is unnecessary to say anything about roofs with principal rafters strengthened merely by curved braces, or by collars and braces. Whatever the type of roof may be, the purlins are invariably tenoned into the principals, and do not rest upon them as in ordinary modern roofs. In roofs with timbers of light scantling this could not be done without endangering their stability, but there was no risk of this with the heavy timbers ENGLISH PARISH CHURCHES — TIMBER ROOFS. 371 used in the Middle Ages. The method has great aesthetic advan tages over the modern one, as it brings all the parts closely together, and mouldings can mitre with one another with excellent results. In many roofs, the main principals of which are con structed in any of the ways aUeady mentioned, there are inter mediate principals of Ughter scantling and fewer parts. The main principals come over piers or solid walling; the others generally over windows or arches. At each end, east and west, is placed one of these half or intermediate principals, to give a finish to the roof and frame in the purUns and common rafters. BRADFORD CHVRCH, SVFFOLK, SOVTH AISLE . FROM BR A/1 DO/1. Pig. 257. WhUst a few aisled roofs are double-pitched, the large majority Aisle are lean-to. In the hands of the mediaeval carpenter this com- ro° s' monest form of roof acquired considerable distinction. There was no room or need for elaborate tie- or hammer-beams ; but nearly aU aisle roofs have braced principals, of heavier scantling than the ordinary rafters, and moulded purUns. There are very few open English timber roofs without some Decora- carving on them, and all were originaUy richly painted and gilded. *£*£* The cornices are crowned by battlementing or cresting, and 372 A HISTORY OF ARCHITECTURAL DEVELOPMENT. frequently are pierced or carved ; the spaces between the timbers of principals are filled with simUar pierced panels or carvings, and at the ends of the beams in hammer-beam roofs are frequently carved angels with outspread wings, whilst in some cases the actual beams are carved to represent angels with their wings folded. Much of the painted decoration has disappeared, but enough remains in many churches to show how rich and effective it was. White always entered largely into the colour scheme, and " barber poling " in white and green, or red and white, was a favourite treatment for three-quarter round mouldings. The boarding between the common rafters was frequently painted white, with sprigs of flowers at intervals, or blue, in which case it was studded with lead stars screwed on and gilded. Similar decoration was applied to the rafters themselves, and to the timbers of the principals, care being taken to counterchange as much as possible, so that although many colours were introduced a broad effect resulted. The richest decoration is generally over the chancel, or in the bay of the chancel immediately over the altar. Either is frequently boarded when the rest of the church is not, and gold is appUed in greater profusion there than else where. In many a small parish church, otherwise of little interest, the roof imparts a distinction to it which parish churches abroad of similar size rarely possess. If the work of the masons deteriorated in England in the fifteenth century, that of the carpenters undoubtedly advanced. This is evident not only in roofs. The work in the wood screens and stalls of many fifteenth- century churches is beyond all praise, and tf Uttle mention has been made of it in this volume, it is not from any want of appreciation, but simply because to have treated these fittings at all adequately would have meant a greater space than could be spared. APPENDIX. TABLE OF DIMENSIONS OF TYPICAL CHURCHES. The foUowing table gives the approximate internal dimensions of some of the principal churches in Europe. The examples chosen are grouped under heads in order to illustrate the differences in proportion in Eomanesque and Gothic ; in churches barrel-vaulted, rib-vaulted, and domed ; in those without either triforia or cleres tories, those with triforia only, and others in which the three divisions, arcade, triforium, and clerestory, are found. The dimen sions are given of some large aisleless churches, and of churches which, although aisleless, have deep side recesses. The table also shows the striking differences in scale and proportion between examples in different countries, and emphasizes the characteristics of each in these respects. WhUst no claim is made that the dimensions given are exact, considerable trouble has been taken to make them as correct as possible. The compilation of such a table is no easy matter. Plans when scaled show striking discrepancies. Dimensions given by authorities vary considerably, and it is seldom that two are found to agree. When stated in figures, as a rule there is nothing to show where they are taken from exactly. From wall to wall seems the natural measurement, but in many cases it is evident that recesses in waUs have been included, or that the tape has been held to doors, or to the glass of windows, in order to gain a few extra feet of length or width. Inaccuracies have doubtless crept in in the following table, due in some cases to the size of the plans and sections available, in others to the calculations necessary to change French feet, Prussian feet, metres, etc., into English feet ; but the intention has always been to state the dimensions in the clear between walls, and to ignore casual recesses. TABLE OF APPROXIMATE INTERNAL DIMENSIONS IN FEET. (a) Exclusive of chapels. (e) Including western narthex. (i) Height of inner aisles. (o) Inclusive of chapels. (/) Double bays covered by one vault. (j) Height of outer aisles. (c) Aisleless. (d) Height of aisles of crossing. (g) Including porch. (h) Including Henry VII.'s Chapel. (k) Including western tower. FigH. in letterpress. Name of example. Width of nave be tween walls. Total width of western arm. Width of bays centre to centre of columns. Length of nave exclu sive of crossing. Width across transepts. Width of eastern arm. Total length. Height. Aisleless Bomanesque churches, S. France barrel-vaulted). k^ Orange Cathedral 45 Abbey Church of Montmajour . . 43 0545 3421 140 97 t*i — 45 — 65 54 Aisleless Bomanesque churches, S. France (domed). s 144-5. 146. Angouleme Cathedral . . Fontevrault Abbey Church . . | 37 ... 52 5252 44 43 148 186 158134 40(a), 58(h) 62 242 285 66 66 Aisleless Bomanesque churches, S. France (rib-vaulted). 47, 162. 56 46 50 46 50 40 155 156 155 56 57(a), 71(6) 300244 8078 S. Radegonde, Poitiers . . Aisled barrel-vaulted churches, S. France, no triforium or clerestory. 151. S. Nazaire, Carcassonne . Fontfroide Abbey Church . 2627 5258 1622 102 107 90 27(c) 177 55 68 Aisled barrel-vaulted churches, S. France, with triforium, no clerestory. 52-3-4-5-6. 157. 83, 160-1. 150. 141-2. 148, 165. 121. 125-6-7-8. 111-2-3. 133-4,137. 135-6. 75, 173. 174. 99 105,178,' 218-9. 176-7, 179, 182. 101, 213,\ 217. Issoire Cathedral . S. Sernin, Toulouse 2530 521)4 1517 121 235 Other aisled barrel-vaulted churches, S. France. Autun Cathedral . . S. Philibert, Tournus S. Front, Perigueux . Le Puy Cathedral . 3123 71 62 1818 125 93 Aisled domed churches, S. France. 40 27 72 67 25 144 92 192 106 90 188126 50(a), 68(6) 60(a), 85(6) 71 52 7032(c) Aisled Bomanesque churches, either timber-roofed or vaulted, Italy, Oermany, Normandy Pisa Cathedral 43 Monreale Cathedral . .... 47 S. Ambrogio, Milan ... . 44 Worms Cathedral ... 39 Speier Cathedral 46 Mainz Cathedral . Jumieges Abbey Churoh S. Etienne, Caen . . 45 3433 106 8787 88 113 / 105 (a) \ 133 (5) 68 75 17 15 44(/)36(/) 41(/) 33(/) 18 20 170 142 175 242 235 250 183 246113 111154 152 131 106113 4439(c)51(c) 82(c) Bomanesque and Gothic churches of England. 204362 264(e)250(e) 220206 308 92 300(o) 80 220 63 356 90 420 108 380 57..71(d) 70 7861 Norwich Cathedral . Winchester Cathedral Gloucester Cathedral Durham Cathedral . Ely Cathedral 3135 343234 7186 8780 79 9266 88 84 72 18 253 174 71 400 73 23 270 215 88 526 78 20 187 148 84 407 68 25 198 169 76 1 415, \ I 472(a)/ 72 17 238 184 79 514(a) 70 ^^£I S TABLE OF APPROXIMATE INTERNAL DIMENSIONS IN FEET— continued. Figs. in letterpress. Name of example. T 100, 102,; 209,210, 212. ] 100, 214. 214. 215.216. 94. 50, 61, 64, 1 103, 196./ 106, 189, \ 197. / 194. 62, 187. 193. 188, 195,\ 208, 210./ 97-8. Width of nave be tween walls. Total width of western aim. Width of bays centre to centre of columns. Length of nave exclu sive of crossing. Width across transepts. Bomanesque and Gothic churches of England — continued. Salisbury Cathedral . Lincoln Cathedral Westminster Abbey Exeter Cathedral . York Cathedral . . Canterbury Cathedral Laon Cathedral . . Notre Dame, Paris . . . Chartres Cathedral Reims Cathedral . . . . . Bourges Cathedral . . . . Amiens Cathedral Le Mans Cathedral (choir only) . Beauvais Cathedral 35 383436 4730 78 S280 74 105 73 20 27 19 20 27 20 198205240209 182 Gothic churches of Northern France. 38 4652 4546 46 / ?2(a)l 1 100 (6)f / 120(a)! \ 149(6)/ 110100 / 135(a)! \ 165(6)/ / 103(a)! I 152(6)/ 30(/) 36(/)23 24 45(/) 25 180 186 194 250 185 204 224191142225130 180 15S 204 162 199180184 Width of eastern arm. Total length. Height. 78 84 77(a) 115(5) 74 100 88 100 450 481 408 origin ally, 505(A) 384486 518 84 '74 choir, 82 nave | 101 70 99 80 / 120(a)! 1 149(6)/152150 / 135(a)! I 165(6)/ 120(a)!149 (6)J 152150135(a)!165 (6)J 156 / 125 (a)l \ 178 (6)j 133 365 409425455 381449 81 110 114 125 fl25,69(0, I 28 (J) 139}110 154 Oothic churches of Southern France. 200. 201-2-3. 204. 226.231. 227-8. 230. 234, 236,\ 238. J 235, 237. 241-2-3. 245. 246-7. Poitiers Cathedral .... Albi Cathedral Ohuroh of Jacobins, Toulouse Cologne Cathedral Ulm Cathedral . S. Mary, Lubeck . S. Peter, Lubeck . 40 1 33 ohoir / 60 63 S. Anastasia, Verona . . . S. Maria Novella, Florence . SS. Giovanni e Paolo, Venioe Como Cathedral Florence Cathedral .... S. Petronio, Bologna . . . Milan Cathedral Gerona Cathedral .... Barcelona Cathedral . . . Salamanca Cathedral (new) . Segovia Cathedral .... Seville Cathedral 107 9983 40 2330 Oothic churches of Germany. 46504133 152160 112 78 2624 21 27 Gothic churches of Italy. 354342555759 58 75 90 92 125126 132 (a)' 188(6) 190 / 132(a)! \ 188(6)/ 334640406264 33 Gothic churches of Spain. 73424344 46 102 82(a)! 118(6)/110(a)! 162 (5)/ 114(a)'158 (6) 194(a)247(6) f 114(a)! I 158(6)/ / 194(a)! I 247(6)/ 4132 33 353S 150 213 310 195 250 204 180 250 386 288 165148 166169187 185 254 110135143 255 124162 250 100 9983 152 51(c) 110 78 35(c)43(c)43(c) 128 85(a)! 106 (6)/ 110(a)! 162(6)/153194(a)! 247(6)/ 310350 (i) 277 452 416312 188 284 340 322 488273284340 342427 75 ohoir 85 nave 9993 150138125 67 85868595 145138 157 117 89 118 125 110 5 INDEX. Abacus — Octagonal, 39 Plan of, changes in, 37-38 Upper (pseudo-dosseret), examples of, 35 Abbeville — S. Ricquier, 105, 278 note ' S. Vulfran Cathedral, 33, 277, 284- 285 ; decoration in, 121, 284 Abbotsbury (Dorset) — S. Catherine's Chapel, 71, 91, 318 Abutments, see Buttresses Addison quoted, 3 note Adel Church (Yorks.), 161 Aix-la-Chapelle Church, 138-139, 333 Alan of Walsingham, 315 Albi Cathedral — Exterior of, 289 Fortified tower of, 106 Gerona Cathedral compared with, 353 Jacobins' Church (Toulouse) com pared with, 292 Plan of, 157, 289-290 Span of, 286 Alencon Cathedral, flamboyant style in, 284 Alfriston church (Sussex), 360 Altars in Churches — Choir widely separated from, 354- 355 Increase in number of, 144-145 Alternation of materials or colours — Brick and marble, 20, 43, 109, 117, 175, 345 Brick and stone, 117 Bricks, glazed coloured, 109 note Marbles, coloured, 117, 175, 341 Stone and tile, 136 Amain, arcading at, 52 Amiens Cathedral — Bourges Cathedral compared with, 274 Chapels of, 153 Amiens Cathedral — continued. Chevet of, 149 Cologne Cathedral modelled on, 325-327 Commemoration of master masons connected with, 256 note Construction of, 275 Dates of, 262, 302 Destruction of, 262 Double aisles of, 156 Exterior of, 276 Flamboyant style in, 284 Fleche of, 276, 303-304 Plying buttresses of, 98 Horizontal lines in, 279 Nave of, 152 Piers in, 32, 355 Plain wall space in, 272 Plan of, 266-267 Proportions of, 265, 272, 302-303 Salisbury Cathedral compared with, 301-304 Sculptured figures in, 124 Site of, 264 Triforium of, 271 Vaulting in, 102 West front of, 278, 300 Andernach Church, 190, 197 note ', 198 Angers — Cathedral (S. Maurice) — Date of, 259 Paintings in, 123 Plan of, 205, 226-227 Span of, 286 Style of, 259 Vaulting of, 78, 86-87 La Trinite— Buttresses of, 269 Plan of, 158, 205, 226-228, 353; illustrated, 227 Restoration of, 117 note, 228 note ' Vaulting of, 79, 81 S. Serge, 230 38o INDEX. Angilbertus, Arohbishop, 168 Angouleme Cathedral- Plan of, 207-209 Proportions of, 210 Towers of, 104 West front of, 230 Animal forms supporting columns, etc., 173 Antwerp Cathedral, 156, 335 Apses in French oathedrals, piers of, 268 Apulia, churches of, 180-184 Aquileja Cathedral, 172 note Aquitaine churches, foreign influence in, 203 Arcading of walls, 51-52, 56 Arches (see also Vaults) — Construction of, Roman and Me diaeval, 14-16- Cusped (foliated), 13, 14 Elasticity of, 14 Enrichments carved on, 127 Extrados of, not concentric with intrados, 183 and note 2 Foliated (cusped), 13, 14 Horseshoe, 13, 183, 219 Intersecting, 10-11, 52 Mouldings of, see Mouldings Ogee, 12 Painting and gilding on, 120-121 Parapets, of, 53 Pointed — Advantages of, 10 Early (pre-Hellenic) use of, 9 Forms of, 11-13 Introduction of, dates of, 10 Vaulting, in, 77-8 and note Romanesque form of, 9 Semicircular — Examples of, 10-12 Popularity of, 13 note Sicily, in, 185-186 Springers of, 15-16 Squinch, carrying spires, 113 Stilted- Prance, in, 13 Sicily, in, 186 Windows, of, 60 Subordination of, 15 Thrust of, see Thrust Transverse, in vaulting, 71-72 and note >, 79 Arohes — continued. Trefoil, 12-13 ; in Auvergne, 219 Unmoulded, examples of, 17 Windows, of, see under Windows Aries— S. Trophime, 216, 230 Ashlar — Facings of, 23 Solid piers of, 29, 45 Walling of, 29, 45 Assisi, Franciscan church at, 339 Asti, Baptistery at, 161 Audenarde — N6tre Dame de Pameleh, 335 Austin Canons, 243 Autun Cathedral — Arches in, 251 note ' Cluny the model for, 223 Doorway of, 123, 222-223 Fluted pilasters in, 32 Plan of, 223-225 Porch of, 122, 222-223 Sculpture in, 123, 124 Vaulting in, 71 Auvergne churches — Arches in — ¦ Foliated, 14, 219 Semicircular, 10 Stilted, 13 Transverse, 71 Lighting of, 218 Piers in, 32 Roofs of, 218 Similarity between, 216 Triforia but no clerestories in, 8, 197, 216 Upper abaous in, 35 Auxerre — Cathedral — Dates of, 264 Vaulting in, 102 View of, 265 Wall thickness in, 34 S. Eusebe, carving in, 127 S. Germain, tower of, 111 S. Pierre — Date of, 285 Plain wall space in, 272 Avallon, 230 Avignon Cathedral, 216 Bacharach— S. Werner, 324, 333 Bacton Churoh (Suffolk), 369 INDEX. 38i Bale Cathedral, 197 Bamberg Church, 324 Bangor Cathedral, 244 Baptisteries, 161-162 Barcelona- Cathedral — Chapels of, 353-354 Interior of, 356 Passage-way triforium of, 357 Plan of, 353-354 S. Maria del Mar, 353 S. Maria del Pino, 352 Barfreston Church (Kent), 60, 161 Bari — Cathedral — East end of, 161 East window of, 173 Frieze of cupola of, 181 Pierced marble window slabs in, 181 S. Gregorio, 181 S. Nicolo— East end of, 161 Exterior of, 182 and note, 183 Glass mosaic in, 181 Plan of, 181-182 Tower of, 103 Barisanus of Trani, 180 Barnack Church, 235, 239 Bases, development of, 40 Basilican churches — Middle storey of, 5 Towers of, position of, 106 Bath Abbey, 65, 92, 317, 318 Batter — Buttresses, of, 50-51, 179 Towers, of, 51, 108, 110, 179 Battlementing, 53 Bayeux Cathedral — Chevet in, 268 Columns in, 32 Date of, 301 Destruction of original, 242 Detached shafts in, 32, 33 Beaune Cathedral, 222, 225 Beauvais — Basse OSuvre, 136-137 Cathedral — Chevet of, 149 Choir of, 274 Collapse of, 263, 275 ; cause of, 304 note Beauvais — continued. Cathedral — continued. Flying buttresses of, 100 Glazed triforium of, 271 Gothic mixed with Renaissance in, 284-285 Proportions of, 272 Transept doorways of, 277 Vaulting in, 102, 263 note Windows of, 69 S. Etienne, 74-75 and note, 259 Belgium (for towns, churches, etc., see their names) — Brick buildings in, 43 Churches of, 334-335 ; niches in, 33 Secular buildings in, 335 Benedictine Order — Building activity of, 1 Cistercian rivalry with, 158 Decline in supremacy of, in France, 254-255 Influence of, 203 Mission of, to England (596), 233 Towers to churches of, 115 Benevento Church, 183 Bergamo, Baptistery at, 161 and note 3 Bernard of Clairvaux, S., 158-159, 254 Berneuil, plan of, 208 Bernieres-sur-mer Church, 80, 243 Beverley — Minster — ¦ East transepts of, 278 Plan of, 153 Windows of, 62 S. Mary, 129 Bilson, John, cited, 76 and note, 251 Bodley, G. F., quoted, 282 Boffiy, Guillermo, 353 note Bologna, S. Petronio — Capital in, 36 Date of, 338 Plan of, original, 344 and note Windows of, 342 ; tracery, 346 Bonn Cathedral, 193, 199 Bordeaux Cathedral, 32 Boscherville, S. Georges de— Date of, 242-243 Eastern arm of, 146, 247 Proportions of, 248 Vaulting of, 79 Bosses, 90-91 Boston Ohuroh (Linos.), 62, 115 382 INDEX. Bourges Cathedral — Apse of, 268 Chevet of, 149 Date of, 262-263 Double aisles of, 156 Exterior of, 274-275 Flying buttresses of, 269 Interior of, unique, 274 Proportions of, 265, 272 Section of, 274 Shafts in, 33 Vaulting in, 81, 102 View of, 265 West front of, 275, 280 Windows of, 133 mentioned, 328, 346 Bowtell moulding, see under Mould ings Boxgrove Priory (Sussex), 91, 263, 314 note ' Bradford-on-Avon Church (S. Law rence) — Dimensions of, 238 East end of, 161, 234 Masonry of, 236 note Plan of, 234-235 Braisne Church (S. Yved) — Chapels of, 325 Lantern tower of, 104, 112 Plan of, 326 Bramford Church (Suffolk), 371 Breamore Church, 238 Brede Church, 64 Brescia Cathedral, 139 note l Brick- Belgian churches, employment in, 335 French churches of, in the South, 289 German churches, employment in, 328 Glazed, towers of, in Italy, 109 note Marble alternating with, 20, 43, 109, 117, 175, 345 Stone alternating with, 117 Brinkburn Priory (SS. Peter and Paul), 18,57 Brioude Cathedral, 216-218 Bristol— S. Mary Redclifie, 91, 358-359, 362 Brittany churches, English influenoe in (14th oentury), 282-283 Brixworth Church, 238-239 Broaches, 113-114 Broadwater Church (Sussex), 361 Bronze doors, 180 Buildwas Abbey (S. Mary) — Date of, 10, 297 Design of, 297-298 Plan of, 159-160 Windows in, 56 note Burgh-by-Sands Church, 106 Burgos — Cathedral, 351 Las Huelgas Church, 159 Burgundy churches — Barrel-vaulted, 223-225 Carving in, 126-127 Clerestory and triforium in, relative importance of, 225 Monastic influence in -construction of, 73, 221 Mouldings in, 19 Narthices in, 222-223, 230-231 Rib- vaulted, 225-230 Transverse arches in, 71 Upper abacus in, 35 Vezelay, see that title West fronts of, 230 Burial inside churches, 162-163 Burton-on-Humber Church, 235 Bury St. Edmunds Church — Bosses in, 130 Date of, 244 Plan of, 147 note 2, 150 Buttresses — Advantages of outside, 46 Batter of, 50-51, 179 Corner, 49-50 Flying- France, in, 51 / Hidden, 96 and note V / Piercing and panelling of, 282 v Position of, in relaifion to vault thrusts, 97-98 v / Sexpartite vaults, to, 269 v Spires, on, 114 V / Stifieners, as, 98-99 v / Thrusts transmitted by, 97 '^ France, Southern, in, 46^47 note Functions of, 46/ 269 -¦ Italy, in, 183 V Niches on, 50 Octagonal, 50-51 Panelled, 50 INDEX. 383 Buttresses — continued. Pilaster, 51 Romanesque, form of, 46 / Sculptured figures on, 50, 124 v Shafts at angles of, 49 Stages of, 47-50 / Towers, of, 110 V Weatherings of, 49 Byland Abbey, 13 Byzantine carving, 126-127 Byzantine churches, galleries of, 6 Byzantine colony of Apulia, 166 Byzantine influence on plans, etc. — France and Germany, in, 137-142, 203 Italy, in, 175, 180-181 Sicily, in, 184 Byzantine-Romanesque work, 167 Caen — Abbaye-aux-Dames (La Trinite') — Aisles of, 146 Apses of, 248 Date of, 242 Gallery lacking in, 7 Nave of, length of, 247 Proportions of, 249 Triforium of, 249 Vaulting in, 71, 80 West front of, 230 Abbaye-aux-Hommes (S. Etienne) — Apses of, 248 Clustered piers in, 30 Date of, 242 Eastern arm of, 247 Galleries of, 7 Nave of, length of, 247 Open space before, 264 Proportions of, 249 Transept altars of, 146 Triforium of, 249 Vaulting in, 101 West front of, 230 Abbaye d'Ardennes, mouldings in, 20, 301 S. Nicolas — Roofing of, 243, 275 Tower of, 111 S. Pierre, spire of, 333 Cahors Cathedral, 207 note, 210 Cambridge — King's College Chapel — Dates of, 317 Cambridge— continued. King's College Chapel — continued. Dimensions and detail of, 318-319 Elevation of, 319, 320 Pan vaulting in, 90, 92 Glass in, 133 Plan of, 319 Templar church at, 142 Campanile towers, 108, 110 Canosa Cathedral, 181, 183 Canterbury — Cathedral— Apse of, original, 247 Chevet of, columns of, 33 note Choir of, 91, 296, 297 Dates of, 243, 315 Plan of, 150, 153 ; original plan, 146-147 and note * Towers of, 104, 107, 110, 112 Vaulting in, 81, 91, 101, 102 S. Augustine's Church (on site of present cathedral), 237-238 S. Martin's Church, 234 Capitals — Carving of, 126-128 Circular, 38 Concave, forms of, 37 Convex, 36-37 Functions of, 34-36 Nature of, 36 Octagonal, 39 Romanesque, derivation of, 35-36 Subordination of, 37 Carcassonne — S. Nazaire, 287 ; vaulting in, 10, 95, 215-216 Carileph, Bishop, 251 and note 2 Carlisle Cathedral — Date of, 244 East window of, 62 Vaulting in, 101 Carvings (see also Mouldings and Sculpture) — Advance in, during 12th century, 258 Ball flower, 129-131 and note Byzantine workmanship in, 126-127 Dogtooth, 129 Italy, South, in, 180 Paris Cathedral, on west front of, 279 Runic crosses, 241 Tudor rose, 131 Caserta Vecchia (Benevento), 14, 183 384 INDEX. Cattaneo quoted, 75 ; cited, 106, 147, 168, 191 note ', 235 Caudebec Church, 284 Cefalu Cathedral- Date of, 186 Exterior of, 230 Mosaics in, 118 Plan of, 188 Stucco on, 45 note l Cerisy-la-foret Church, 146, 243 Chalons-sur-Marne — Notre Dame de l'Epine, 284 Chamalieres Church, 216-218 Chancels, raised, 163 Chantry chapels, 155 and note, 163, 311 Chapter houses, English, 315 Oharlieu Church, 230 Charroux Church, 140-142, 147 Chartres — • Cathedral — Apse of, 267-268 Arcading on, stilting of, 268 Carving in, 131 Choir of, 268 Crypt of, 147 Date of, 262 Double aisles of, 156, 266 Nave of, 152 Open space before, 264 Piers in, 32 Plain wall space in, 272 Plan of, 276 Porches of, transepts of, 122, 277, 280 Proportions of, 265, 272 Soulptured figures in, 124, 125 Spires of, 114 Towers of, 111, 276-277 Triforium of, 271 Vaulting in, 34, 102 Wall thickness in, 34 West front of, 278 Windows of — Glass, 132 Tracery, 58, 59, 69 mentioned, 153 S. Pierre, 271 Chatres, 208 Chester Cathedral — Date of, 244 and note East end of, 247 Font in, position of, 162 Chester Cathedral — continued. Grooved shafts in, 24 and note Plan of, 147 note2, 151 South transept of, 359 Vaulting of, 86, 91, 101, 102 Chevets, 148-150, 267 Chiaravalle Church (near Milan), 159 Chichester Cathedral — Alterations in, 246 Date of, 244 Decoration in, 118 Eastern arm of, 247 Foreign influence in, 297 Nave of, 246 Plan of, 146, 151, 154 Shafts in, 28, 298 note Side chapels of, 154 Towers of, 107 Triforium of, 118, 249 Vaulting of, 84 and note *, 90, 91, 101, 102 Christ Church (Hants) — Chantry chapels in, 155 Date of, 244 Sculptured figures in, 124 Triforium decoration in, 249 Cistercian Order, influence of, on plans, etc., 145 and note, 158-161, 295, 306 Claypole Church, 63, 64 Clerestory, functions of, 5 Clermont-Ferrand — Notre Dame du Port, 148, 216 Cluny Abbey — Narthex of, 222 Nave of, length of, 247 Plan of, 143, 148, 150, 153, 223 mentioned, 220 Cobham Church (Kent), 361 Coblentz— S. Castor, 198 Coire Cathedral (Switzerland), 14 Coleshill Church, tower windows of, 111 Cologne — Cathedral Date of, 325 Nave of, 152 Plan of, 325-328 Western towers of, 333 Church of the Apostles — Apses of, 199-200 Vaulting in, 195 S. Gereon, 121 INDEX. 385 Cologne — continued. S. Maria in Capitolio — Piers and oolumns in, 31 Plan of, 199-200 Vaulting in, 195 S. Martin- Apses of, 199-200 Decoration in, 121 Tower of, 200 Columns — Diminution of, 137, 179 Doubling of, 32, 33 note Entasis of, 23, 137, 179 Flutings of, 23 Monolithic, 137 "Norman," 23 Piers — Alternation with, 26, 137, 175, 242, 251 Supersession by, 24 Rubble, of, 23 Como — Cathedral, 341 S. Abbondio, 105 Compostella — S. Iago, 219, 280 and note 2, 350 Conques Abbey, 10, 217, 219-220 Constantinople — S. Sophia — Dosseret absent from, 35 Transverse arches in, 72 note ' Corbels of parapets, 53 Corroyer,M., cited, 74 note 2, 77 note, 269 Cottingham Church, 62, 63 Coutances — Cathedral — Columns in, 32 Date of, 301 Detached shafts in, 32, 33 Double aisles of, 266 Lantern tower of, 104, 111 Mouldings in, 20 Ordinance of, 274 Side chapels of, 153-154, 263 Spire of, 114 Towers of, 111, 276, 277 S. Nicolas, 285 S. Pierre, 285 Cresy and Taylor cited, 178 note z Crick Abbey", 366 Cromer Church, 30 Croydon — Whitgift Almshouses Chapel, 319 VOL. II. Crypts, 163-164 Cupolas, 171 and note Curves in Gothio ohurches, 178-180 and note Daly, Cesar, cited, 139 note * Daphni Church (near Athens), 183 note l Dareth Church (Kent), 161 d' Argent, l'Abbe Marc, 282 de Cormant, Thomas and Renaut, 257, 279, 304 de Honnecourt, Villard, 256 ; sketch book of, 158-159 and note de Luzarches, Robert, 256 note, 257, 279, 304 de Noyer, 299 Decoration — Billet ornament, 185 and note Carving, see that title Colour schemes, 121-122 Contrasting materials, by, 117 (see also Alternation of materials) Counterchanging, 187-188 Diaper work,'218 and note 2 Dog-tooth ornament, 129, 269 Egg-and-dart ornament, 37, 127 Enrichments carved on arches, etc., 127 Fresco painting, 119 Inlays, 186 Mosaics, 24, 118 and note, 175, 180- 181 Mouldings, of, 22 Oil paintings, 120-121 Paint — English roofs, on, 372 General use of, in early times probable, 116-118 Italy, in, 342, 348 Sculpture — Archaic, 125 Examples of, 123 Framing of, 123-124 France, in, 124-125 Tempera paintings, 119 Timber roofs, of, 371-372 Deerhurst — Church, 238 Odda's Chapel, 45, 236 note, 238 Dieppe Church, 284 Dieulafoy, M., cited, 215 2c 386 INDEX. Dijon — Notre Dame — Porch of, 222 Sculpture in, 123 Style of, 264 West front of, 279 S. Benigne, 140-142, 147 S. Michel, 225 Dimensions of principal churches in Europe, table of, 373-377 Dogtooth ornament, 129, 269 Dol Cathedral, 263-264 Domed churches, 183, 189 Domes, mosaics on, 118 note Doorways of churches — Development of, 40 France, in, 279-280 Dosserets, 35 Dover Castle Church, 161, 238-239 Driibeck, columns and piers in, 137 Dryburgh Abbey, 321 Durham Castle crypt, 91 Durham Cathedral — Apse of, original, 147 note " Arches in, 17 Chapel of the Nine Altars — Apse superseded by, 245 Circular windows in, 69 Date of, 161, 252 Marble employed in, 26-27 Plan and dimensions of, 306 Clerestory windows of, 252 Columns in — Alternation of piers with, 26, 251 Variety of, 23 Dates of, 76, 244, 250, 251 Eastern arm of, length of, 150, 247 Galilee porch of, 252 Nave of — Length of, 247 Vaulting of, 82, 91, 250 Plan of, 250 Proportions of, 248-249 Towers of, 107, 252 Transepts of, 248 Vaulting of , 76-78, 82, 101, 102, 250-252 Earls Barton Church, 235-236, 240 Early English architecture, 306-307 Eastbourne— S. Mary, 14, 360 Ebrach Church, 160 Eltham Palace, 369 Ely- Cathedral — Date of, 243 East end of, 247, 305-306 Font in, position of, 162 Galilee porch of, 300 Galleries in, 7 Lady Chapel of — -Crocket in, 130 Sculptured figures in, 124 Lantern of, 152, 315 Nave of — Ceiling of, 91, 245 Length of, 246 Piers of, 250 Niches on buttresses of, 50 Piers in, 250 Plan of, 151-153 Proportions of, 248 Shafts in, 28 Towers of — Central, fall of, 245 Western, 106 Transepts of, 248 Triforium of, 249 Vaulting in, 91, 101, 102 Prior Crauden's Chapel, 128 England — Black Death (1349), effect of, 12, 64, 129, 284, 310-311 Cathedrals of, contrasted with French, 281 Cistercian influx into, 295 East ends of churches in, 161, 237- 238, 247 Gothic architecture in — (1200-1250), 300-307 (1250-1300), 307-310 (14th and 15th centuries), 310-321 Early English, 306-307 East ends, 305-306 Triforia, 309-310, 314-315 West fronts, 304-305 Parish churches in — Lighting of, 363, 364 Number and variety of, 358 Plan of, usual, 359 Proportions of, 360-361 Roofs of — Construction of, 367-371 Decoration of, 371-372 Material of, 358, 362, 366 INDEX. 387 England— continued. Parish churches in— continued. Roofs of — continued. Pitch of, 363, 366-367, 371 Seotional ordinance of, 362-363 Spaciousness of, 361-362 Peasant Revolt (1381), 311 Perpendicular style of 15th-century work in, 284 Romanesque architecture in, early examples of — Long and short work, 236 Pilaster treatment, 235-237 Plan of, general, 238 Proportions of, 238-239 Square east ends, 237-238 Windows, 239-240 Romanesque cathedrals and churches of 11th and 12th centuries in — Alterations frequent in, 245-246 Eastern arm, length of, 247 Internal divisions, proportions of, 248-249 List of, 243-244 Naves, 250-252 ; length of, 246-247 Transepts, 247-248 Triforium design, 249-250 West fronts, 252-253 Severance of, from Normandy, 300 Transitional work in (12th century), 296-298 West, group of churches in, showing architectural progression, 322 Enlart, M., cited, 242, 255-257, 259, 264 note, 284 Enrichment, see Decoration Entasis — Columns, of, 137, 179 Spires, of, 114 Entrances of churches, English and French contrasted, 305 Erfurt- Cathedral, 333 S. Severus, 330 Escomb Church (Durham), 234, 238 Eton College Chapel, 88, 317 Evreux Cathedral, 104, 259 Exeter Cathedral — Alterations in, 246 Carving in, 129 Date of, 244 Piers in, 29 Exeter Cathedral — continued. Plan of, 146, 151, 152, 312 ; illus trated, 313 Towers of, 105 Vaulting of, 86, 91, 101, 102 Wall thiokness in, 34 Fairpord Church (Gloucs.), 133 Figeac— S. Sauveur, 282 Fire, frequent destruction of churches by, 245-246 Flamard, Bishop, 251 Flamboyant style, 283-284 Flint- Towers of, 108 Walling of, 43 Florence- Campanile, 107, 339 Cathedral — Dome of, 338 Piers of, 340 Plan of, 343 Sections of, 341 Windows of, 342 Loggia dei Lanzi, 339 Loggia del Bigallo, 11, 339 Rise of, 166 S. Croce — Chapels of, 343 Columns of, 340 Fresco paintings in, 119 Gallery of, 340 Plan of, 345 Wmdows and roof of, 342 S. Maria Novella, 343 S. Miniato — Ceiling of, 251 Date of, 25 note * Interior of, 172 Marble veneering in, 117, 136 Piers of, 25 Plan of, 173-175 Raised chancel of, 163 Roof of, 366 Flour Cathedral (Auvergne), 272 Fontenay Church, 159 Fontevrault Abbey, 207, 208, 210 Fortified churches, 106 Fountains Abbey — Addition to (13th century), 305-306 Date of, 10, 297 Design of, 298 388 INDEX. Fountains Abbey — continued. East end of, 161 Interior arrangement of, 145 note Vaulting of, 84, 91, 215 Fontfroide Abbey, 216 France (for districts, towns, churches, etc., see their names) Arches in, 78 Buttresses in, flying, 51 Carving in, 128 Galleries of churches in, 7 German early buildings under influ ence of, 143 note Gothic in, late survival of, 285 (see also Prance, North — Cathedrals and Prance, South — Gothic) East ends in, compared with Eng lish, 152 Early churches in, 136-137 Flamboyant style of 15th-century work in, 283-284 Height of churches in, 46, 98, 100, 103 Hundred Years' War, effects of, 12, 20, 262 Mouldings in, 20-21 Piers and columns in, examples of, 31-33 Pinnacles in, 100 Plain wall-space in churches of, 272 Sculpture in, schools of, 124-125 Side chapels in cathedrals of, 153-155 Spires in, 114 Vaulting shafts in, 93-94 Vaults in, 74, 82, 83 Windows of churches in, 58-62 France, Central — Arches in, 16 Gothic architecture originating in, question as to, 259 Priority of, in architectural art (12th to 13th century), 258 France, North — Architects and craftsmen in (12th to 13th century), 256-257 Cathedrals of — Chevets of, 267-268 Dates of, 262-263 Doorways of, 279-280 Double aisles of, 156, 265-266 English oathedrals contrasted with, 281 Interiors, spaciousness of, 266-267 France, North — continued. Cathedrals of — continued. Proportions of, 265 Sites of, 264-265 Towers of, 275-278 Transepts of, importance of, 277 Triforia of, 269-272 West fronts of, 278-281 Later Gothic work in, 281-282 Romanesque churches in, west fronts of, 230 Windows of churches in, 55 note, 56 France, South — Aisleless churches in, 156-158, 205, 208, 210 Barrel vaulting in, 212-216 Brick buildings in, 13 Buttressing in, 47 ; niches, 50 Capitals in — Concave, 37, 39 Convex, 37 Corinthian, 204 Carving in, 126 Classic influence in, 203-204 Columns in, 23 Domed churches in — Aisled, 211-212 Aisleless, 205-210 English influence on 14th-century churches in, 282-283 Gothic churches in, Roman influence in, 286 Transepts, omission of, 275 note ' Vaulting in, 79, 86, 95 Windows of churches in, 56 France, South- West, domed churches in, 9 France, West, English influence in 14th-century ohurohes in, 282-283 Frankfurt cathedral, 333 Freiburg cathedral, 333 Fresco painting, 119, Furness abbey, 159 Gables — Buttresses, on, 49 Lucarnes, over, 114 Towers, on, 112 Galleries, external open — Bari and Trani, at, 183 Germany, in, 194 Galleries in churches, 5-7 INDEX. 389 Gedney, S. Mary Magdalene, 361 Germany (fen- towns, churches, etc., see their names) — Alternation of large and small sup ports in churches of, 26 Apses in, 199-200 ; double, 191 Arcading in, 197-198 Arches in, pointed, date of, 11 Backwardness of architecture in, after 12th century, 190 Brick buildings in, 43, 328 Capitals in, 36-37 Carving in, 127 Choirs of churches in, 332-333 Colour decoration in, 121 Columns in, 23, 137 Cupolas and turrets of ohurches in, 103, 171 Early churches in, 136, 137 Exterior of churches in — Compactness of, 192 Simplicity of, 194 Western facades, 198 French influence on buildings in, 143 note, 325-328 Galleries- External, 194 Internal, absence of, 196 Hallenkirchen, 8, 330, 355 Lateral entrances of churches in, 194 Masonic dexterity exhibited in, 333 Mouldings in, interpenetration of, 21 Piers in, 31 Roofs of churches in, 332 " Round-arched Gothic " work in, 78 Spires of churches in, 112, 114-115 ; pierced, 333 Towers and turrets in, 193, 198, 200 ; single western towers, 333-334 Triforia, omission of, 7, 331 Vaulting in, 195 Windows of churches in, 56 Germigny-les-Pres Church (Loiret), 139-140 Gernrode Church, 196 Gerona Cathedral, 352-353 and note Giotto, 107, 339 Glasgow Cathedral, 323 Glass in church windows — Colour in, 133-134 Early examples of, 131 Figures in, 132 Glass mosaio, see Mosaic Glastonbury Abbey, 298 Gloucester Cathedral — Abaci in, 34 note Capitals in, 37 Carving in, 131 Chantry ohapels in, 155 Columns in, 23, 25 Date of, 243, 244 note Eastern arm of, 247 Galleries in, 7 Interior arrangements of, 145-147 note 2 Nave of, 246, 250 Panelling in, 246, 315 Proportions of, 248 Reconstruction of, 246, 312 Roofings of, 245 Tower of, 107 Transepts of, 248 Vaulting of, 88, 91-92 Windows of, glass in, 133 Gloucestershire, towers in, 111 Goodyear, Mr., cited, 110, 179 and note, 180 and note Gothic, meaninglessness of term, 3 note Gothic architecture — Curves in, theories as to, 179-180 and note England, in, see tinder England France, in, see France, North — cathedrals, and France, South — ¦ Gothic Germany, in, see under Germany Italy, in, see under Italy Renaissance style mixed with, 284- 285 Rise of, 259 Spain, in, see under Spain Whitewashing of interiors, 335, 337 Green, J. R., quoted, 307 note Greenstead Church, 233 Guingamp Church, 282 Gurgoyles, 54 Hand centers, 85 Heckington Church (Lines.), 368 Hereford Cathedral — Clerestory windows of, 309 Columns of, 23 Date of, 243 East end of, 247 39° INDEX. Hereford Cathedral— continued. Nave of, 246, 250 Tower of, 107 Vaulting in, 91 Hexham Church — Crypt of, 234 Date of, 306 Windows of, 57 Hildesheim— S. Michael — Columns and piers in, 31, 137 Turrets of, 193 Hitcham Church, 63 Hodges, C. C, quoted, 251 Holland — Brick buildings in, 43 Churches in, 335, 337 Houghton-le-Dale Chapel, 43, 44, 64 Huesca— S. Pedro, 349-350, 355 Hugh, Bishop of Lincoln, 299-300 Huish Episcopi (Somerset), 111 Ipeley Church, 359 Ingolstadt — Frauenkirche, 333 Ireland, Christianity in, 232 Issoire Cathedral (Auvergne) — Angouleme Cathedral compared with, 209 Arcading of, 182 Arches in, 14, 219 Capital in, 36 Date of, 216 Exterior of, 218-219 Interior of, 219 Plan of, 217 Windows at crossing of, 218 Italy (for towns, churches, etc., see their names) — Alternation of large and small sup ports in, 26 Arches in — Extrados of, not concentrio with intrados, 183 Horseshoe, 14 Pointed, date of, 11 Tradition influencing, 16, 78 Buttressing in, 46 Capitals in, 37 Churches in, supports of, 82 and note Classic traditions dominant in, 2, 4 Clerestories omitted in churches in, 8 Columns in, 23 Dosserets in, 35 Italy — continued. Gothic architecture in — Capitals in, 340 Characteristics of, 338-339 Choirs in, 343-344 Early examples of, 339 Internal ordinance in, 340-342 Mural decoration in, 342 Scale of, 342-343 West fronts, 344-345 Mouldings in, 20 Piers of churches in, 24 Porches in, 172-173 Shafts inlaid with mosaics in, 24 Spires in, 114 Towers of churches in, 107 Tuscan churches, 173-180 Windows of churches in, 56 Italy, North- German influence in, 166 Lombards, 165 Italy, South — Byzantine and Norman rule in Apulia, 166-167 Byzantine-Romanesque work in, 167 Carving in, 126 Churches of, 180-184 Jedburgh Abbey — Interior of, 197 note 3 ; illustrated, 322 Piers in, 26 Plan of, 32 Jerusalem, Church of the Holy Se pulchre at, 141-142 Jervaulx Abbey, 306 Jumieges Abbey — Date and style of, 242 Towers of, 111 Triforium of, 249 Kelso Abbey, 244, 321-322 Kirkstall Abbey, 57 La Chaise Dieu (Auvergne), 288, 290 La Martorana Church, 185 note Laach Abbey Church — Apses in, 192 i Entrances of, 194 [ Western atrium of, 198 INDEX. 391 Labels, 17, 22 Lanterns, 104 Laon — Cathedral — Buttresses of, 96, 269 Carving in, 129 Chapels of, 154 Columns in, 31, 81 note, 261 and note Date of, 262 East end of, 269 English features of, 269 External arcading of, 279 Galleries in, 7 Horizontal lines of, 52 Mural decoration in, 116-117 and note Plain wall space in, 272 Plan of, 153 Rebuilding of, funds for, 255-256 Shafts in, 32 Towers of, 111, 275-277 ; effigies of oxen on, 255 Triforium of, 260, 269-270 • Vaulting of, 79-81, 96, 102 West front of, 278 Windows of, 69 Templar church at, 142 Laud, Archbishop, 319 Lausanne Cathedral, 197 Laval — La Trinite', 227 Lavenham Church, 30, 311 Le Mans — Cathedral — Apse of, 267-268 Chevet of, 149 Choir of, without triforium, 271 Dogtooth ornament in, 269 Double aisles at east end of, 156, 266 Exterior of, 263, 265; illustrated, 149 Interior arrangement of, 274 Lady Chapel of, 152 Mouldings in, 20 Nave of, 260-261 Pinnacles of, 101 Plan of, 149 Proportions of, 272 Sculptured figures in, 125 Windows in, 132 S. Pierre de la Couture, 131, 226-227 and note Le Morvan, churches of, 221 Le Puy — Cathedral — Diaper work in, 218 note'' Piers in, 32 Plan of, 212 West end of, 231 S. Miohel de l'Aiguille, 231 note Lead, gurgoyles of, 54 Leeds, S. John Baptist Church, 319 Leominster Church, 67, 131 and note Leon — Cathedral, 351-352 S. Isidoro, 219 and note, 350 Lerida Church, 350 Lessay Abbey, 242-243 Lichfield Cathedral — Clerestory windows of, 309 Choir, etc., of, 313-314 Shafts in, 28 Spires of, 114 Towers of, 107 Vaulting of, 86, 90, 91, 101, 102 Liernes, 86, 88 and note 2, 91 Lighting of churches, 71, 169-170 Limburg-on-the-Lahn Cathedral, 197 note \ 198, 324 Lincoln — Cathedral — Angel choir of — Cusping in, 67 Dimensions of, 309 Sculpture in, 124 Section of, 308 Vaulting of, 82, 300 Window screens of, 309, 314 Arcading in, double, 300 Buttresses in, 97 Capital and crockets in, 130 Chantry chapels in, 155 Clerestory windows of, 310 Dates of, 243, 299 East transepts of, 278 Eastern arm, length of, 153 Labels, interior, in, 22 Plan of, 151-153, 299; illustrated, 151 Roof window of, 305 Screen of, 178 Shafts in, 28, 29 Side chapels of, 154 392 INDEX. Lincoln — continued. Cathedral — continued. Towers of, 104, 107 Triforium of, 271 Varieties of Gothic exemplified in, 300 note Vaulting in, 81, 83, 86, 88 note1, 91, 93, 101, 102, 310 West front of, 230, 304 Western porches of, 253 Windows of, 69 ; roof, 305 S. Mary, 232 S. Peter, 232 Lisieux— Cathedral (S. Pierre) — Chevet of, 301 Columns in, 261 Mouldings in, 20 Nave of, 258-259 Style of, 261 Vaulting in, 80 S. Jacques — Date and style of, 285 Decoration in, 121 Plan of, 150 Llandaff Cathedral, 244 Loches— S. Ours, 212-213 Lombardic designs — Influence of, on 11th-century English Romanesque, 242 Roofs, 170 Towers, 109, 110 Lombards, 165 London — ¦ Charterhouse, roofing of, 369 Inns of Court, roofing of, 369 Parliament, Houses of — Panelling of, 30 and note Victoria Tower of, 51 note l S. Augustine's (Kilburn), 158 S. Mary Aldermary, 90 S. Paul's (old)— Date of, 243 Tower of, 107 S. Saviour's (Southwark), 91 Temple, the, 142 Tower of— S. John's Chapel, 71, 91, 95 Long and short bonding, 236 Long Melford Church (Suffolk), 360 Louth Churoh (Linos.), 114 Lucarnes, 114 Lubeck — Marienkirche, 328-331 S. James, 330-331 S. Peter, 330-331 Lucca, arcaded gallery fronts in, 171, 176 Lucca Cathedral (S. Martino), 338-339, 340-341 Liineberg Church, 330 Magdalen La ver Church, 358 Mainz Cathedral — Arcading in, 197-198 Columns in, 31 Crypt of, 191-192 Double apses of, 191-192 Entrances of, 194 Proportions of, 197 Turrets of, 193 Vaulting in, 195 Malmesbury Abbey — Date of, 10, 297 Doorway of, 16 Nave of, 298 Parapet of, 53 Pinnacles of, 100 Triforium of, 249 Vaulting of, 78 Malvern Priory, 243, 317 Manchester Cathedral, 156 Mantes Cathedral, 262 Marble- Alternating colours of, 117, 175, 341 ; roundels of green, 109 Brick alternating with, 20, 43, 109, 117, 175, 345 Facing with, 43 Stone in combination with, 117 Tesserse, 175 Veneer of, 43, 117, 175 Marburg— S. Elizabeth, 330, 333 Masons and masonic guilds — Influence of, 3, 5, 20 Position of master masons in France (12th to 13th century), 256-257 Maulbronn Abbey, plan of, 159 Mediaeval architecture — Origin of, 1-2 Period of, 3-4 Style of, 2 Melrose Abbey, 314, 321-323 Mettlach Church, 138 Metz Cathedral, 271, 328 INDEX. 393 Milan- Cathedral — Apse and transepts of, 344 Exterior of, 346, 347-348 Interior of, 346, 348 Lighting of, 346 Piers in, height of, 355 Plan of, 343 note, 346 S. Ambrogio contrasted with, 348 Sections of, 347 Window tracery in, 346 S. Ambrogio — Arches of, 16, 239 Burial crypt in, 163 Carving in, 127 Cathedral of Milan contrasted with, 348 Date of, 168 Galleries in, 7 Lighting of, 169-170, 218 Piers in, 30 Pilasters on tower of, 236 Plan of, 168-169 and notes Squinch arches of, 212 Tower of, 109 Vaulting of, 8, 75, 79-80 West front of, 170, 344 S. Lorenzo, 140, 147 note l S. Satiro — Pilaster decoration of, 235 Plan of, 140 Tower of, 109 S. Vincenzo-in-Prato, 170 Moissac Abbey, 127 Molfetta Cathedral, 181, 183-184 Monastery churches, 145 Monastic Orders, see Benedictine and Cistercian Monkwearmouth Church, 240, 241 Monreale Cathedral — Arcading at, 52 Carving in, 126 Cloisters of, 185 Date of, 186 Detached shafts of, 186 Exterior of, 230 Mosaics in, 5, 118 Plan of, 187 Windows of, 64 Montier-en-der Church, triforia in, 136 note Montmajour — Abbey, 216 S. Croix, 140 Montreal Church (Burgundy), 179, 288 Monumental slabs and brasses, 133 Moore cited, 75 note Morienval Church (Oise)— East end of, 74 and note 3, 148 Towers of, 105 Mosaic, glass, 24, 118 and note, 175, 180-181 Mouldings (see also Carving) — Arches, of, painting and gilding of, 120 Bowtell — Pointed, 18 Rounded, 17, 20 Contemporary, similarity between, 21 Different planes, on, 19-20 Dogtooth, 129, 269 Hood, superfluous nature of, in Eng lish cathedrals, 22, 272 Interpenetration of, 20-21 Material determining form of, 18, 21 Ogee, dates of, 18-19 Plinths, of, 52 Roman and mediaeval, contrasted, 16 Romanesque period, of, 22 Sections of mediaeval, nature of, 17 note Timber roofs, of, 367 Torus, see subheading Bowtell Transitional period, in, 18 Vault ribs, of, 84-85 Windows, of, 59 Mouthiers Church, 208 Much Wenlock Abbey — Bay of arcade in, 40 Bay of triforium in, 270 Date of, 298 Piers in, 26 Minister Cathedral, 137 Mural decoration, see Decoration Narthices or porches of churches — Burgundy, in, 222-223, 247 Italy, in, 172-173 Normandy, in, 243 Naumburg Church, 324 2c3 394 INDEX. Naves of churches — Divisions of wall of, 5 Length of, in early Romanesque in England, 246-247 Netley Church (Hants), 63 Neuss— S. Quirin, 199 Neuvy — S. Sepulchre, 140 Nevers— S. Etienne, 217, 218 note » Nibs (pilasters), 24 Niches, canopied, on piers, 33 Nimes, Baths of Diana at, 71, 138 Norman artificers, 180-181 Norman influence, 184-185 Normandy — Capitals in, 36-38 Carving in, 127-129 Clustered piers in, 26, 30 Romanesque churches of 11th cen tury in — Examples of, 242-243 Internal divisions, proportions of, 248-249 Transepts of, 247-248 Severance of, from England, 300 Towers of, 106 Northampton, Templar Church at, 142 Norwich — Cathedral — Choir of, 247 Columns of, 23 Date of, 244 Eastern arm of, 247 Interior arrangement of, 146, 147 note, 150 Nave of, length of, 246 Plan of, 150 Proportions of, 248 Roofings of, 245 Spires of, 114 Tower of, 106, 107 Transepts of, 248 Triforium of, 249 Vaulting in, 88, 91 S. Peter's Mancroft, section of, 365 Churches in, number of, 358 Novara — • Baptistery at, 161 Cathedral, 168 Noyon Cathedral — Choir of, 260 note ' Columns and piers in, 81 note Galleries in, 7 Noyon Cathedral — continued. Nave of, 258 Plan of, 260 Triforium of, 270 Vaulting of, 81 Niirnberg — Liebfrauenkirche, 121 Lorenzkirche, 332 S. Sebald, 332 Oak, painting and gilding of, 121 Odda's Chapel (Gloucs.), 45, 236 note, 238 Orange — Cathedral, 216 Theatre, 51 Orcival Church, 216 Orleans — Cathedral, 268, 279 S. Aignan, 272 Orvieto Cathedral- Gallery and arches of, 341 Marble in, 175 West front of, 344 Windows and roof of, 342 Ottmarsheim Church, 139 Ouistreham Church, 77, 243, 249 Oversailing courses, 53 Oxford — Christ Church Cathedral, 25, 91 Christ Church CoUege, 319 Painting, mural, see under Decoration Palermo — Cappella Palatina, 184 Cathedral, 185 note, 186-187 La Zisa Palace, 184-186 S. Catalda, 186, 188-189 S. Giovanni degU Eremeti, 188 Palestine influence, 185-186 PaneUing — 14th and 15th century, 315, 317 Buttresses, of, 50 Exterior walls, of, 317, 364 Flint, 43 Marble, 188 Piers and arohes, of, 30 Pilasters and string courses, by, 51 Stone, 71, 111, 246 Wood, on ceilings, 342 Parapets, 53, 114 Paray-le-Monial, 222, 225 INDEX. 395 Paris- Cathedral, see subheading Notre Dame La Sainte ChapeUe— Date of, 262, 281 Window tracery in, 59, 300 La Trinite (Place Blanche), 158 note Notre Dame — Apse of, 268 Buttresses of, flying, 97, 269 Carvings in, 128 Chevet of, 149 Choir of, 268 Columns in, 31, 81 note, 261 Date of, 262 Double aisles in, 156 GaUeries in, 7 Horizontal lines of, 52 Interior of, 268 Open space before, 264 Plain waU space in, 272 Proportions of, 265, 272 Roof of, 278-279 Sculpture in, 123-124 Shafts in, 32 Side chapels of, 153 Towers of, western, 327 ; transept towers omitted, 277 Triforium of, 260, 269-270 and note Vaulting of, 79, 81, 85 note, 102 ; original, 215 WaU thickness in, 34 West front of, 278, 279 Window tracery in, 59, 69 S. Etienne du Mont, 35 note S. Eustache, 225 S. Germain des Pres, 31 Parma — Baptistery at, 161 Cathedral, cupola towers of, 171 Patrington (Yorks.)— S. Patrick, 114, 361-364 Patrixbourne Church (Kent), windows in, 60, 62, 161 Pavia — S. Michele, 168 Carving on west front of, 2 Cupola towers of, 171 East end of, elevation and plan of, 171 GaUeries of, internal, 7; external, 171 Roof of, 170 Pavia — continued. S. Pietro in Cielo d'Oro, 168 Pendlebury — S. Augustine's, 158 Perigueux — S. Etienne, 205-206, 207 note S. Front Cathedral, 203, 205-209 Perpendioular style, 284 Perpignan Cathedral, 158, 290 Pershore Abbey — Date of, 243 Tower of, 104 Triforium omitted in, 8, 314 note l Peterborough Cathedral — Capitals in, 25, 37 Columns in, 25 Date of, 244 and note East end of, 247 Font in, position of, 162 Galleries in, 7 Nave of — Ceiling of, 91, 245 Ely nave compared with, 250 Length of, 246 Proportions of, 248 Towers of, 107 Triforium of, 249 West front of, 304 Windows of, 55 note, 60 Piacenza Cathedral, 172 Piers — Alternating, of differing section, 250 Alternation of, with columns. 26, 137, 175, 242, 251 Apses, of, in French cathedrals, 268 Clustered — Development of, 26 Forms of, 26 Normandy, in, 30 Crossing, at, in cruciform churches, 103-104 Early use of, 24 Evolution of, 24-25 Niches on, canopied, 33 Oblong, 26, 29 Rebated angles of, 26 Shafts attached to, see Shafts Spanish churches, in, 355-356 T -shaped, 25 and note l Thickness of, 2 Pilasters — Fluted, 32 396 INDEX. Pilasters — continued. Panelling by, 51 Towers, on, 235-236 Transverse arches supported on, 24 Pinnacles, 49, 100, 114 Pisa — Arcaded gallery fronts in, 171, 176 Baptistery at, 161 Cathedral — Arcading in, 176 Exterior of, 178 Funds towards building of, 166 note Mural decoration in, 175 Plan of, 177-179 Upper abacus in, 35 Windows of, 177 Grouping of buUdings in, 177 Leaning tower of, 109 -110 S. Caterina, 338 S. Pietro-a-Grado, 173 Pisano, Niccolo and Giovanni, 339 Pistoja churches, 176 Planning of churches — Aisleless plans, 156-158 Ambulatory, eastern, 147 and notes Burial inside as affecting, 162-164 Byzantine, 137-142 Chancel, raised, 163 Changes in, after 11th century, causes for, 144 Chantry chapels, 155 and note Cistercian influence on, 158-161 Double aisles, 155-156 East ends, 146-153, 161 Entrances, lateral, 193-194 Fonts, position of, 162 GaUeries, external open, 170-171 Porches, 172-173 Roman, 137-138 Side chapels, 153-155 Transepts, 153 Plinths, 52 Poitiers — Cathedral — Curves in, 179-180 Lighting of, 287 Piers in, height of, 355 Plan of, 287-288, 330 Vaulting of, 79, 86 S. Hilaire, 211 Polebrook Church (Northants), 120 Polignac Churoh, 32, 216, 217 note Pont l'eveche Church, 150 Pontaubert Church (Burgundy), 222, 288 Poore, Bishop, 301 Porches of churches, see Narthices Prior cited, 298 Quimper Church (Brittany), 282, 364 Ravello Cathedral, 180 Ravenna — S. ApoUinare-in-Classe — ¦ Arcading of, 235 Burial crypt in, 163 and note s Tower of, 107-108 S. ApoUinare Nuovo — Mosaic decoration in, 5 Tower of, 107, 193 note S. Giovanni EvangeUsta, 109 note S. Vitale— Carving in, 127 Plan of, 139 Towers of churches in, 106, 107, 109 note, 193 note Reddes— S. Pierre, 138 Reims — Cathedral — Apse of, 267 Buttresses of, flying, 98-100 Chevet of, 149 Commemoration of master mason connected with, 256 note Construction of, 275 Date of, 262 Double aisles of, 156 Exterior of, 279 Horizontal lines of, 52 Nave of, length of, 152 North transept doorways of, 277 and note Piers in, 32 Plan of, 266 Proportions of, 265 Sculpture and carving in, 124; sculpture removed for windows, 280 note 1 Section of, 99 Side chapels, absence of, 153 Stone seat round, 275 note 2 Towers of, 111, 276, 277 Triforium of, 271 Vaulting in, 34, 102 West front of, 278 INDEX. 397 Reims— -continued. Cathedral — continued. Windows of — Glass in, 133 Sculpture removed for, 280 note » Tracery of, 58, 69 S. Remi — Capital in, 36 GaUeries in, 7 Plan of, 136 Renaissance architecture, Gothic style mixed with, 284-285 Responds, function and forms of, 30 Restoration of churches, 116-117, 122 Ridge ribs, 86, 88 note ' Ripon Cathedral, 234, 296 Rivaulx Abbey, 306 Rochester Cathedral — Crypt of, 71, 91 Date of, 243 East end of, 153, 161, 247 Nave of, 91, 244, 246 Roman influence on plans, 137-138; few supports, 210, 286 Roman towers, 109 Romanesque, signification of term, 3 Romanesque architecture — Byzantine-Romanesque, 167 England, in, see under England Normandy, in, during 11th century, see under Normandy Rome — BasUioa of Constantine— Proportions of, 169 Recesses in, 157 Vaulting of, 74 note l ; abutment of vaults, 96 note, 157 CaracaUa's baths, 79 S. Agnese, 6 S. Clemente, 24, 119 S. John, 156 S. Lorenzo, 6 S. Maria in Cosmedin, 24, 109 S. Maria in Trastevere, 136 S. Maria Sopra Minerva, 339 S. Paolo fuori le Mura, 136, 156 S. Peter (BasiUcan Church), 156 SS. Giovanni e Paolo, 109 Romsey — Abbey — Choir bays of, 322 East end of, 147, 161 Romsey — continued. Abbey — continued. Nun's Church, 244, 247 Plan of, 148 Windows of, 58 Roofing of ohurohes — Germany, in, 332 Gothio cathedrals, of, 52 Timber, 187 (see also under England- Parish churches) Roslyn Chapel, 71 Roueiha, church at, 25 note ' Rouen — Cathedral — Apse of, 267 Date of, 262 Proportions of, 274 Sculpture in, 123 Towers of, 276, 277 Transept doorways of, 277 West front of, 230, 280-281 Detached shafts in churches of, 32 S. Maclou— Flamboyant style of, 284 Lantern tower of, 104 Mouldings in, 20 Supports in, 33 S. Ouen— Apse of, 268 Dates of, 282 Glazed triforium of, 271 Niches in, 33, 123 Proportions of, 282 Roofing of, 271 Sculptured figures in, 124 Shafts in, 33 Towers of, 333 Vaulting in, 102 WaU thickness in, 34 Window of, 69 Royat Church (Auvergne), 106 Rubble- Columns of, 23 Regular courses of, 45 Vault webs of, 83 WaUs of, 15, 45 Runic crosses, 241 S. Albans — Cathedral — Alterations in, 246 Arches in, 17 398 INDEX. S. Albans — continued. Cathedral — continued. Ceiling of, 91, 366 Choir of, 247 Date of, 243 Eastern arm of, 151, 247 Galleries in, 7 Incongruities of style in, 262 note Interior arrangement of, 145, 147 note2 Nave of — Altars in, 144 Length of, 246 Paintings in, 120 Piers in, 26 Roman brickwork in, 43 note Shrine of S. Alban in, 121, 152 Tower of, 106 Vaulting of, 91 S. Martin's Church, 34 S. Avit Senieur (South France), 77 note S. Cross Abbey, 91, 161 S. David's Cathedral, 243 S. Denis (near Paris) — Date of, 255, 259 Eastern transept of, 136 Relics in, 147 Style of, 259 Vaulting of, 278 S. GaU Monastery Church — Double apses of, 191 note 2 Plan of, 144 Towers of, 193 S. Generoux, 204 S. Martin de Londres, 140 S. Nectaire Church, 216-217 S. Paul, M. Anthyme, cited, 142 and note 1, 255 S. Saturnin (Auvergne), 148 Salamanoa New Cathedral, 355-356 Salamanca Old Cathedral, 350-351 Salisbury Cathedral — Altars in, 153 Amiens Cathedral compared with, 301-304 Arcades of, 103 Buttresses of, 49, 97 Chantry chapels in, 155 Choir of, 146, 152 Colour effects in, 299 Columns of, 28, 117 Dates of, 301-302 SaUsbury Cathedral — continued. East end of, 305 ; Ulustrated, 152 East transepts of, 153, 278 Plan of, 151, 152, 302 Proportions of, 302-303 Shafts in, 26-28 Spires of, 114 Symmetry of, 302 Tower of, 107, 276 Triforium of, 271 Vaulting of, 34, 90, 91, 101-102 View of, from north-east, 304 WaU thickness in, 33 West front of, 302, 304 Salonica — S. Demetrius, 24, 178 note 1 Santiago de Compostella Cathedral, 219, 280 and note 2, 350 Saracenic influence, 183-186, 219 Saulieu Church, 127 Saumur — Notre Dame de NantiUy, 216 S. Pierre- Plan of, 226, 229-230 Vaulting of, 79, 86, 199, 228 Saxon masonry, 236 and note Scoinson, 68 Scotland, ruined abbeys in, 321 (for particular churches see their names) Scott, Sir GUbert, quoted, 70, 88 Sees Cathedral — Clerestory window screen of, 314 note* Paintings in, 123 Shafts, detached, in, 263 Vaulting in, 102 Segovia Cathedral, 355-356 Selby Church, 62, 305 Semur Church, 222 SenUs, S. Frambourg, 58 Sens Cathedral — Columns in, 32 Date of, 259, 264 Flying buttresses of, 269 Open space before, 264 Transept doorways of, 277 Sessa Cathedral, 180 SeviUe Cathedral, 355 Shafts- Angle, 75, 76 Attached, examples of, 26, 33 Buttress angles, at, 49 Counter-changing of, 32 INDEX. 399 Shafts — continued. Detached — ¦ Examples of, 26-28, 32 France, in, 32 Joints of, 27-28 SicUy, in, 186 Vaulting, 93-94, 101-102 Weight on, 28-29 Windows, of, 56 Sharpe cited, 131 note, 203 note Sherborne Abbey — Alterations to (1436-1504), 317 Date of, 244 PaneUing of piers in, 30 Vaults of, 92 note Shoreham, New, Church (Sussex), 43 Shoreham, Old, Church (Sussex)— East end of, 161 Material of, 43 Plan of, 360 Shrewsbury — S. Mary, 18 SicUy, churches of — Arcading on, 52 Arches of, 185-186; date of pointed arch, 11 and note Carving in, 126 Decoration of, by inlays, 186-187 Domes of, 189 Examples of, 184-189 Mixed architecture in, 167-168 Shafts inlaid with mosaics in, 24 Variety of influence in, 184-185 Siena Cathedral — Alternating marbles in, 175, 341 Dome of, 338 GaUery of, 341 West front of, 344 Windows of, 341 SUchester Church, 232, 237 Soissons Cathedral — South transept of, 260 Triforium of, 270 Windows in, 58 note SoUgnac Church, 208 Somersetshire, towers in, 111 Sompting Church, 237, 240, 241 SouiUac, church at, 208 Southwell Minster, 129-130, 244 Spain — Classic traditions in, 39 note Gothic churches in — Choir, position of, 354- 355 and note Spain — continued. Gothio churches in — continued. French influence on, 349-353 Piers, weight of, 355-356 Saracenic influenoe in, 219 note Speier Cathedral — Arcading in, 197-198 Columns in, 31 Entrance of, west, 194 Galleries of, external, 194 Plan and section of, 196 Proportions of, 197 Turrets of, 193 Vaulting in, 195 Western apse, absence of, 191 Spiers, R. Phene, cited, 208 note Spires of churches- Angle of, 114 Base of, 113 ; broaches at base, 113- 114 Entasis of, 114 Flying buttresses on, 114 French examples of, 111 Germany, in, 333 Lucarnes of, 114 Octagonal, 112, 115 Roofings of, 113 note Sixteen-sided, 115 Stone, 111, 113 ; construction of, 114 Timber, 111-113; construction of, 112-114 Stanton Harcourt Church, 51 Stone — Brick alternating with, 117 Marble in combination with, 117 Pierced panelling of, 246 Pierced slabs of, for windows, 239 Spires of, see under Spires Tiles alternating with, 136 Walls of, 43 Stone Church (Kent), 60, 61 Strassburg Cathedral, 328, 333 String-courses, 51-53 Street cited, 219 note, 275, 349, 353 note ; quoted, 352, 354 Stucco, 237 ; tempera painting on, 119 Suger, Abbe", 255 Supports, Roman method as to, con trasted with Gothic, 210, 286 Switzerland, churches in — Arcading of, 197 Entrances of, 194 400 INDEX. Switzerland, churches in — continued. Windows of, 56 Symbolism, semi-barbaric, in carving, 2, 127 Tag e'Ivan, 215 Tarragona Cathedral, 350 Tas-de-charge, 93 and note, 94 Taunton— S. Mary Magdalene, 111 Templar churches, 142 and note - Tewkesbury Abbey — Abaci in, 34 note Capitals in, 37 Chantry chapels in, 155 Chevet of, 150 Columns of, 23, 25 Date of, 243 Nave of, 246, 250 Proportions of, 248 Roofings of, 245 Tower of, 104, 106 Vaulting of, 84 note \ 88, 91 West front of, 252-253 West window of, 253 note Thaxted Church (Essex), 360 Thrusts — Arches, of, 47 Buttresses' transmission of, 97 Concentration of, 71, 72, 95 Counterpoise of, summary regarding, 101 Timber roofs, of, 99 Vaults, of, compared, 88, 98 WaU ribs, of, 95 Tiercerons, 86, 88, 91 Tiles alternating with stone, 136 Timber- Churches of, 233 Roofs of, 187 (see also tinder Eng land — Parish churches) Spires of, see under Spires Vaults of, 314 note 2, 315 Tintern Abbey, 28 Tiverton Church, 111 Toledo Cathedral, 351-352 Torcello Cathedral — Burial crypt in, 163 Mosaics in, 118 Upper abacus in, 35 Toro, Collegiate Church at, 350-351 Torregiano, 318 Toscanella— S. Maria and S. Pietro, 173 Toulouse — Augustines, Church of the, 292 Cordeliers, Church of the, 292 Jacobins' Church at, 157, 290-292, 355 S. Sernin — Apsidal chapels of, 148 Eastern arm of, length of, 247 Nave of, length of, 247 Plan of, 217, 219, 220 Relics in, 147 Vaults of, 10 mentioned, 292 Tournai — Cathedral — ¦ Apses of, 200-202 Spire of, 112 Towers of, 105, 276-277 Transepts of, 260, 335 View of, from the south, 201 Churches, towers of, 202 Tournus, Burgundy — S. Phuibert— Arcading of, 133 Arches in, 298 Date of, 223 Dome of, 212 note Exterior of, 194 Narthex of, 222 Plans and sections of, 214 Vaulting of, 95, 213-215 Tours — ¦ Cathedral- Capitals in, 38 Date of, 263 Gable of, 279 Piers in, 32 Shafts in, 33 Tower of, western, 277 Vaulting in, 102 S. Martin, 147 Towers of churches — Batter of, 51, 105, 179 Beacon or watch towers, as, 108 Belfries, as, 107, 108 Buttressing of, 110 Campanile, 108, 110 Central, 103-104 France, in, 275-278 Gables on, 112, 243 Italy, in, three schools of, 108-109 Louvres of, 111 note Octagonal, 115; square with octa gonal top, 111 INDEX. 401 Towers of churches— continued. Pilaster treatment of, 235-236 Positions of, 103 Refuges, as, 106, 108 Round, 108 Saddle-back, 111-112 Transepts, over, 104-105, 334 West end, at, 104, 106 Tracery of windows, see under Windows Trani— Cathedral- Bronze doors of, 180 Columns of, 181 note Crypt of, 163 East end of, 161 Exterior of, 182 and note, 183 Tower of, 107 S. Maria Immaculata, 183-184 Churches in, pierced marble window slabs of, 181 Transepts of churches, 277-278 Transoms, 64-65 Treguier Cathedral, 104-105, 282 Trier- Cathedral — Entrances of, 194 GaUeries in waU of, 194 Plan of, 198-199 Vaulting in, 228 note 2 Liebfrauenkirche, 325-326 Triforium — EngUsh cathedrals, in, varieties of, 249-250 Nature and function of, 5 Omission of, 8, 196 Troja Cathedral, 175-176, 180-181 Troyes Cathedral, 156, 262, 266 Truro Cathedral, 162 note, 361 Tudor rose, 131 Tuscan churches — Examples of, 173-180 GaUeries, outside, 175-176 Tynemouth Priory, 57 Ulm Cathedral, 331-333 Valpolicella district (near Verona), double apse of church in, 191 Vaults (see also Arches) — Barrel, 70-71 Bays in, marking of, 71 and note Dates of, 74 Vaults — continued. Fan, 88-90 Flying buttresses' position in relation to, 97-98 Intersecting (groined), 71 List of English examples, 91 Oblong spaoes, over, 79, 82 Painted, 348 Plastered, 83, 86 Pointed arches of, 77-78 and note Ribbed, 72-77; Roman, 73-74 and note l Ribs of — Extra, 86 Sections of, 84 Setting out of, 92-95 Windows in relation to, 90 Sexpartite, 80-81; flying buttresses to, 269 Shafts of, 93-94, 101-102 Tas-de-charge, 93 and note Thrusts of, 95, 97 (see also Thrusts) Tie-beams' position in relation to, 99-100 and note Timber, 314 note 2, 315 Transverse arches of, 71-72 and note ', 79 Web of, 83-86 Vendome Abbey, 278 note 2, 284 Venice — Frari, Church of the, 339 S. Mark's, 166— Carving in, 127 Mosaics in, 118 S. Front compared with, 206-208 Tower of, 108 and note S. Stefano, ceiling of, 172 note, 342 SS. Giovanni e Paolo — Arches in, 20, 340 Chapels of, 343 Towers of Venetian school, 108- 109 VerceUi— S. Andrea, 339 Vermenton Church (Burgundy), 125 Verona — ¦ Cathedral, porch of, 172 S. Anastasia — Arches in, 20, 340 Chapels of, 343 Decoration in, 121, 342 Internal ordinance of, 341 Plan of, 344 402 INDEX. Verona — continued . S. Anastasia — continued. Sections of, 340 Tower and buttresses of, 345 S. Fermo Maggiore — CeiUng of, 172 note, 342, 366 Pilasters of, external, 179 S. Stefano, plan of, 147 S. Zeno— Arches of, 16, 172 Bronze doors of, 180 Capitals in, 1, 36 Plan of, 172 Porch of, 172 Raised chancel of, 163 Roof of, 172, 366 Supports in, large and smaU, 172, 251 Tower of, 105, 109, 112 Vezelay Abbey — Choir of, 264 ; columns in, 23 Date of, 221-222, 264 Narthex of, 222, 259 Nave of, length of, 247 Piers in, 32 Plan of, 221-222 Sculpture in, 124 SimpUcity of, 158 Vaulting of, 79 Vienna Cathedral — Plan of, 334 Proportions of, 330 Roofing of, 332 Towers of, 105, 334 Vignory Church (Haute-Marne), 148 ViUefranche-de-Rouergue Church, 47 note VioUet-le-Duo cited, 83, 85 and note, 86, 124-125 Vitre, gurgoyle at, 54 Von Steinbach, Irwin, 257 Walls — Arcading of, see under Arcading Decoration of, see Decoration Materials of, 43-45 Plastered, 45 Thickness of, 72 note'; above sup ports, 33-34, 45, 213-214 and note Walpole S. Peter (Norfolk), 361 Waltham Abbey — Columns of, 23 Waltham Abbey — continued. Date of, 244 Triforium of, 249 Warmington Church (Northants), 361 Warwick Church (Cumberland), 360 Waverley (near Farnham), Cistercian settlement at, 295 Wells- Bishop's palace at, banqueting hall in, 65 Cathedral — Buttresses of, sculptured figures on, 50, 124 Chantry chapels in, 155 Construction of, 299 Date of, 298-299 East end of, 146, 151 Piers and shafts of, 298-299 Presbytery and Lady Chapel of, 313-314 Screen of, 178 Towers of, 107 Triforium storey of, 299 Vaulting in, 91, 101-102 ; vaulting shafts, 299 West front of, 304-305 Westminster Abbey — Chapter-house of, 59 Chevet of, 150 Choir- Position of, 146, 247, 355 note Section of, 308 Vaulting of, 91 Cloisters of, vault over, 83-84 Columns of, 28, 117 Cusping in, 67 Doorway, north, of, 305 Flying buttresses of, 307 French influence in plan and propor tions of, 300, 307 Galleries in, 7 Henry VII.'s Chapel- Carvings in, 131 Fan vaulting in, 88-90, 92 PaneUing in, 30 and note Side chapels of, 318 Tomb in, 317 Labels, interior, in, 22 Nave of — Comparison of, with eastern arm, 312 Piers in, 29 INDEX. 4°3 Westminster Abbey — continued. Nave of — continued. Vaulting of, 86, 91 Width of, 239 Piers in, 29, 355 Proportions of, 272, 309 Shafts in, 26-28 Tas-de-charge in, 93 Tomb of Edward the Confessor, 24 note Triforium of, 309 WaU thickness in, 34 Windows of, clerestory, 307-310 Westminster Cathedral, mosaics in, 118 Westmmster HaU, roofing of, 869, 370 Whitby Abbey, 306 Whitewashing of churohes, 119-120 WiUiam of Edington, Bishop, 311 WiUiam of Malmesbury, quoted, 45 WiUiam of Sens, 33 note, 81, 260, 296 WUUam of Wykeham, Bishop, 246, 311, 317 Wimbotsham Church (Norfolk), 368 Winchester Cathedral — Alterations in, by WiUiam of Wyke ham, 246, 317 flailing of transepts of, 245 Chantry chapels in, 92, 155 Choir of, position of, 146 Crypt of, 76 Dates of, 243, 315 Font in, position of, 162 Length of, 223 Nave of — Date of, 315 Length of, 246 Remodelling of, 317 Vaulting of, 91 Piers in, 355 Proportions of, 248 Sculptured figures in, 124 Shafts in, 101 and note ; angle shafts, 75-76 Tas-de-charge in, 93 Transept aisles of, 220, 248, 250 Triforium of, 249 West front of, 315 Windows of, 65 Windows of churches — Arches of, 67-68 ; stated, 60, 62 note Cuspings of, 66-67 Decoration of jambs, etc., 59, 66 Windows of churches — continued. Development of, 55 ; (1200-1350), 261 Early Romanesque in England, 239 German Gothic churches, in, 333 Grouping of, 55-56, 59 Increase in size of (15th century), 364 Mouldings of, 59, 66 MulUons of, 55 note, 59, 65-66 Pierced slabs in, 55, 136, 239 Romanesque, 55 Rose, 68, 278 Tracery of — Bar, 59 English and Frenoh periods of advancement in, eompared, 300 Lincoln Cathedral, at, 309, 314 Origin of, 60 Plate, 59 Screen arrangement of, 314 Transoms of, 65 WaU ribs in relation to 90 Wheel, 60, 68 Windsor— St. George's Chapel, 88, 317, 318 Wing Church, 238 Wittering Church, 238 Wood, see Timber Worcester Cathedral- Buttresses in, 97 Columns in, 28 Date of, 243 East end of, 247 Marble work in, 117 Nave of, 312 Piers in, 29 Presbytery of, shafts in, 28 Retro-choir of, 299 Shafts in, detached, 28, 298 Tower of, 107 Triforium of, 310 Vaulting in, 91 Worms Cathedral — Arcading in, 197-198 Columns in, 31 Exterior of, 194-195 Plan of, 193 Turrets of, 193 Vaulting in, 195-196 Worth Church, 238-240 Wren, Sir Christopher, 90 Wulfstan, Bishop of Worcester, 295 Wyckham Church, 239 404 INDEX. Xanten Church, 325 Yatton Keynell Church, 110 York- Cathedral — • Chapter-house of, 315 Choir of — Date of, 315 Height of, 151 Length of, 153 Position of, 146 Triforium of, 314 note * Vaulting of, 91 Flying buttresses of, 314 note 2 Nave of, date of, 313 South transept of, 306 Towers of, 104, 107 Vaulting of— Material of vault, 314 note 2 York — continued. Cathedral — continued. Vaulting of — continued. Vaulting shafts, 101 WaU ribs, 90 and note, 95 Windows of — East, 65 note ', 314 Glass in, 132-133 North transept — " Five Sisters " window, 58, 132, 306 South transept, 60 West, 62 Cathedral(old^destroyed),dateof,243 Churches in, number of, 358 Ypres — Cathedral, 335 Cloth HaU, 335-336 Zurich Cathedral, 194, 197 THE END PRINTED BY WILLIAM CLOWES AND SONS, LIMITED, LONDON AND EECCLES.