vl THE LIBRARY OF THE UNIVERSITY OF CALIFORNIA LOS ANGELES GLASS MOSAIC AT THE BACK OF A NICHE IN THE IMPEimAL A. Twisted glass rod (fom border. Full size. From Arcliacologia LX/fl. ^^^N^ PAUSILYPON THE IMPERIAL VILLA NEAR NAPLES WITH A DESCRIPTION OF THE SUBMERGED FORESHORE AND WITH OBSERVATIONS ON THE TOMB OF VIRGIL AND ON OTHER ROMAN ANTIQUITIES ON POSILIPO BY R. T. GUNTHER, M.A. FELLOW OF THE LINNEAN AND ROVAL GEOGRAPHICAL SOCIETIES AND OF MAGDALEN COLLEGE, OXFORD JVilh a Map, Two Coloured Plates, and iq6 other Illustrations OXFORD PRINTED BY HORACE HART FOR THE AUTHOR AT THE UNIVERSITY PRESS 191 3 Library TO HIS MOST GRACIOUS MAJESTY KING VICTOR EMMANUEL III THIS DESCRIPTION OF A COUNTRY VILLA FORMERLY OWNED BY HIS PREDECESSORS IN ITALY THE EMPERORS OF ROME IS BY HIS SPECIAL PERMISSION MOST RESPECTFULLY AND HUMBLY DEDICATED 727155 PREFACE The primary object of this work is to provide students of Archaeology with a comprehensive description, as complete as possible, of the extant remains of the Imperial Villa on Posilipo. With this aim the site was surveyed, an inventory of the antiquities taken and the results of re- searches, carried on in vacations between the years 1893 ^"^ 1907, com- bined with an account of what little may be gleaned from ancient writers and with the records of previous investigators. To all this I have added a resu7ncoimy discoveries of the nearer submarine antiquities, previously described in Earth-movemeiifs in the Bay of Naples, both because they formed an integral part of the Imperial Villa and because that book, now nearly out of print, is likely to become rare. Those studies are still the only ones in the domain of submarine archaeology that have been pub- lished, but the recent rumour of the finding of a submerged Greek town on Pharos Bank off the Island of Lemnos in the /Egean seems to indicate that there may be other and wider areas for research. No description of the Pausilypon has yet appeared in English, nor indeed are the Italian accounts much more than meagre reports on the early excavations of di Pietro. It has been our good fortune to find there a classical site of first-rate importance almost untouched by modern scientific archaeologists. To many it is less attractive because the port- able antiquities have been pilfered : the local scholars and antiquaries are so fully taken up with the amount of work still to be done in Pompeii and on other Campanian sites, that they cannot e.\tend their operations ; yet in point of importance the site of Pausilypon is second to none in all that classic region, and is worthy of the closest and most scientific study and investigation. Some apology may be thought necessary for the amount of con- structional detail and for the numerous measurements with which the text bristles, but in a first investigation it is impossible to judge what detail may prove to be important and what may not ; and in this case much will assuredly be destroyed before the next survey is ever made. My chief aim throughout has been truth and accurac^^ and many of the measurements have been verified on subsequent visits, but I am only too well aware that many errors must have remained undetected, and now I am no longer in a position to check any of the statements. My opportunities for the prosecution of these studies were very VI PREFACE exceptional for an Englishman in Ital}-. The most important of the sites were in the possession of friends and fellow countr^'men. I had the advantage of being able to make my home on the spot, to have the use of boats of good construction, to secure the read}' advice of experts, and to preser\'e amicable relations with the Neapolitan fishermen, a jealous class of men who might otherwise have viewed our proceedings with a livel}' suspicion. I gladly take this opportunit}- of pa3-ing a tribute of gratitude to all those who afforded these facilities, particularly to my father-in-law the late Mr. Neville-Rolfe, H.B.M. Consul-General for South Ital}', and to the Earl of Roseber}' in whose beautiful villa on Posilipo we lived, and whose Neapolitan librar}', put at my disposal, was of the greatest help ; but, when looking out over the sunken foreshore of the Villa, what wouldn't we have given for five minutes' chat with one of Lord Rosebery's Struldbrugs ' whose undimmed meniorj' might bridge the abj-ss of time and reconstruct for us the scene as Virgil might have seen it! I must also express my obligations to my kind friend the late Mr. Nelson Foley and to Signor Acampora, the proprietors of the Gaiola Islands and of the buildings of the Imperial Villa on the mainland, to Master Innes Foley for showing me several interesting fragments of terra-cotta or marble that his quick eye had perceived in the heaps of debris and which he had hoarded under the appropriate name of ' keeping-stuff', and to nij- wife for wading through the proof-sheets. The plans, maps, photographs and drawings have mostly been made by mj-self, but those which have been borrowed from others are acknow- ledged in the text. Thanks are due to the Council of the Society' ot Antiquaries for permitting me to borrow the illustrations to m}' previous papers published in Arcliacologia in 1903 and 1912 respective!}-. The expenses of part of the investigation, including that of the Baths, were met by a grant from the Craven Fund of Oxford Universit}' in 1907, while the President and Fellows of Magdalen College very generous!}' contributed towards the cost of reproduction of the illustrations : indeed without their assistance this work would not have been completed. Nor while acknowledging various sources of assistance must those friends be forgotten who have promised to subscribe for copies of this volume, without their help I should fare but sorrowfull3\ Magdalen College, Oxford. June 23, 1913. ' ' The Struldbrug. A Rectorial Address delivered to the Students of St. Andrews, September 14, 1911, on the occasion of the five hundredth anniversarj' of the founda- tion of the University' by Lord Roseberj-. Privately printed 1911. CONTENTS PAGE Introductory i History of the Site .......... 6 Routes from Rome 13 The Grotta di Sejano 21 The Vallone di GiMOL.^ ......... 27 The Theatre 29 The Odeon ............ 40 The House by the Theatre 48 The ' Nymphaeum ' .......... 51 The Temple 56 The Ruins between the Odeon and the Belvedere . . . 6r The House of Pollio 63 The Tunnel on the Valley Road 77 The Eastern Substructions and Terraces 82 The Gaiola Beach and Lower Baths 91 The Upper Baths or Baths of Hadrian 99 Buildings to the West of the Upper Baths 119 The 'Belvedere' ........... 123 The Water Supply .......... 126 The Vineyard Site . . ........ 135 Graeco-Roman Building 140 SCOGLIO DI ViRGILIO I4I The Submerged Regions of Posilipo 145 The Trentaremi Region 147 The Gaiola Region 149 Scuola di Virgilio 155 The Harbour of Pausilypon 163 The Marechiano Region . . • ^73 The Rosebery Region 187 The Burial-Places of Posilipo 197 The Tomb of Virgil 201 Conclusion 205 viii CONTENTS APPENDIX PAGE Inscriptions 211 Architectlre 219 Columns and Bases 219 Capitals 225 Mouldings ............ 229 Architraves ........... 233 Building Construction ......... 239 Bricks and Tiles 241 Pavements 246 Marbles ............ 249 Mural Decorations 252 Mural Paintings . . . . . . . . . . 253 Sculpture 257 Stone Utensils 273 Terra-cotta 277 Glass 287 Uranium Glass 290 Objects of Met.vl .\nd Bone 291 Coi.Ns 292 Index 293 LIST OF ILLUSTRATIONS ILLUSTRATIONS IN COLOUR Roman Glass Mosaic Frontispkcr Wall Painting in the House of PoUio Facing p. 6g Details of Borders of Mural Paintings Page -jo Ceiling Decoration - 74 Mural Painting ,, 255 LIST OF PHOTOGRAPHS AND DRAWINGS PAGE 2 Fig. I. The Western Cliffs of Posilipo 3. The Ridge of Posilipo and Island of Nisida from the West ... 15 4. Grotta di Posilipo i-. 5. Grotta di Sejano, Western Entrance 23 7. ,, „ Eastern Entrance 25 9. View from above the Odeon, looking North 30 10. i he Theatre 31 12. „ Steps leading up from the Precinctio 35 13. „ Eastern Side 36 14. „ Mural Decorations 38 15. The Odeon .^o 16. „ Imperial Box and Passage 41 17. ,, Western Stairway 42 19. „ Restoration of Stage 44 20. A Corner of the Peristyle in a Pompeian House 45 21. The Odeon. Walls 46 24. Nymphaeum. Hemicycle 52 26. ,. ,, Eastern Side 54 27. Temple 56 29. „ South-west Corner 58 30. „ Substructions 59 31. Ruins on the Cliff' over Trentaremi Bay 60 32. Gaiola Valley, Western Side 62 36. House of Pollio. Room F 66 37- >> Room A 67 38. ,, Painted Borders 70 39- » Fresco 72 42. ,, Ceiling Decoration 74 43. Tunnel, Upper End 76 44. ,, South End 77 50. Lower Baths 90 53. „ Suspensura 95 55. The Upper Baths. Lower Store}' 98 56. ,, From the West 98 59. ,, Corridor 104 62. „ The Old Caldarium no 63. Frigidarium of Baths at Pompeii in 66. Upper Baths. Heated Bath 114 X LIST OF ILLUSTRATIONS PAGE Fig. 67. Upper Baths. Heated Bath 114 68. „ Furnace Room 115 70. Roman Baths 116 71. Upper Baths. Stokehole of Furnace 117 75. The ' Belvedere • 122 76. Hill-side below the ' Belvedere' 124 78. Lead Water-Pipe 129 79. Vineyard Site. North Side 134 81. ,, South Side -138 85. Scoglio di Virgilio. Stucco Columns 143 86. ,, and Grotta dei Tuoni 144 87. Quarry Caves in Trentaremi Bay 146 88. Tufa Quarries „ „ 146 89. Gaiola Channel 148 90, 91- " '50 92. Statue of S. Francesco 152 93. „ a Roman Seaside Divinity 152 94. Building No. XXIX, below Mr. Foley's House 153 96. Scuola di Virgilio in 1768 156 97- .. " 1894 156 98. .. ,. 1907 157 99. The Gaiola Island 158 100. „ Eastern Caves 159 101. Isolotto di Gaiola 160 102. Shore of Pausih-pon Harbour 164 104. A Roman Harbour 166 105. „ „ 167 106. Pausilypon Harbour 169 107. House with the Three Buttresses 170 109. Casa degli Spiriti 1 73 110. „ „ 174 112. Marechiano Harbour 176 113. The Old Baths at Marechiano 176 115. Site XVI 179 116. The Temple of Fortune 183 117. Artificial Caves at Villa Rosebery 186 119. Wall Painting of Seaside Constructions 190 120. „ „ „ 191 121. „ „ „ 192 122. The Cavalcanti Tunnel for the Coast Road . . . . . 123. Catacombs near Ranipc del Coroglio 124. Tomb with Amphoras .......... 197 125. Interior of Catacombs 198 126. A Sepulchral Chamber 198 127. Sepulchral Niches 199 128. The ' Tomb of Virgil ' 200 129. ,, „ North Wall 200 130. Inscriptions 212 131- .. 215 132. Brick Stamps 216 i» ,. „ 218 138. Carved Marble Fragments 225 139. Corinthian Capitals 226 140- •> '• 227 142. Pilaster Capitals 228 14.3- .. '. 229 148. Ornamented Mouldings 236 150. Panel Ornament 238 151. Tiles, Suspensura brick, &c 243 152. Leg-tiles 244 153. Mosaic Pavements 246 155. Plaster Reliefs 252 156-7. Mural Paintings 254-5 LIST OF ILLUSTRATIONS xi Sculpture PAGE Fig. 158-9. Bearded Dionysos 258-9 160. Foot and Ann 260 161. Fragments 261 162. Head of Aphrodite (?) 262 163. Draped Figure 263 164. The Nereid 265 165. Candelabrum Base 266 166. Table Supports 267 167. Sarcophagus 267 168. Pedestal 268 169. Thyrsus-Head 268 170. Ring and Knotted Rope 269 171. Woman and Steer 269 172. Battle Frieze 270 173. Sacrificial Relief 271 174-5. Feet of Carnivore and Bull 272 176-80. Stone Utensils 274 178. Marble Patera 275 181. Pedestals 275 182. Terra-cotta Statuettes and Mural Reliefs 276 183. Mural Relief 278 184. Antefix 279 185-7. Antefixal Tiles 280-2 188-9. Miscellaneous Pottery 284-5 190. Lamp 286 191. Glass Vessel by Ennion 287 192-3. Glass Vessels 288 194. Bronze Hinge and Iron Hook 291 195. Bone Veneer 292 LIST OF MAPS AND PLANS 2. Campanian Roads 13 6. Grotta di Sejano 24 8. A Water Reservoir 28 II. Theatre, Viridarium and Odcon 32-3 18. Odeon, Western Stairway 42 22. Chambers beneath the House of Signor Acampora .... 49 23. Nymphaeum 51 25. „ Vertical Section 53 28. Temple ^57 33. House of Pollio 64 34- •. 65 35- .. 65 40. „ North Wall 73 41. ,, South Wall 73 45. Tunnel and Vaults 79 46. Walls near Mosaic Niche 83 47. Eastern Substructions 84 48- •- I, 85 49. Mural Frescoes near Mosaic Niche 87 51. Walls on Gaiola Beach 92 52. Lower Baths 94 54. Thermae of Caracalla, method of heating . . ... 97 57. Upper Baths. Lower Storey loi 58. Walls to East of Upper Baths 103 60. Upper Baths. Middle Storey 106 61. „ Steps 108 64. Heated Basin of Upper Baths 112 65- I) ,. 112 xii LIST OF ILLUSTRATIONS PACE Fig. 69. Restoration of Buildin" over the Furnace Room 116 72. Building to North of Upper Baths 118 73. Chamber X to West of Upper Baths 119 74. Room with Four Niches 120 77. Belvedere and Water Reser\-oir 127 80. Vineyard Site. Western Side 137 82. „ South Side 139 83. „ West Side .... .... 139 84. Scogho di VirgiHo 142 95. Building No. XXIX 154 103. Pausilypon and Marechiano Harbours 165 108. Water Reser\-oir near Casa Marotta 172 111. Casa degli Spiriti 175 114. Baths at Marechiano 178 118. The Roseberj' Region 197 128. The 'Tomb of Virgil' 200 The Gaiola and Marechiano Regions in 1775 208 134. Bases of Columns 219 135. Fluted Columns 221 136. Stucco Column on the Scoglio di Virgilio 223 137. Stucco Columns, &c 224 141. Fluted Pilaster Shafts 227 144. Mouldings 230 145. Cornices, Mouldings and Bracket 232 146. Architraves 234 147. Ornamented Mouldings 235 149. Plaster Cornices and Mouldings 237 154. Marble Pavements 247 Sketch-Map of the Principal Buildings of the Imperial Villa and of the Gaiola Region At End INTRODUCTORY The sweep of the coast of the Bay of Naples between the promontories of Misenum and Minerva is interrupted by but one important headland, which still retains the ancient name of Pausilj'pon or Posilipo. Without going so far as to accept a Greek name as conclusive evidence of the site of a Greek settlement, it may fairly be urged that the position of this headland, situate as it is so far to the windward of the head of the ba}-, is a very favourable one geographically, and would certainly not have been overlooked by the earl}' Greek colonists when they were selecting sites for their townships. There are reasons for the view that the early colony of Parthenope or Palaepolis, the predecessor of Naples, maj- have been established here. Tradition takes us further back still, to cave-dwelling Cumaei, remembered for their tunnellings ; but we have no definite record of antiquities on Posilipo belonging to that hazy past. The principal Roman buildings date from imperial times, and are connected by history and tradition with Vedius Pollio, Virgil, Augustus, and also with Hadrian. In recent 3'ears the fame of the antiquities of Posihpo has been obscured by the discoveries of the more perfectly preserved treasures of Pompeii and Herculaneum, yet it should be borne in mind that while the name of Pompeii is known to us from Roman literature only as a name in a bald list of Campanian villages, that of Pausilypon is associated with the greatest of Roman emperors and poets. In our survey of the principal antiquities on Posilipo we shall include the buildings of the adjoining Regions of Gaiola and Marechiano, now parti}' beneath the sea, the submerged buildings of the Rosebery Region, the roadways and tunnels of Sejanus and Posilipo, the tombs, the stone quarries, and the harbour and its buildings. B 2 INTRODUCTORY The movements of the land which brought man}^ of these buildings under water appear to have taken place between the fifth and the twelfth centuries. They affected a wide area, inckiding, as we have been able to show, the whole of the coast of Campania and Lucania, and, in the Bay of Naples at any rate, have involved a change of level h'lG. 1. The Western Cliffs of Posilii'o. of 40 feet in a downward direction, followed by one of 20 feet in an upward direction. The Imperial Villa of Augustus, the Pausilj'pum, the principal subject of this memoir, is situated on a spot the charms of which are hardly surpassed anywhere in this localit}', though its beauty is so famed that it has passed into proverb. It was placed at tiic south-west extremity of Posilipo, the long green promontory wliich seems, when viewed from the city of Naples, INTRODUCTORY 3 to mark the end of the baj', so completely does it shut out the promontories and islands beyond. The hill-ridge of Posilipo constitutes a natural boundary between the town and that region of burnt rocks whose soils testify to the battle-fields of the giants of subterranean fire and water, the Phlegraean Fields. Along the shore of the eastern or Naples side of Posilipo the sea ripples upon the grey sands of tinj' coves and at the foot of the intervening headlands of tufa rock which slope gently, verdure clad ; and the same gentle features continue where the coastline, bending at the point or 'capo', trends south-west for a mile; then at the Gaiola there is an abrupt change, and the coast, running due west, is exposed to the full onslaught of the scirocco : it meets the enemy with serried front, falling deep and precipitous as a buttressed wall into the sea. Here, where the character of the coast changes, the buildings of the Villa stood, sloping down to the sea on the one side, dominating the cliff on the other. A cluster of ruins crowns a prominent rock still associated with the name of Virgil, the ' Scoglio di Virgilio'. His rock overhangs the islets of the 'Gaiola' { — caveola), which in Roman times were one with the mainland and part of a low promontory extending for a quarter of a mile out into the sea. These, together with the wider foreshore, were all covered with numerous buildings. The space occupied is not large, hardly more than 16 acres, yet few ancient sites have more natural beauties than the Pausilypon ; on one side of the hill the giddy precipices fall sheer into the rippling waters of the cliff-bound Bay of Trentaremi ; on the other, is the luxuriant fertility of the smiling valley of Gaiola. Here, untilled ridges clad in wild copse oak ; there, neat rows of carefully tended vines, and beyond all, the more distant views over the blue sea : strung out in an unrivalled sweep are Vesuvius and the hills of Sorrento, Capri, Nisida, Misenum, and Ischia. Sometimes in clear weather against the sunset sky maj' be seen the rock isle of Ponza. It is by no means easy to reconstruct with accuracy a picture of the landscape as the Romans saw it, or of the complex of buildings with which they crowned the heights. On the western side the buildings have shared the fate of the cliffs on which they stood, have crumbled with them and slid with landslips into the sea : a few remaining ruins are in constant peril on walls of rock undermined b}' quarry-caves of mysterious interest, in which the sea ebbs and flows with hollow gurgles far into the heart of the hill. Inland to the east the ancient valle3'-bed has been filled up with B 2 4 INTRODUCTORY soil washed down from higher grounds and is now cultivated. On level sites the fallen debris of buildings ma}' be found lying by their foundations ; of those which have fallen into the sea no trace remains. But if we miss in these ruins the precision of outline to be found in Pompeii, we have on the other hand natural beauties with which Pompeii cannot compare. The scenerj- is lovel}', everj- ten yards gives a new landscape, rocks mingle with foliage on the hill slopes, a charming chaos of crags and trees ; crumbling walls half draped with iv}*, capped with genista and lentiscus, and linked with long, twining wreaths of br3'on3' ; in spring orchids and anemones are bright like jewels among the grasses : the vegetation is luxuriant and lovely at all seasons. The tufa cliffs though high and barren have a beauty of their own in the variety of shading that time has laid on their surface, in their rugged height, and in the tints of water of varied depths at their feet. It is remarkable that a site with such varied beauties and interests should exist so near a large town and jet have remained completely out of the world. It is only within the last few years that new houses have sprung up, and that older ones have been again inhabited. A vivid picture of the former desolation of the place is given by the Duke of Buckingham, who visited Posilipo in April, 1828, ' A lonely hermit is now the onl}' inhabitant of the site of the luxurious Roman's proud domain. Where millions were spent in enjoyment, a begging friar is the onlj' specimen of mortality ; and on the spot where the festive shouts and the classic songs of thousands were raised in honour of their princel}' entertainer, a mendicant asks for alms to support a wretched existence. On the shelves of the lonely rocks on which stands the low-browed hermitage, with its simple cross and figures of Franciscan friars cut in stone before its door, we saw the hermit watching the passing vessel, and giving his solitary blessing to the fishermen as their barks glided by his abode.' Since that date the hermit has disappeared, his hermitage has been turned into a cottage : ragged children play where he fasted and prated, but a wooden effig}' of St. Francis survives, and it will perhaps be long before modern ideas sweep this descendant of a Roman sea-god from his crag. Yet to this very isolation is due that it is possible to make out anything of the ancient buildings at all. Ground plans can for the most part be revealed by excavation, but the superficial treasures have been long removed not onl}' by the INTRODUCTORY 5 systematic search of antiquity-hunters, but also by the incessant pilferings of the tenders of the woods and vine^'ards. The unlettered ploughboy wins The casual treasure from the furrowed soil. Still the ground is always worth looking over and especially after heavy rain, when films of moisture polish the fragments of precious marbles and once again revive their decorative colours. HISTORY OF THE SITE ' But,' some objector will ask, ' what evidence is there that this site is really that of the Augustan Pausilypum?' Frankl}', we have no proof: there are no inscriptions or boundarjf stones, but we have tradition. There is no other site in the vicinity with buildings of like importance, and, what is more, we were fortunate enough to find evidence that a part of the property was in the imperial possession at the time of Hadrian. The presumption is that he succeeded to it b}' inheritance. But we must not anticipate history by the con- clusions of archaeology. The Villa is known to us from a passage in Pliny,' who praises it as 'a faire house of retreat and pleasure called Pausil3-pum, in Campania not far from Naples ' : but such are the unexpected turns in a reputation, that it has been remembered not for the supreme beaut}' of its position, nor for the luxurious perfection of its architec- ture, but as having- been the scene of an inhuman displa}' of callousness by the first owner, one Vedius PoUio, who with Mopsus the Lydian and with the grandfather of Agamemnon has often been held up as a monster of cruelty for having caused his slaves to be rent and torn b}- fish. Vedius Pollio, a Roman freedman, was a courtier, and a great favourite of Augustus. ' He devised experiments of cruelty by means of this creature {nitiracim) ; for he caused certain slaves condemned to die, to be put into the stewes where these muraenes were kept, to be eaten and devoured bj- them : not that there were not wilde beasts ynow upon the land for this feat, but because he tooke pleasure to behold a man torne and pluckt in pieces all at once : which pleasant sight he could not see by any other beast upon the land.' - On one occasion when Augustus was a guest at Pausilj'pum a crystallJHum or glass vessel was broken by one of the slaves. The wretch would assuredly have paid the penalty had he not begged his ' Xat. Hist. ix. 53. • nr. Holland's translation of Pliny ix. 23. HISTORY OF THE SITE 7 life of the Emperor, who not only granted the petition, but caused all his host's glass vessels to be smashed and, instead of the culprit, to be thrown into the fish-tanks, where the fragments may still be lying for the joy of future archaeologists. This stor}', which has perhaps as much foundation in fact as any other item of cliruiiique scandaleusc, was current in Roman Society, and Seneca in a brochure, De Clcmentia, intended for the perusal of Nero, asks, though not foreseeing the grim humour of addressing the question to that particular prince, ' Who could fail to execrate such a monster as this Vedius Pollio, beyond even the aversion felt by his ill-starred slaves, or to deem a wretch who could fatten fish upon living men, himself worthy of ten thousand deaths?' But though no respecter of human life, Pollio would appear to have spared no trouble or expense for all that aff"ected the welfare of his fish, and it is not likely that he would be behindhand in the knowledge of the art of pisciculture which was thriving exceedingly in his day. Like many others, the fish-tanks of Pollio were probably divided by partitions so that such fish as lived in a state of enmity might be kept apart. Notorious among these were the mullet and the sea-pike which ' hate one another and be ever at deadly war ', and the conger and the lamprey ' which gnaw off one anothers taile '. Special arrange- ments would have been made for the renewal, aeration, and cooling or warming of the water, and for the provision of shady grots in which the fish might lie when, in dog days, the direct rays of the sun beat strongly. Lucullus is said to have been even more extravagant with his pisciculture at the neighbouring villa, the Neapolitanum, nearer Naples. He made a cutting through a mountain to let in an arm of the sea into his fish-pools, the doing whereof cost him more money than the house itself which he there had built. After his death his fish sold for three million sesterces, or about .^8,000. We have no record that Pollio was as attached to any of his fish as Hortensius at Bauli, who ' could not hold but weep for love ' of a lamprey which died in his pool, or that he decked his fish with golden ear-rings as did Antonia the wife of Drusus. These fish lovers, says Cicero, ' deemed no moment of their lives more happy than when the creatures first came to eat out of their hands.' But it is certain that in the Pausilypon aquaria, fish throve remarkably well, and some reached an age not often surpassed by fish living in captivity ; one of the fish dying sixty j-ears after it had been put in by Vedius Pollio, 8 HISTORY OF THE SITE and the fish-pools had become the property of Caesar : two more of the same age and same kind were left, which lived still longer.* PoUio made a will by which at his death in 15 B.C. the Pausilypon estate passed to Augustus, who also appears to have acquired the Neapolitanum of LucuUus. And so during the first and second centuries most of the finest and largest villas round about Puteoli, Baiac, and Misenum passed into the possession of the Crown either by inheritance or bj' confiscation. Augustus himself appears to have resided for lengthy periods in or near Naples, and for part of the time at all events the Pausilypon property' was in the hands of procuratorcs, the name of one of whom M. VLPIVS EVPHRATES is preserved to us in an inscription. That the propert}' w^as still in the imperial possession in the second century a.d. is indicated b}- a leaden water- pipe, which was found in the upper baths, stamped with the name and title of Hadrian. The associations of \'irgil with Pausil3'pum are of a more legendary character. History throws no light upon his visits to the Augustan Villa, and anj' glimmer of recollection of the great poet which might have persisted through the Roman period has paled before the false lights of necromancy which have gathered round Virgil, the great magician of the Dark Ages. It is foreign to our present purpose to discuss the various traditions concerning Virgil and his magical works, but it should be mentioned that many of the stories about him are believed at the present day as implicitly. as ever. Not long ago an old fisherman is reported to have told the following story concerning the ruins commonl}- known as the ' Scuola di Virgilio '. ' Sit on that wall,' said the old man ; ' that is where Virgil used to sit. One often saw him there with his book in his hand, he was a handsome, fresh-looking man ; he knew how with his magic to preserve his youth, these walls were covered with circles and lines. He used to come here with Prince Marcellus and teach him the secrets of the spirit world. Often in the wildest storms, when no fisherman would have dared to go out, they used to put to sea in a boat. No rower was ever afraid when \'irgil was with him ; the fiercer the storm, the better he liked to be there. Often he sat up there on the mountain and looked out towards the gulf. Many of his books he wrote there. No doubt they were prophecies which he wrote, for there was never a storm but he foretold its coming. Then he visited the gardeners and field labourers and gave ' Plinv ix. 53. HISTORY OF THE SITE 9 them good advice/ and taught them when to sow their corn. Often when cloud and storm were coming down from Vesuvius he would turn them back with a powerful spell, and often he would spend whole nights with his face towards the mountain when the lightnings were beginning to flash about its head, perhaps in silent converse with its spirits. There had long been talk of making a road from Naples over Posilipo ; he came to our aid, and in one night his spirits had built the road through the cave. Another time he helped us in a wonderful way ; the gnats had become as great a plague here as they were in Egypt in the days of Moses. So he made a great golden fly, which rose at his command into the air and drove all the gnats away. So too once all the wells and fountains had become infested with leeches ; he made a golden leech, threw it into a well and the plague was stayed.' It is not easy to believe that a people to whom these stories have meant so much, should have forgotten the site of the grave of their hero, especially as at a later period the very presence of his grave was believed to be intimately connected with the welfare of the city. But the consideration of this subject as well as of his tunnelling must be reserved for the appropriate chapters. We were not without hope of finding some graffito upon a w'all which might have further corroborated what is already known con- cerning the extraordinary popularity of Virgil in the neighbourhood of Naples, witness the numerous quotations from the Aencid on Pompeian walls, many of which were undoubtedly due to the fact that the works of Virgil were used as reading-books for schoolboys at an early date.- Almost certain it is that his poems were recited,^ and the drama- tized versions of parts of them were acted either in odeons or in theatres. We know, for instance, that a Virgilian pantomime, Tiinuis, had found such favour with Nero that he vowed he would perform it himself in the event of his escaping from his troubles.* The first owner of the property in more recent times of whom we have any record as having interested himself in the local antiquities was a Greek gentleman, Antonio Paleologo,'' who nourished in the ' Virgil possessed an estate near Nola (Aulus Gcllius). A garden in whicli he grew medicinal plants was said to have been situated on Monte Vergine near Avella— originally known as Mons Vergilianus. ' ' Primus dicitur Latine ex tempore disputasse, primusque Virgilium ct alios poetas novos praelegere coepisse.' Suetonius, De Graiuiu. et Rlutt. 16. ^ ' Auditis in theatro Virgilii versibus,' Tac. Dial, de Oratt. 13.7. ' Suet. vi. 54. ^ Celano 9th day. lo HISTORY OF THE SITE sixteenth century-, and who decorated his Posilipan villa at Marechiano with statues and Greek and Roman inscriptions.' The site must have been much encumbered with brushwood and disjointed masses of masonr}-, but notwithstanding all this, Fabio Giordano knew of the existence here of a theatre or circus, of a nymphaeum, of the ' piscine ', and of the half of a portico clad with marbles and adorned with statuar}', including a statue of Mercury and other divinities.- From Paleologo the property passed bj' purchase to a noble family of Salerno of the name of Maza, who for several generations showed a commendable interest in archaeology. Francesco Maria Maza [circ. 1680) was himself the author of the indifferent inscriptions which he caused to be affixed to the so-called ' Piscine of V. Pollio' and to the ' Temple of Fortune ', and which are still in sittij^ The fate of the Maza collection was dispersal, and the loss to archaeological science was irreparable, for a catalogue raisoimc had never been prepared of them. Several objects of art from Posilipan sites found their waj' into the hands of Spanish collectors, and are still no doubt among the Roman antiquities in Spain. Man}^ fine pieces were taken to Mergellina and lost among the other ornaments of the villa of the Viceroy, the Duke of Medina. Under date Jul}' 5, 1755, Carlos Weber recorded in his diar}- that a large and ver}- beautiful marble capital in the Corinthian style, which had been found at Mar piano (Marechiano), was to be sent to Portici to be put into the palace which Cannavari had planned for Charles HI in 1738. (Ruggiero, 1888.) About 1820 the southern portion of the property was purchased b}' a well-known Neapolitan archaeologist, Cavaliere Guglielmo Bechi, at public auction, and his name was associated with the Villa for more than half a centur}-. During Iiis proprietorship much excavation was accomplished, but again without publication of results. In 1841 more methodical excavations were commenced on the adjoining property to the west of the course of the ancient lane that led down the vallej' from the Grotta di Sejano to the sea. Here, his Excellenc}' Monsignor di Pietro, Archbishop of Berito and Apostolic Nuncio at the court of Naples, being much interested in recovering records and treasures of ancient art, entrusted the examination of the region to an architect, Pietro Bersani, and the results of the work ' Aldo Manuzio, Grutero. " Fabio Giordano's sixteenth-century MSS. preserved in the Naples Library. ^ Fr. Guiscardi, Di tin aittico kmpio a ' Marc C/iiaio', Napoli, 1906. HISTORY OF THE SITE ii fully justified his hopes. The precious marbles lying in the soil must have been so conspicuous that a man would have been blind indeed not to have comprehended their meaning, and the spade having been applied in the winter of 1841, the principal buildings on that part of the Villa were soon brought to clearer light. In a letter to Cavaliere Visconti, dated March 12, 1842, after the work had been going on some five months, Bersani gives a brief resume of their progress. He reports that a fragment of a fluted column of cipolline marble was found at a spot which was strewn with tiles and masses of opus reticulatum. This led them to expect that more might be found here; nor were they disappointed, for they soon dis- covered anew the Theatre, an Odeon, and the remains of a Portico overlooking the sea. The oblong building now called the Temple was also found and the remains of an aqueduct. On January 13, 1842, a statue of a good period was unearthed ; the workmanship was proclaimed excellent, half life-size, Greek in style ; and though headless and without arms it was judged sufficiently attractive to please members of the court, and they were invited to view it. An interesting note upon the site is recorded in the Diary of my brother- Fellow of Magdalen, Nassau W. Senior.^ We went on Monday, Dec. 30, 1850, with the Gladstones to see the excavations made two or three years ago by an architect, M. Bechi, and by the Papal Nuncio, below the promontory of Posilipo, on the site of the great Villa of Augustus ; the villa in which Pollio received Augustus and disgusted him by ordering a slave to be cut up to feed the Murenae; the villa in which Cicero visited Marcus Brutus, who occupied it as the guardian of joung Lucullus. A few days later he was invited to a private view of the statue of the Nereid at the Museum in Naples. At the demise of Cavaliere Guglielmo Bechi the property passed to his daughter, who sold it to a speculator named De Negri, who doubtless inspired by the tradition of the teeming fish-tanks of Pollio hoped once more to enclose a portion of the sea and to make great profit for himself out of the proHfic capability of fish. With this object he endeavoured to form a company, the ' Societa dclla Pescicoltura del Regno d'ltalia nel Mar di Posilipo', the success of which was forecast in a Imall brochure written in 1864 by Dr. Sorito. In spite of the fact that the arguments were backed up by figures showing that ' Journals kept in France and ISaly from 1842 to 18^2, London, 1871. la HISTORY OF THE SITE the profits for the year 1865 would amount to more than 760,000,000 lire, it could have needed no great discernment to perceive that a class who are as jealous of their rights, or as unscrupulous in dealing with those who would interfere with them, as the Neapolitan fishermen, would not be likely to allow their daily bread to be taken without a struggle, and the failure of the Piscicultural Society was the result. The Company went into liquidation and the property was purchased by the Marchese del Tufo, an amateur canoeist, who knew every inch of the coast and was smitten with the idea of an islet home on the Gaiola. During his occupancy a great deal of digging of all sorts went on. The Marchese opened a quarry for pozzolana between the Villa Bechi and the vineyard above the Scoglio di Virgilio, thus clearing away the central part of what had been a broad continuous terrace along the south front of the property in Roman times, and had lasted until 1870. The buildings that stood against the hill-side above the terrace, including the southern part of the baths, were just tumbled down the slope into the sea, the pozzolana upon which they stood being shot on board vessels moored in the straits below. During these operations the statue of a female figure was found above the Grotta dei Tuoni, about the year 1872, but the antiquities found, even if we now had them, would be small compensation for the destruction of the buildings. The next change brought the property into the possession of the neighbouring landowner Signor Acampora, who afterwards parted with it to its late owner Mr. Nelson Foley. The Villa Bechi on the mainland, with a small part of the estate immediately contiguous, passed for a few 3'ears into the possession of Mr. Norman Douglas who, when extending the garden, made several discoveries of ancient marbles which will be dul^' noted later. Our own investigations with the kind permission of the proprietors were made at various times between 1893 and 1907. Once more the property has changed hands and the Onorevole Paratore is the fortunate owner of the estate on the mainland, while Dr. Ernest Praun has acquired the Island of the Gaiola. ROUTES FROM ROME Before the increased intercourse of the towns of the Phlegraean Fields with Rome had made a direct and permanent road a necessity, travellers from Rome to Naples, having followed the Appian Way as far as Capua, would have proceeded b^^ the straight road which ran due south through Atella and entered Naples b}' the Porta Fig. 2. Campanian Roads. Capuana, on the central thoroughfare, the Decumanus major, then probably known by the name of Dioscurias after the temple of the Dioscuri (on the site of the Church of San Paolo), and now the Strada dei Tribunali. But when under the early Empire the commercial importance of Puteoli, the allurements of Baiae, and the military' position of Misenum combined, made a direct highway imperative, the old track leading along the coast south from Sinuessa to Cumae was greatl}' improved. The waj' ran along a dull, flat, sandy fore- shore, in places narrowing to a mere bank of sand between the lagoons r4 ROUTES FROM ROME and the sea. The only engineering difficulties were the securing the permanence of the causewaj^ among drifting sands, and the bridging of the mouths of the rivers. Of these the Volturnus alone would have presented serious trouble. So great was the volume of its yellow waters when swollen, that before they were eflfectively bridged b}' Domitian, the coast route does not appear to have been practicable at all seasons of the j'ear. At the same time, Statius would seem to have given Domitian rather more than his due share of credit in the well-known panegyric on the Via Domitiana, for an inscription ' found at Castel Volturno is proof of the existence of a road of sorts at an earlier period. It is a record of repairs executed by the Duoviri, M. Arrius and M. Sextius. By the Via Domitiana the journey from Rome to Puteoli was shortened by some 13 miles and that to Neapolis by some 4 miles. The actual distances from Sinuessa to Puteoli and Neapolis b}' the Way of Domitian were about 22 and 28 miles respectively, whereas by the consular roads via Capua they were about 35 miles to Puteoli and 32 miles direct to Naples. B3' the shortest route the Pausilypum would have been about 133 miles from Romc.'- Communication between Puteoli and Neapolis in quite early times cannot have been easy owing to the natural barrier formed by the steep western escarpment of the Posilipo ridge, which for 3 miles extends like a rampart from Coroglio to the Camaldoli massif and M • ARRIUS M • F • M • SEXTIUS ■ M • F ■ DVO • VIRI • D • C • S • VIAM ■ FACIVND • ET ■ REFICIVND • COERAV The distances between the stations were thus stated in the Roman itineraries: Tabula Peuiingeriaua. AiUomtte Itinetary. Sinuessa Sinuessa Safe VII Volturno XII Literno XII Literno XXIIII Cumas VI Cumas VI Puteohs III Putcolis III Neapoli V Neapoli X The distance from Sinuessa to Capua was XXVI. The length of the consular road Capua-Puteojis was XXI, and that of Capua-Neapoli XVII. ROUTES FROM ROME 15 effectively bars the way. In Roman times there were only three practicable routes. The oldest, the Via Antiniana, led over the northern end of Posilipo at the only point where the gradient allowed of a carriage road : for the rest of its length the western precipices were so deterrent that it seemed easier to carry two other roads through the hill by tunnels each close on half a mile in length, than to engineer roads up and over the obstacle. The tunnel on the middle of the three roads, known as the Crypta or Grotta di Posilipo, measured 774 yards and the southern tunnel or Grotta di Sejano 844 3-ards in length. Both tunnels are described Fig. 3. The Ridge of Posilipo and Island of Nisida from the West. by classical writers, but partly in consequence ol their not being both mentioned in the same passage and partly no doubt to the fact that the existence of the southernmost had been forgotten until its rediscovery not so very long ago, a good deal of confusion has arisen as to which of the two tunnels is referred to in each case. Fortunately, however, the two tunnels were not alike. The Grotta di Posilipo was originally very dark owing to its lovvness, narrowness, and the absence of any shaft through which light might enter. The Grotta di Sejano, wider and higher, was partly lit and ventilated by lateral shafts through the cliffs at the head of Trentaremi Ba}^ We may therefore take it that Strabo was describing the Grotta di Sejano when he wrote, 'There is a hidden passage through the mountain between Dicaearchia and Neapolis, which is pierced like the one near Cumae, so that the way is open for carriages to pass for several furlongs and light from the outside of the mountain is brought down through shafts from many directions,' and that by the term Fauces, Seneca was referring to the northern tunnel, in which two chariots could not pass one another and which was also so low that, as Pctronius graphically puts it, wayfarers had to duck their heads as they walked through. The comparison instituted by Strabo between the Posilipan tunnel and that at Cumae is unsuited to the Grotta vccchia di Posilipo in its original condition. i6 ROUTES FROM ROME No mention of the Grotta di Posilipo has been found which can be regarded as being earlier than the time of Nero, so that it is impossible to confirm or to refute the theory of the older Neapolitan antiquarians who have attributed it to the Cumaei, those mysterious early inhabitants of this region who in the remote past were noted for their burrowing proclivities. The tunnel is mentioned in the Itinerary of the Geographer of Ravenna, and a rough diagrammatic sketch of its two entrances is attempted in the Tabula Peutingeriana. In mid-tunnel a small chamber hewn in the rock contained a Mithraic shrine, with a dedicatory inscription to the god : OMNI POTENTI DEO MITHRAE APPIUS CLAUDIUS TARRONIUS DEXTER V • C • DICAT. /. -\-. 2481. A Christian chapel has now succeeded to the Mithraeum and often exercises a veiy important if passive function, viz. that of preserving order in what might otherwise be a dangerous locality. Mithra may also have been a good influence on the side of order. In its present state there is room in the tunnel for three carriages to pass abreast without inconvenience, save from the dust which the wheels of the carriages and horses put in motion, and which in addition to the choky nature of all dust, exhales a disagreeable, pungent odour. Images and pictures of saints used to be hung on the sides of the cavern, with small votive lamps burning dimly before them ; but the presence of these symbols of religion did not prevent the loud imprecations of the coachmen, muleteers, and lazzaroni, which sounded lugubriouslj' amid the reverberations produced by the noise of the carriages. The tunnel, as we now know it, is ver^- different to what it was when first constructed. It has been shortened, heightened, widened, levelled, and lit since the Middle Ages, so that it is about twice as high and as wide and more than four times as well illuminated. Judging from its present gloom, it is not surprising that in its original condition it was so dark and dust}'' that Seneca termed it a prison and Petronius complained of not being able to stand upright in it. Indications of its earlier gradients and dimensions are marked on the walls and have been made the subject of a searching study by Paoli. Owing to improvements having been principally effected by the lowering of the roadwa}', the oldest marks on the wall made by the dragging of cart-axles are those which are highest above ROUTES FROM ROME '7 o > b. O ffi S O H o C u o 13 •a c C3 C o o O 6 i8 GROTTA DI POSILIPO the ground. The curvature of the original vault is also indicated in places. The various alterations of which records have been noted on the walls of the tunnel are shown in the elevation, fig. 4. The present roadway a b has been cut through the rock. At about a yard above the ground a groove, cc, like a cart-rut has been ploughed in the southern wall by the axles of carts which have been allowed to rub against the soft rock when passing other vehicles, or as a simple method of putting on the brake when going downhill alia Marina. This groove rises from a height of 8 feet above the present floor of the tunnel, which is nearly level, to 17 feet further in, indicating a gradient of about one in eight. A similar groove d d, starting at 14 feet and rising to 34 feet in a lengtli of 130 feet, may be taken as the indication of an older and higher road level. Finall}', at the level e e is a projecting ridge which has been taken to indicate a still older street-level, 61 feet above the present. The two tombs, one shown at f, bj- the entrance to the tunnel must have been constructed when the roadway' was at this level, for ever since the lowering of the road the doorway to one has been left inaccessible and the other has onl}' been visited by birds and bats. Arranged in chronological order the following stages may be recognized : 1. The original tunnel answering to the description of Seneca and Petronius and reached by an approach of sufficient gradient and elevation to give access to the tombs. Some improvement is said to have been effected in fifteen days b}^ Cocceius. 2. The roadway lowered to d d. 3. „ „ cc. 4. ,, ,, present level. The length of the tunnel has been variously stated b\' different authors. Passing over the much e.\aggerated estimate of fifteen miles of Benjamin of Tudela in the twelfth centurj', we find that some of the older measures are greater than those made nearer our own time. The dimensions, unless otherwise stated, are in English feet. Year. Authorily. 1570- F Giordano. 1607. Capaccio. 1600. Sanfelice. de Magistris. 1692. Celano. 1768. Paoli. Length. Width. Height. 600 piedi. 24 ft. — nearly one mile. — — 34 ft. 2.300 ft. 20-6 ft. 81 ft E. 83 W. 64-22 ; 5 in centre GROTTA DI POSILIPO 19 Year. Aijt/iori/y. Length. Width. Height. 1776. Carletti. 2,283 ft. 19 ft Ends. Middle. 1793- Lorenzo Giustiniani. 2,262 1841. Mendia. Original Dimen- sions. 2,322 10-3— 8 17-8-6 ,, Actual Dimen- — 205-15 80 14 sions. 1862. Murray. 2,242 21-5 69 E. 25 1890. Baedeker. 2,271 32-21 87 E. 50—20 One of the benefits to Naples and especially to the traffic with the Phlegraean villages, due to the Spanish Dominion, was the reconstruc- tion of this tunnel. Alfonso I of Arragon relevelled and widened the roadway in 1442, and caused two circular air-shafts to be cut in an oblique direction through the hill so as to allow light to illumine the middle of the tunnel. Pietro Razzano, however, is quoted as having said that there was no ventilation, that the tunnel was buried in a dense thicket, and that people were afraid to pass through. In 1546, Pietro di Toledo paved the roadway, improved the illumination by two new ventilating shafts and by enlarging the ends, and erected the little chapel to Santa Maria della Grotta in the middle of the tunnel, probably with the intention that it might serve like its predecessor the Mithraic shrine as a check to robbery and other crimes to which the place might be a temptation. By the alterations of this Viceroy the roadway appears to have been levelled to such an extent that the light at one end could be seen from the other as a star,^ and on a few days about the equinox, the sun actually shines through. Further repairs were undertaken by Charles III in 1754, internal supporting arches being inserted where the roof showed signs ot giving way ; the pavement was also renewed. In the Middle Ages the construction of the tunnel was attributed to Virgil, not to the Virgil as he is known to schoolboys, but to a mythical presentment of him, Virgil the Necromancer. The popular legend may be read in Scoppa with several others borrowed from the Cronica di Partcnopc. The story has travelled into the literature of every nation in Europe,- and a reference to it also occurs in Marlowe's Doctor Faushis, Act I, sc. 26: There saw we learned Maro's golden tombe, The way he cut an english mile in length Through a rock of stone, in one night's space. ' Scotto, Itinerario. ' Cf. Thersander, Schaiiplats vieLr iiiigcrciiiiten Meinungen, ii. 308, 554: Jean d'Auton, Cliioniqiies, i, p. 321. C 2 20 GROTTA DI POSILIPO In Naples the story was told that the Grotta di Virgilio, b}' which name the tunnel was commonly known, was once visited b)- King Robert and Petrarch. On being asked what he thought of the popular belief, Petrarch answered jestingly that he had nowhere read that Virgil was a magician.' To this the king confessed that the place shows traces not of magic but of iron quarryman's tools, ' non illic magici, sed ferri vestigia confessus est.' Not much less fabulous is the storj- repeated bj' Giannettasius (Hist. Neap. Dec. i, c. i) and Celano (Notizie, giorno g) that Cocceius completed the whole tunnelling in fifteen days ; it is, however, possible that this period was devoted to altering or widening a pre-existing tunnel. It is not surprising when we consider the great development of Virgilian legends in Naples that this wonderful tunnel, so near his grave, should have been attributed by Neapolitans to their m3-thical benefactor, and it was said to have been rendered inviolable by its proximity to his tomb. ' Petrarch (Ilin. Syr., i, p. 560 (ed. Basil, 1581). We made our first acquaintance with this story in Benecke's English translation of Comparetti's learned work on \'irgilian Legends, but the quotation given by that authority in a footnote is incorrect. For ' nusquam meniini me legisse marmorarium fuisse Vergilium', read 'nusquam me legisse magicarium fuisse Virgilium '. GROTTA DI SEJANO The second tunnel passes through the hill but a few j'ards distant from the vertical face of the j-ellow cliffs in which Posilipo abruptly ends. Although called the Grotta di Sillano by Giordano, it is now known as the Grotta di Sejano ; but how it came to be associated with the minister of Tiberius, unless by a confusion with Sillano, is a mystery. There are good grounds for believing it to be somewhat earlier, coeval in fact with the tunnel near Avernus, constructed by order of Agrippa to connect the Lake with Cumae by a direct road (37 B.C.) and designed by the engineer M. Cocceius Nerva (Straboj. But of course, Sejanus {circ. a. d. 20) may have improved or repaired the tunnel, and as we shall point out later, the sides of the tunnel show work of different periods. That the Grotta di Sejano was used for public traffic until a comparatively late period is shown by two inscriptions that have been found within it. The first on a milestone refers to some part of the long reign of Constantius Pius (335-361) : the other has been attributed, though with some uncertainty, to the reign of Honorius (395-413)- VII D • N FL ■ VAL CONSTANTINO PIO • FEL INVICTO • AVC DIVI • CONSTANTI Pll • FILIO- C.I.L. 6930. The figure VII may relate to the distance from the gate of Cumae, now called the Arco Felice, to the western end of the tunnel. From the fifth centur}', until its reopening in the nineteenth, we have no information concerning the tunnel excepting a few scattered references to it as a large cave ; all knowledge of its former use as a roadway having been lost : but in 1840 the making of the new road from Coroglio to Bagnoli, and the interest in the adjacent antiquities, drew attention to it. It was found that one end of the tunnel 22 GROTTA DI SEJANO had become closed b}- earth-falls washed down bj' water from the hill above, and the entrance was hidden bj' vegetation. Influenced by a report of the engineer Mendia, King Ferdinand II ordered the clearing of the tunnel in Ma}-, 1840. The workmen were at first much impeded in their work bj' the heat and absence of ventilation, and the falling in of portions of the vault was a constant danger. It was only after the clearing of the ancient air-shafts, opening into Trentaremi Baj-, that the air was rendered pure enough to support the burning of the candles. For a considerable space the sides and vault of the tunnel had to be supported by new masonr}'. The passage was found to be quite blocked b}- earth-falls at nine points, and one occurred during the progress of the work. Notwithstanding all difficulties the tunnel had almost been cleared by Jul}-, 1841.' We have unfortunately no indication of the nature of the original approach to the western end of the tunnel. The road-level, now about 22 feet above sea-level, would, allowing for the subsidence of the land, have been at least 40 feet above the sea in Roman times. The new road from the plain of Bagnoli at the Rampe di Coroglio, has been engineered up to the level of the entrance to the tunnel by a double zig-zag traverse, when any indications of the course taken by the Roman roadway were probably buried or removed. As we now see it, the west end of the tunnel is greatly obstructed by massive arches of tufa which have been recently put in to hold up the vaulting, and which by reducing the width of the opening by one half have rendered the interior much darker than it formerly was. It is a great pity that some plan of restoration, not involving the introduction of these unsightly piers and arches, was not adopted, and even now it would not be impossible so to strengthen the original work in other ways so as to permit of their removal. As it is, it is not easy to appreciate the original effect of the fine entrance, quite 45 feet high, which would be more impressive were it not for the cramped nature of the approach from the road and the partial concealment by a commonplace entrance-lodge. On examining the walls round the little court before the tunnel it will be seen that the great entrance arch is flanked by strong retaining walls built to hold back the soil of the hill-side. These are faced with carefully shaped stones which are rather smaller than those used for the internal work, and are believed to be of diff^erent date. Here, too, it will be noticed that the original vaulting is joined to the later ' Mendia's final report in August, 1842. was reprinted by Lancellotti, see Promenade, 1842. GROTTA DI SEJANO 23 work with a lower arch to which painted plaster was still adhering in 1842. On the side walls the masonry is in two different styles, the one coarse reticulate, the other in smaller and more carefully trimmed blocks. The internal dimensions and nature of the work are by no means uniform from end to end. The eastern or Gaiola end quarried out of the solid tufa needed no further support, but the greater length being excavated in the less compact pozzolana required an internal masonry lining which appears to have originally been executed throughout in opus incertum and tufa concrete, though repairs have Fig. 5. Grotta di Sejano. Western Entrance. been carried out in other styles. The roadway rises from west to east. According to the report of the Ro3-al Commission of Sept. 24, 1842, the length of the tunnel is given as 2,526 feet (Baedeker has ggo yards), and is therefore 208 feet longer than the Grotta di Posilipo. The width varies from 21 feet to 13 feet and the height from 28 feet to 14 feet, but the Bagnoli entrance is much higher, being 48 feet high by 23 feet wide. This western entrance facing Puteoli is formed by a very high and strong semicircular arch of brick resting on supports 4I feet thick. Inside, the walls are of opus reticulatum for some distance. In the north wall is a niche, 13I x 18 feet long, in which a fragment of an inscription was found referring to repairs carried out in the tunnel, and believed to contain the name of the 24 GROTTA DI SEJANO Emperor Honorius, which would indicate a date early in the fifth centurj'. lOITORIO lAVSILIP GLECTAA RAIVSVCCAMP PUBLICO REDDIT However, we prefer the reading given in the Appendix, p. 211. For a length of some 260 feet the thickness of the crown of the semicircular vault is 2^ feet and the side walls are a little over i| feet thick. The vaultinL,*^ then becomes thinner, i^ feet being its average thickness for the greater part of the tunnel. After the reticulatum follows a long stretch of opus inccrtum. Further on, where the rock |L» .'??^5^'- 'V* ^^^-^^^^^ '^ . /•u'r,* !"» Fig. 6. Grotta di Sejano. Plan. was less compact or where the original masonry has failed, the lining had to be reinforced by stronger work, and so we meet with a lining of isodomic tufa masonry of narrow blocks of tufa traversed by a horizontal band composed of three courses of red brick. A portion of the vault is supported by arches carcfullj' constructed in flat red brick, which has been supposed to have been part of the repairs referred to in the ' Honorian ' inscription. The varjung height of these brick arches would seem to indicate that a good deal of material had fallen from the roof of the tunnel before the}' were built. Some human skeletons were found near them. A length of the original opus incertum (largish blocks) follows, which gives place further on to opus reticulatum of inferior execution. In the succeeding length of opus incertum opens the middle ventilating shaft, and this is succeeded bj^ the last or Gaiola section of the tunnel, which has been hollowed out of the solid tufa for a distance of 488 GROTTA DI SEJANO 25 feet (180 feet, Berard), and is about 16 feet wide b}' 28 feet. Of the original air-shafts only three remain. The others may have been destroyed by falls of earth, and have been concealed by the later work of restoration. The ventilation is effective, but the illumination is unsatisfactory. Three horizontal galleries open in the face of the cliff over Trentaremi Bay. The one nearest the end of it is excavated in the tufa, the other two in the pozzolana. They are respectively situated at 216 feet, 562 feet, and at 965 feet from the Gaiola entrance to the tunnel. Their dimensions are shown in the following table. Distance from entrance. Lenglli. Width. Height. 216 ft. 130 ft. 5J Ct. 9 ft. 562 93 5i 8i 965 645 4 7 Fig. 7. Grotta di Sejano. Eastern Entrance. The Cave of Polyphemus of Berard. The eastern entrance is half blocked by a modern wall of tufa with a doorway : it is about 30 feet high by 20 feet broad. As has been said, the last section hewn through the solid rock is somewhat higher than the preceding portion, which, traversing looser strata, required an interior lining. This simple material explanation of the different character of the two sections of the tunnel does not appear to have been given its due importance by Berard, who, trusting to the proximity of a few of the common umbrella-pines, identifies this Roman quarry with the Cave of Polyphemus. There may of course 26 GROTTA DI SEJANO have been a cave here in very early times, a cave used b}' primiti\e shepherds and their floctcs as a place of refuge, and to keep their wine and milk cool. The existence of a cave maj' even have suggested the thought of piercing the mountain ; but much clearer evidence will be necessary before we can accept it as the scene of the Homeric legend. In any attempt to appreciate the cost of these reallj- great undertakings, it should be borne in mind that it was the practice in Roman times as at the present, to quarry tufa and pozzolana for building, bj' excavating large galleries underground without disturbing the surface soil ; and to drive one of these quarr}^ galleries straight on through the mountain could have involved but little extra expense. It is probable that the Grotta di Sejano was one of the quarries whence was drawn much of the material used in the extensive constructions above and below water to be described later: the material excavated was certainly not thrown out to form large spoil- heaps like those near the ends of modern raihvaj' tunnels. The labour must have been immense, for it has been estimated that the quantity of material excavated from the Grotta di Sejano alone was not less than 100,000 cubic yards. In addition we have the expense of investing most of the tunnel with the masonry lining. One problem still remains for discussion : wh}- should the Romans have gone to the trouble and expense of constructing two long tunnels through the hill so close to each other? The view commonl}- held is that the Grotta di Posilipo was for the public traffic between Puteoli and Neapolis, whereas the larger Grotta di Sejano was for the private service of the Villa Pausil3pon. If this were so, it is strange that such a magnificent appendage to an Imperial Villa should not have been a theme of eulog}- to any of the court writers who were acquainted with the locality, and who never lost an opportunity of praising ever}^ luxurious contrivance that might display the magnificence of their patrons. Taking this omission into account, with the general belief that the Grotta di Sejano was of an earlier period than the other, that it was larger and better lighted, and coupling with this the evidence that the road that passed through it was continued along the coast of Posilipo to Neapolis, and was a public way during the fourth centur}', we are led to the conclusion that in its earlier and later periods the Grotta di Sejano was a public thoroughfare, although it is possible that in the interval, after the shorter route by the Grotta di Posilipo had been opened, effective measures may have been taken to secure privac\' to the Imperial Villa. X'ALLONE DI GAIOLA Just below the point at which the tunnel opens into it, the Vallone di Gaiola widens, no doubt owing to the little cutting which sets the tunnel back from the bend of the valley. The steep hill-sides, neatly terraced and planted, cottages nestling among the fruit-trees, the thick growth of canes along the bottom, make a typical picture of peace and fertilit}'. Along the ridges of the hills aloes and old and stunted oaks, though now confined to hedgerows, are here as elsewhere on Posilipo, the last survivors of a more widely spread coppice vegetation which preceded the vines. The hill which divides us from Trentaremi Bay has been greatly reduced on the seaward side b}' landslips which have left it but a ridge, so narrow and so precipitous that the workmen can only reach their pozzolana quarr}- by a small passage through the rock. Beyond the tunnel we could not perceive, nor could we learn, that there were any remains of paving to the roadway ; but then, neither has a paved floor been detected within the tunnel itself. It must, however, be remembered that the natural forces of erosion act with extraordinary rapidit}' on the light and loose materials which overlie the more solid yellow tufa of the hill. Observations made on the valleys which seam the similar slopes of the Camaldoli massif, have shown with what rapidity beds of valleys such as these may change, becoming deepened in some places, silted up at others. The road, after emerging from the tunnel, pursued the downward course of the valley, passing some buildings of minor importance of which nothing but the foundations under some modern cottages now remains. Still descending, the road passed below the little Temple which occupied an imposing position above on the right. In front and at a greater altitude still, was a magnificently situated house, which we have named the House of Pollio, lying high above the road near the point at which it passed through a short artificial tunnel 45 feet long by 11^ feet wide. Many caves have been cut in the tufa sides of this valley, which are still useful to the colo)ii as barns or stables. Of the many 28 VALLONE DI GAIOLA buildings which have disappeared in recent times we noted one with reticulated walls at the foot of which many fragments of coloured marbles might have been picked up ten )-ears ago, but the site has since been obliterated by an alteration of the road leading down to the Gaiola. On the east side of the valley, a large cave hewn in the rock has been adapted as a water reservoir, by the construction of a massive retaining wall 9 feet in height across the entrance. Some Roman brickwork and ancient cement indicate the antiquity of a work which was so well executed that it is still in use. The capacity cannot have been less than 30,000 gallons. Fig. 8. A W.\ter Reservoir. THE BUILDINGS OF THE PAUSILYPUIM The most important antiquities on the Gaiola ridge are so situated that, although not far from Naples as the crow flies, the}' are by no means easy of access. The ruins stand on private property and may only be visited by permission of the owner. They may be reached by sea, or by stumbling through the tunnel of Sejanus, or b}^ a lane, roughly paved and so narrow and dusty that tourists do not go there as often as the interest of the antiquities or the charm of the site would lead one to expect. THE THEATRE Among the ruins, the one which attracts most attention and which appeals strongly to the imagination on account of its remoteness from any possible audience, is the Theatre, now in the grounds, part garden, part vineyard, of a private villa. We come upon it at a turn in the path that leads to the house: it is a surprise when emerging from between the vines, to behold an open space, surrounded by a semicircular auditorium in the hill-side. The seats are discernible here and there through a thick clothing of weeds and grasses, genista, fennel, and hemlock ; while some stunted oak-trees are growing on the top of the hillock, at the back of the ruined wall of the gallery, the green standing out against the blue sky. The arena is before us, and though overgrown with weeds is clearly defined ; beyond it, above a clump of green shrubs and over- looking the upper tier of seats of the theatre, stands the modern white-plastered, unpretending country' house of the owner, Signor Acampora. The site was first cleared in April, 1842; but the existence of a structure of such conspicuous form and dimensions must have been well known to the successive tenants of the property : and failing any other in the vicinit}', must in all probability have been the theatre o is o 5 O THEATRE 31 mentioned in the sixteenth-century manuscript of Fabio Giordano.' And so, although the credit of its rediscovery cannot be given to the architect emploj'ed by Monsignor di Pietro, yet to him is the greater merit of having been the first to clear it, and render it available for study. At present, the western side is again encumbered by debris. The theatre has a twofold claim upon our attention. In the first place it appears to have been the private theatre of an Imperial villa, and very few other examples of private theatres are known,- and secondly, there is a remarkable absence of any stage building, from which we may infer that it was constructed during a transition- period, before the Romans had entirely overcome their early prejudice Fig. 10. The Theatre. (From Collina di Posilipo.) against stone-built theatres, which they considered a yielding to Greek luxury. It is pretty certain that for a long time, in consequence of this feeling, no theatre was built in Rome of anything but wood : the first stone theatre, Pompey's theatre, was not constructed before 55 b.c. At the same time, like our modern Elizabethan Stage Societ}', they may have adapted the structure of the stage to the nature of the play ; and we know that Menander was played in Naples in the first century. It would not be unnatural for the patrons of the stage, even when accustomed to the greater comfort of stone seats under an awning, velarhtm, to remain conservative in their views about stage fittings and ' A. Gervasio, Osservazioiii iitlonto alciiiic aniiche iscrizioni die soiio o furotto gia in Napoli, 1842, p. 62 ; F. M. Avellino, Bull. arch. Nap. i, p. 86. * Heinrich, su Juieital, Sat. vi. 70, p. 228. THEATRE ■^s:^<-:: 5 lie oj" House Fig. II. Pla.n of Theatre, n ODE ON '^ r,[i;jMmm§^ ViRIDARIUM AND OdEOiN. D 34 THEATRE furniture and to desire to retain the wooden stage ; which might, moreover, have advantages in the way of rapid alteration or of easy removal. The theatre measures about 51 yards in diameter, and its orientation is about 4 E. of true south. The orchestra was paved with white marble, of which fragments have been picked up near the main stairwaj-, but none of the slabs are now in situ. In the middle of the arena is a shallow rectangular pit, shown in fig. 13 and marked c on the plan, measuring 20 feet long b}' 13 feet wide and sunk out to a depth of 2 feet. Its sides are faced with tufa (opus retiailatum), but originallj' it undoubtedly had an inner lining of marble veneer, and, as Ruggiero has alread}' suggested, might have been a compluvium into which rain-water was conducted ; but, as there was no lack of water ' laid on ' to all the Pausilypon buildings, it is more likely to have been a tank to receive the jets from fountain figures like those of the House of the Vettii in Pompeii. On either side of the tank are two deep holes [v, v, v, t')> the purpose of which seems to have puzzled the original excavators. We believe them to have been part of the contrivances for supporting stage scener}^ curtains or awnings, carried on wooden masts or uprights stepped in the holes. The channels down the sides of the holes even suggest the positions of heel-ropes used for raising or lowering the masts. Further excavation might possibly reveal their purpose more clearlj'. In the construction of the cavea, the builders, following the Greek practice, took advantage of a hiU-side facing southward, which thej' hollowed out and lined with the seats of the auditorium. The details of the plan are not quicklj' grasped, for although seats and stairs are discernible here and there, yet the thick clothing of weeds, with roots driving the stone-work asunder, no less than the well-intentioned restorations of Pietro, have so obscured the original work, that even the exact number of the rows of seats cannot be counted without much care. Further, owing to the entire removal of the marble slabs with which the lower benches at anj' rate were once covered, it is impossible to make measurements with accuracy, which is the more regrettable as the two sides of the theatre are not exactl}' alike, and so the commonly accepted symmetrical plan published by Fusco must be regarded as unreliable in detail. The auditorium consisted of 18 or 19 rows ' of seats surrounded b}' a gallery. The seats are divided into two concentric tiers {cavea ima ei media) by aprecinclio, 4 ft. 4 in. wide, with a wall 4 ft. 4 in. high ; 12 rows of seats being below and 6 or 7 above the precinctio. The stairways ' The number 17 given by Fusco and repeated by Beloch is certainly too small. THEATRE 35 are all above ground ; there are no vomitoria. The seats of the inia cavea are divided into three f //«M >\L~^ l L-J i /.~X i >Tc^r Fig. 14. Mural Dfcorations of Room k. painted in red. Then a row of small panels arched over, in red, and traversed by a green band of simple kej- pattern. The lower red border below the panels was further decorated by a yellow running ornament of foliage and scrolls. The wall of the corresponding room r' on the west side was decorated with a fresco representing a two- horsed chariot being driven bj- a woman wearing a cloak, which was THEATRE 39 depicted fluttering out to show the pace of the horses. The back- ground was painted dark red. Of the furniture of the theatre there is not much to relate. The marble seats had all been removed before 1840. But among the objects unearthed by Camillo di Pietro were some of the stone brackets which supported the poles to which the velarium was fixed. The head of a youth, perhaps an Apollo, was the only piece of sculpture found here. The number of spectators for whom accommodation could have been provided has been computed as 2,000, allowing 23 inches per person. There would have been 600 seats below the precinctio, 850 above it, and at least 200 more under the portico, and 350 on the precinct, arena, and elsewhere, with a full house. Under the blaze of the noonday summer sun, for the theatre faces south, the gold and brown of the burnt grasses of the slope, the oaks above and the white house, green and white against the sky, the bright green of the vines below, be3'ond which, at a stone's throw, rises a cliff in the adjoining vineyard, enhancing the brilliance by the dullness of its unlit tufa face — this all makes a little picture to delight, and the antiquarian interest becomes, for the time being, quite secondary. No pleasanter site or more charming landscape could be found for the erection of the massive buildings clad in white marble which no doubt rose among masses of green, for the Romans well understood how to gain the full effect of architectural beauty by the contrast of masses of evergreen and other vegetation. THE ODEON Some 50 3'ards south of the theatre, where the ridge between Trentaremi Bay and the Yallone di Gaiola becomes higher, another large building has been raised against the hill-side. This was the Fig. 15. The Odeon. (From Collinti di Posilipo.) Odeon. Odeon and theatre both, were carved out of opposite hill- sides, so that the auditoria face one another ; their chords also were parallel, and there is but little doubt that galleries of communication, the foundations of which are still partly traceable, connected the ends of a portico which la}' behind the stage of the odeon with the wings of the theatre opposite. See the plan on p. 33. Thus a quadrangle was enclosed which may have served on occasion for extended spectacular performances viewed from the theatre, or at any rate as a viridariuin or garden between the portico of the odeon and the wooden stage building of the theatre. Such a viridarium was attached to the odeon of Herodes Atticus at Athens. It is traversed by an underground aqueduct, whose course is indicated in the large map. The whole of the area is at present under vines and fruit-trees. The late writer Hesychius tells us that tlie purpose of an odeon ODEON 41 was for the performances of rhapsodists and of players on the cithern, but it is more than Hkely to have also served for the recitations and declarations of poets and orators and for public readings. For this purpose Hadrian had built the Athenaeum in Rome, and, whatever the tastes of his predecessors may have been, would not have considered an Imperial Villa complete without a building for this purpose. The odeon was all built of concrete faced with opus lateritium and reticulatum ; the more important parts were overlaid with ornamental marbles, among which rosso autico, giallo antico, Parian, and pavonaz- zetto were abundantly used, while the walls of the passages and smaller rooms were coated with plaster and were coloured. y;^ k .•« Fig. 16. The Imperial Bo.x and Passage (p). Unlike the theatre, the odeon was furnished with a well-pro- portioned permanent stage building, at the back of which ran the portico, which formed the southern side of the garden quadrangle. In the auditorium the positions of the cunei and of the number of the gradus were carefully recorded at the time of the excavation, and have been indicated on our plan. But the chief interest lies in the presence of a very large space in the centre of the auditorium, at the level of the fourth gradus, which must be regarded as the Imperial box. The back wall, with a large apse faced with tufa for a height of 18 inches and then with red bricks carefully laid, and originally wainscoted with coloured marbles, is one of the best preserved parts of the building and now holds up the hill-side behind. It is now about 16 feet in height. 42 ODEON Fig. 17. Western Stairway to Auditorium and Entrance to the Small Room under it. PASSAGE - BOX V-^VWV^^'JStVUV^V^-^s/K,^'^^ Room -«dfr 1^ •^'""^/■.'•j C^fvU-i :^ SUi 4:' PASSAGE " STAGE q J( A A A A fl ,^ A " 5fadt \.-t "00" " -->^AAJkAK■•.V^ •:v? Fig. 18. Plan of Structures shown in Fig. 17. The Imperial seat or pulvinar (s) is indicated by a small elevation in the middle, and in the apse behind, was a pedestal (/>) that would have supported a statue. The walls are still studded with the iron holdfasts which kept the marble overlay in place, and curved pieces of the giallo ODEON 43 antico mouldings have been found. The approach to the Imperial box is, from behind the lateral cunei on both sides, by two passages (p, p'), each 6 feet in width, with a door at either end of each passage. The walls of the passages are reticulated, plastered and coloured red and yellow. On either .side of the central box were two lateral cunei, affording room for six more gradus. The means of access to the seats of the auditorium is ample and excellentlj- planned. Two passages {-a^iohoC) (t, t'), 6 feet in width, and quite separate from those leading to the Imperial box, lie between the auditorium and the stage. Three small radial stairways rising from the orchestra led to the lower seats, and two others (x, x') in laterally placed voiiiHon'a provided for access to the uppermost of the seats of the cunei, which could thus be reached from the parodoi (t,t'), without persons passing in front of, or intercepting the view from, the Imperial box. Two small and dark rooms of awkward, triangular shape were con trived under the cunei next to and partly under the lateral stairways. Their walls are faced with reticulate work, the angles of which are formed not of the usual brick lateritium, but of tufa, as shown to the left of the stairway in figs. 17 and 18, worked in three-course ashlar. The interior was plastered white. The ceiling is formed by the underside of the sloping concrete floor above, and has a steep pitch in consequence. It is about 18 inches lower along the curved side than at the other side. Upon the underside of the ceiling may be clearly seen the impression of the scaffold boards, upon which the newly mixed concrete had been thrown. Raised a little above the orchestra, a long and narrow stage, measuring 56^ feet long by 9 ft. gin. wide, was reached by a few steps from the passage (parodos) at either wing. It has been suggested that such steps were used by actors pretending to iiave arrived from a distance. The foundations of the stage building, of opus lateritium and reticulatum, are in a tolerably good state of preservation : they show the elevation of the proscenium to have been of the usual type, with niches, three semicircular and two rectangular, in which messengers, or those charged with the maintenance of order during the performances, may have sat. At the back of the stage is a hemicycle, once ornamented with six fine, fluted Corinthian columns of cipolline marble about 9 ft. 6 in. in height and 14 inches in diameter, the lower 3 feet of the fluting being filled with a bead. Two of these columns were found entire, and a photograph of one of them which has been moved to the 44 ODEON house of Signer Acampora, was used for the accompan3-ing sketch as a suggestion of the original appearance of the hemicjxie. At the two ends of the stage building are two small chambers, about 6 feet square, which may have served as dressing-rooms for the performer, who would have stepped forth ' well combed, draped in his new toga, wearing rings on his fingers, his larynx made supple by an emollient potion, and gazing on the audience with a caressing eye'.' Fig. 19. Slx.gested Restoration ok tiik Back of the Stage. Both parodos t and the passage to the Imperial box p open out of a passage g which led out to the portico at the back of the stage building. The walls of the passage were painted pink. The Portico. — Upon a pavement raised i ft. 3in. above the level of the garden, stood eighteen brick-built columns covered with fluted stucco, in the Tuscan style. These columns were 21-5 inches in diameter and 7 ft. 5 in. apart and stood on squared marble bases. The centre of the portico, for a length of about 14 j-ards, projected slightl}' into the quadrangle, thus making an effective break of line. The general appearance of the columns and of the western corner of the portico as seen from tlie viridariuni, would probabl}' have ' Persius i. 18. quoted troin Boissiei". ODEON 45 resembled the view of the comer of the Pompeian peristyle shown in fig. 20. The pavement was formed partly of little chips of marble of varied colour, in which African breccia is abundant, and partly of white mosaic, and between the columns were inlaid bands of white marble sloping towards the wall of the odeon. The pavement lay about 3 feet below the top of the ruined wall of the hemic^-cle. The white mosaic tesserae, which were roughly cut, have been for the most part turned over by the mattock of the cultivator, but their abundance at certain spots still shows the extent of the ancient pavement. Fig. 20. A Corner of the Peristyle in a I'ompeian llousi;. Buildings extend for many yards on both sides of the auditor- ium. Those to the east have not as yet been excavated, but on the other side to the west, overhanging Trentaremi Ba}', several chambers D, E, F lined with opus reticulatum have been cleared of earth. They appear to have communicated with the passages g and p leading to the Imperial box. Between the south wall of the passage p and the hill the ex- cavators have opened a long narrow chamber b, nearly 5 feet wide and covered with a barrel vault. The construction of the west end of this block of building is shown in fig. 21. The reticulated parts of the wall were covered with red painted plaster. On the floor above were other apartments, probably on both sides of the Imperial box. The floor of one of these, c, of opus signinum, may just be seen in the corner of the passage, resting upon solid ground at a height of about .y^if^^^^^^^^^^^^^^^^^^^^^^^^m^^ i ^PP' ' 'V ,1 ^^^^^^^^^^^H^^^jS^SflflESflfl^^^^^^^^^^l 1 eJIi^^^^^^^^^Il .^5 1 '■ ■■■•■■:.-. m o < .J a o u S H O ■ ^ *♦ • ».'.■■ N-::^ ODEON 47 13 feet above the floor. The rest of the building is on the edge of the cliff, and much has been lost by the falling of the cliff side. At the time of the excavation of the building it was considered that there was clear proof that the odeon had originally been roofed over, as was usually the case. The measurements of the odeon are : length of stage building over all, 116 ft. 6 in. ; radius of hemicycle of stage building, 16 ft. 4 in. ; depth of stage at sides, 9 ft. 9 in. The Imperial box measures 27 feet bj' 20 feet; apse, 10 ft. 6 in. across, by 3 ft. 10 in. deep. Construction. The south side and the apse are faced with brick- lateritium, excepting for the lower 18 inches, which is of tufa. The passages leading to the box are faced with reticulate work, but the angles are of brick-lateritium, of which thirteen and six courses alternately are ashlared into the reticulate work. There is a good example of this kind of work on both sides of the eastern end of the passage p, where the thirteen courses occupy 2 feet, and the six courses rather less than I foot, the average being i-8 inches per course. The jambs of the entrance to the triangular chambers under the cunei are built in tufa-lateritium laid in three-course ashlar in 12J inches, which inclines us to the belief that this part of the building is older than the brick- faced door jambs of the small rooms b, b' on the opposite side of the parodos. The seats of the auditorium, which are 2 ft. 3 in. wide, are also built in tufa work. So that in the absence of evidence to the contrary, we refer the auditorium to an older period, perhaps to the time of Pollio, while the stage building and the Imperial Box and the passages leading thereto, may be referred to a later, perhaps Hadrianic, period. The reticulated wall of the vomitorium at the west end of the auditorium appears to have been thickened at the later period, its original thickness having been i ft. 6 in., final thickness 2 ft. 2 in. THE HOUSE BY THE THEATRE The Sfaincay. We now proceed to describe the antiquities on the hill to the west of the theatre round about and under the house formerly occupied by Di Pietro, but now in the possession of Signor Acampora. It is one of the many instances on Posilipo of the permanence of dwelling sites, of a Roman building being still utilized as a modern habitation. During the course of the excavations of 1842 a stairway consisting of several short flights of steps and landings was found to lead from a passage from the arena of the theatre westward up the hill-side to another large building, supposed to have been a Nvmphaeum, and probably also to the galler}- of the theatre. After a re-examination of the site, we have, however, come to the conclusion that the stairway led to a dwelling-house situated alongside the theatre, and of which the so-called Nymphaeum, to be described presentl}', may have been the peristj'le. At the bottom, near the corner of the viridarium, the early excavators thought that the}' recognized a few steps and a landing, whence a slight incline some 14 yards in length led higher. The incline gave access to a small chamber with a reticulated vault, and then to a second, though in a more ruinous state. From the top of the incline a wide flight of four steps, 11 feet in width, led to another landing; the lower three steps were covered with tiles in fair condition, but the tiles on the top step were much broken. On the left of the landing, at an elevation of i ft. 3 in. and between two pilasters, was a marble threshold indicating a doorwaj^ ; and on the right another threshold-stone of piperno with a socket for a bolt. Eight steps led higher in a northerly direction ; the first four were covered with tiles and were 8 feet wide, but the other four were only 6 feet wide, the topmost being formed of blocks of piperno. Unfortunately the old stairway' has been replaced, or at any rate is quite hidden, bj' the modern steps and path leading to Signor Acampora's house : we can onlj' approximately follow Fusco's descrip- tion of its ancient course ; to examine it critically is now impossible. HOUSE BY THE THEATRE 49 The Rooms under Signer Acampora's House. A ground-plan of the cellars under the house is shown in fig. 22. Such parts of the walls as can be seen are faced with reticulate work, as indicated by the toothing in the figure, and are finished off at the corners and door jambs with tufa-lateritium. Unfortunately the further excavation necessary to reveal the complete ground-plan Fig. 22. Chambers beneath the House of Signor Acampora. Roman walls, black: later walls, dotted. cannot be undertaken without danger of weakening the foundations of the dwelling-house above, but enough of the several parts of the building has been cleared to show that they have varied orientations, indicating a modification of the original plan by later additions and alterations. On the west a large L-shaped modern storeroom [a) has been formed by building a new wall across the Roman work : it is about 9 feet in width, the length of the further part being about 30 feet. 50 HOUSE BY THE THEATRE A narrow vaulted passage b, 6 feet high (compass-bearing 170°), leads past a walled-up doorway r, flanked by lateritium jambs, into a small cross chamber d with a barrel roof 9 feet high, perforated by a small shaft 5. Opening out of 2 d J: < o < THE HOUSE OF POLLIO At a distance of rather more than 50 3'ards east of the odeon is a piece of rough ground like a small rectangular plateau, standing high above the valley of the Gaiola on the north and east, and raised about 15 feet above the vineyard to the south. The ruined walls and rubble with which it is surrounded mark it as a site of importance, and, while affording a passive protection from the mattock of the cultivator, have been a happy hunting-ground for grubbers in search of antiquities. Both hill-side and summit are covered with a light wood of copse oak. The old builders found it necessary to strengthen the steep slopes with retaining walls and arched substructions, massive remains of which bear enduring testimony to the weight of the buildings above. From the top, 130 feet above the sea, there are fine views up and down the Valley of the Gaiola and across to Vesuvius. It was the site of the largest dwelling-house yet discovered here, and we believe it to have been the principal house of the Villa Pausilypana. We name it the House of Pollio. Our identification has been partly suggested b}' the fact that in the larger villas of the Romans, or for the matter of that in so manj' villas of the Renaissance, the principal dwelling-house was built upon the most elevated part of the grounds, both with the intent that the view might be extensive and varied, and also that as much air as possible might blow through the apartments during the summer heats. Pollio would have been more than likely to have put his house in such a position Ubi gratior aura leniat et rabiem Canis et momenta Leonis. Hor. Ep. i. 10. 16. Of the upper storej's nothing remains but scraps of partition walls. At the western end are a couple of reticulate tufa walls a, about a yard apart, which run parallel for a space of some 14 yards : the innermost is of interest, in that it shows a doorwaj' in a tufa wall, into room b, filled up with later brickwork. ,J^^^^Ml/i^^#Cl^.^ JL.4^,^^ #.^^^^^,^'t Fig. 33. House of I^ollio. HOUSE OF POLLIO 65 At the north-east corner a large room was floored with a fine pavement 6 inches thick, with a pohshed facing of coloured marble fragments very like, and probably coeval with, that in the Scoglio di Virgilio building. Remains of newer construction in brick, ee . indicated in the upper right-hand corner of fig. 36, rested directly upon the top of this pavement ; and inasmuch as the vault over the chamber F underneath has been strongly underpinned, it may not unreasonably be inferred that substantial additions and alterations were made on Fig. 34. Walls to East. Drawn to the scale on fig. 33. Fig. 35. Rooms to the South of the House of Pollio. this side of the house : the later work being generally in brick, like the aforementioned filling of the doorway. Several of the rooms on the lower storey, in a fair state of preserva- tion, are shown in plan in fig. 33 ; but there are many others which are so much encumbered with the debris of fallen vaults and roofs, that further excavations are necessary. For the most part the walls have lost all traces of colour, and scraps of pigment or moulding on detached fragments of plaster are too often the sole clue to the nature of their former decorations. Along the southern side several small dismantled chambers and passages (c-g, k, l, m) may be made out : others lie deeper in the hill- side and deserve further exploration. On the side overhanging the valley to the east are some larger apartments (e, f, g). Their doorways faced east ; that of the middle chamber f being 9 ft. 6 in. wide and arched over. This chamber, F 66 HOUSE OF POLLIO which is more complete than the others, has been greatly diminished in size by the introduction of massive piers and arches for under- pinning the vault, which is evidence in favour of the addition of a heav}' superstructure upon the floor above. Chamber v was originally a plain square room with reticulate tufa walls, plastered over. The east wall has been strengthened by brickwork and concrete ; a pier, 6 feet through, being raised on either side of the archway ; and at the back of the room three other piers, each a yard square in plan, were built of tufa lateritium in 4|-inch courses, which Fig. 36. House of Pollio. Roo.m F. piers carried two brick-faced arches, resulting in the production of two alcoves at the end of the room. Under one of the arches a small opening leads into the passage dd, now choked with debris. Two fragments of a fine fluted alabaster column (No. 19, p. 222) were found among the debris of this room. The work clearly belongs to two if not to three periods. The reticulate tufa-faced walls belong to an earlier period, the abundant use of brick to a later period ; and at certain points we believed it possible to make out that patches of masonry, built of two courses of brick alternating with one of tufa, had been inserted after the completion of the adjoining reticulate wall. HOUSE OF POLLIO 67 The north wall of room g is faced with tufa lateritium. and may therefore be referred to the earher period. About 20 or 25 feet down the steep hill-side and some 6 yards east are the ruins of a lower storey h, j : the construction is of the earlier period. Here we found mosaic cubes, pieces of rosso antico marble, and much brilliantly painted plaster, silver-grey and bright blue being the predominant colours. The wall j is faced with two courses of brick alternating with one course of tufa. Along the north side are other rooms. The principal one b stood Fig. 37. House ok Pollio. W.all between Rooms A and B. a httle back from the north front of the building : and the two flanking walls on either side of a may have been formerly spanned by a wide arch. By the entrance, two relieving arches have been built in the wall. The wider one spanning the doorway has been partly replaced by a patch of reticulate work clearly shown in the photograph. It is an interesting case of where a defect in the older masonry has been made good by work of a different character, which has been carefully' executed notwithstanding the fact that the whole was soon to be concealed by plaster. The door jambs were constructed at different times, that to the left being faced with alternating courses of two bricks and one tufa stone, the three measuring 9 inches ; the other on F 2 68 HOUSE OF POLLIO the right is built of" brick only, the bricks being 9 inches square, and taking 4! courses to rise 9 inches. The latter is believed to be the more recent of the two. This doorway leads into the Painted Room. THE PAINTED ROOM Within room h sufficient portions of the painted plaster have been preserved to enable us to form a tolerably complete notion of the main outlines of the mural decorations. Tiie room, very nearl}' square, 19 by 18 feet, is covered by a low- pitched barrel-vault. The side walls measured to the spring of the vault are 13 feet high : the end walls, following the curve of the vault, rise 3 feet higher. There are doorways through the north, east, and south walls, and a small square window close to the ceiling in the middle of the north wall. The doorwa}- in the north wall has been partly blocked by a modern flight of six steps, leading down to the floor, four feet below the accumulated debris outside ; the south doorway, spanned b}' a horizontal lintel, with a window over, opens into a narrow passage-like chamber c bej'ond, plastered with fine white plaster, and lit b\- two small window openings in the party-wall ; and it is ceiled bj' the same vault as the larger room. It no doubt originall}^ communi- cated with the rest of the house b}- a long vaulted passage dd, into which the third door in the east wall also opened, but the way is now blocked b)' some rough masonry of a later period. An angle in the masonry' ma}- be distinctly seen in the south wall opposite the entrance. Another feature which must be mentioned is a singular recess to the right of the entrance (fig. 40, and shown as an inset in fig. 33), measuring some 3 feet in width by 6 in height, with a narrow bracket painted black and projecting slightl}' bej'ond the wall line. It appears to have been used as a cupboard, and, though cutting into the scheme of decoration, to have been a part of the original construction of the wall. The aperture may conceivably have been closed b}' a wooden door upon which the mural decoration was continued. The cupboard is deeper above than below. For the middle third the plastered back is sloped out towards the room. The back of the lower portion appears to have been covered with marble. The scheme of decoration of this room is of a familiar Pompeian architectural type. o a. O O o a. LL o o X LLl I o z < CL < 5 HOUSE OF POLLIO 69 The wall space below the cornice is primarily divided into the three usual areas by two horizontal bands, which delimit the principal area from the dado beneath and from the frieze above. On each wall the area is still further subdivided into the three panels, so character- istic of mural decoration of the Pompeian period, and of which the origin has been traced to the conventional tripartite division of the proscenium of the classic stage. The cornice is of plaster, 5 inches deep, and is divided longitudinally into three divisions like an architrave by bead mouldings. The north and south walls are continued up to meet the barrel- vaulting, so that between the top of the frieze and the cornice, lunette spaces are enclosed. The west wall is not pierced by any doorway or other opening ; and so, as the painting upon it is also the best preserved, we may begin with a description of it. The West Wall. The Dado, in Pompeian red, is 31 inches high. Above a ' skirting ' of 10 inches in height, the dado is divided into three spaces by two white-framed panels, each representing an arch supported by two piers, between which flies a white swan ; the whole panel is repre- sented as supporting the tall, slender columns which form part of the architectural decoration of the wall above the dado. \ er}- close parallels may be found in the paintings of many a Pompeian house, notably in those on the walls of the House of Diomed, wiiich resembles this one in having also been a country villa. Across the middle of the red side-spaces of the dado, from panel to panel, a narrow olive-green wreath supports a central, yellow, foliated ornament ; in the middle space the wreath is replaced by a straight band of a conventional pattern of yellow rectangles enclosing flowers (fig. 38, c). The dado is divided from the main wall-space by a white band, i| inches wide, which, hke the narrow white skirting line, runs all round the room at the same level. The Wall-Space. The principal part of the wall-space rises from the dado to 9 ft. 4 in. above the floor, and also is divided into three black panels b}- an architectural motif. The left side-space is decorated as follows. A red band (i§ inches wide) runs up the corners of the room ; this is divided by a narrow 70 HOUSE OF POLLIO white line from a broad red band, bordering the large black panel, measuring 6 ft. 4 in. by 4 ft. 8 in., which is contoured bj- a thin white line, and bears an inner yellow frame, i§ inches wide, of a fan pattern b. Within the frame was a small picture (12 in. by 20 in.), recently removed. The right side-space is similarly treated, but the middle black panel is bordered by two vertical greenish lines, and the design of the yellow Fig. 38. Details ok Painted Borders. House of Poluo. frame is different (fig. 38, a) and the picture was square (155 bj- 15^ in.). A hole in the plaster of the third panel, indicating a looted picture, measures 192 x 12 inches. The three spaces are divided b}' two columns, shaded white and cream. The bases of the columns rest upon the plinth-like, white- arched panels of the dado, and the upper ends are carried up on to the frieze. This, the main scheme of the decoration of the wall, is divided from the frieze b}' a white band, 2| inches deep. HOUSE OF POLLIO 71 TJie Frieze. In the centre an oblong pedestal supports a tall yellow cande- labrum with vase-like base, and from the corners of the pedestal two yellow thyrsi lean in a slanting direction against the backs of two white canopies, spread back to back over the columns. The canopies are connected by a yellow band of fan pattern ornament, which is stretched in front of the candelabrum. Of the two pillars supporting the back of the canopy, the nearer is surmounted by a yellow capital and is continuous with the detached column, the full relief of which is made conspicuous by an oblique bracket painted above the capital. The roof of the canopy is supported on longitudinal rafters let into four transverse joists. Above the foremost pillar of the canopy is an ornament, a cande- labrum-like pinnacle ending in five ' flames ', to which is tied a garlanded festoon, which stretches across to the black band near the corner of the room. Lower down, a second band of green foliage tightl}- spans the same space and supported a central figure, perhaps a bird, which has been removed. Around all, runs a black band with a thin white bordering, which is continued across the top and down the sides. The scheme of decoration of the opposite wall was apparently identical, and though less complete owing to the existence of the doorway to the passage, yet the mural painting is less mutilated, probably because owing to the indistinctness of the small paintings on the black panels, no vandal has hitherto considered them as being worth}' of removal. Their prevailing ground tints are blue and green ; they may have been marine views. The North and South 1 1 'alls. The north and south walls were treated alike, though with a scheme of decoration differing in detail from that of the east and west wails. The middle dado space of the north wall is crossed by a yellow band of conventionalized arch pattern, fig. 40. The middle black wall-space is flanked b}' two j-ellow columns standing over the swan panels and terminating in the frieze in highly ornate candelabra-heads, upon which rest the same inverted crescentic objects So seen in a Pompeian Fountain Mosaic in the Naples Museum, fig. 39. Above the crescent is some six-rayed object, and over it depends a white bordered banner hanging from the cornice, a 72 HOUSE OF POLLIO decoration not infrequently found in other frescoes of the period. To the shaft of the pole is attached a band of conventional geometric pattern (a). Near the corner of the room a yellow column and two pinkish pilasters carry a canopy surmounted by a winged seated figure and two candelabrum-heads as finials, to which a green garland is bound. Lower in the frieze, two pheasants, rcgarda)it, perch upon upturned scrolls. The red-painted lunettes (north and south) are surrounded by the white plaster cornice-moulding and black and 3'ellow lines. Each was decorated b}^ two s\-mmetrical paintings which have been neatly cut out of the plaster. Fig. 39. Fresco ox North Wall. The Cei/inp: 'i- The ground colour of the vaulted ceiling is red ; it is outlined at the spring b}' the white plaster moulding already mentioned, and this is thrown into further relief above, just as it was below, by a broad black band edged with a thread of yellow. Within a plain yellow border, lined with white, are four square green frames containing decorative figures; only one of which, a winged horse, now remains. The green frames are connected to Fif.. 40 North Wall. DOOR TO PASSAGE Fig. 41. South Wall. 74 HOUSE OF POLLIO one another by three straight decorated bands. The swastika in circles, alternating with a diamond ornament with a spot, is the pattern on the outermost of the bands fig. 38, d ; it is painted in yellow on the red sround, 2 inches wide, and is continued to the outer yellow border. The middle band is formed by a dark line relieved by two lighter lines, with a beaded ornament along the upper one. Unfortunatel}- a great piece of the plaster has fallen from the middle of the ceiling, and so, much of the decoration is lost : the rest of the painting is faint with age and damp. At least one of the central figures in the green squares has been intentionally cut out, but, let us hope, may, with the other missing paintings, be )-et traced in some collection of antiquities. KiG. 42. Halk of riiK Ceiling Decoration. The Background should be red instead of black. The general stj'le and scheme of the mural paintings in this room so clearl)' resemble that which we know to have been in fashion in Campania, especially in country houses, before a. d. 79, that it is eas}' to assign this as their approximate date. It is impossible to believe that the ver}^ close resemblance of the minutest details of the decorations in the Villa Pausilj'pon and in that of Diomed and others at Pompeii could bear anj' other reasonable interpretation than that of contem- poraneit\- ; and unless it be a case of deliberate archaism, we certain!}- cannot attribute the execution of these mural paintings to the later days of the Empire, as has been suggested b}' some. HOUSE OF POLLIO 75 Attention may be drawn to the following points of resemblance. An ornament like that on the ceiling is depicted in Niccolini, pi. LVIII, where we see a similar border without the swastika, but with the alternate diamonds. The fan border was not uncommon in Pompeii: it is associated with a green foliated band in the house Reg. IX, ins. vii, pi. III. The thyrsi and corner scrolls are of frequent occurrence, as for instance in the Pompeian Lararium figured by Overbeck, fig. 146, and a similar altar-motif as a decoration occurs in the House of the Centenar}'. Fresco No. XXII in the House of the Centenary combines several of the architectural elements of the Gaiola paintings with the fan border, the foliated band, and the twisted scroll finials. The candelabra have been regarded b}' Mau as a distinctive feature of his 'candelaber-st^^l' or fourth Pompeian st3'le. On the other hand, the perfection of the preserN^ation of these paintings at the time of their discovery would seem to need further explanation. Unlike Pompeii, the Pausilypon site was not buried up and preserved by the eruption of a volcano. A long-continued habitation- of the house during the centuries that Imperial Rome survived Pompeii would assuredly have ended in the effacement of these wall paintings or in their replacement by others. We must therefore suppose that the lower rooms were not in use for long, perhaps other rooms were built over them, and that when the upper storeys fell in ruins, possibly as the result of an earthquake, the lower rooms got buried and preserved under the debris. The room is now used as a shed for faggots; and damp having found its way through the 6 feet of superincumbent soil and concrete vaulting, is slowly but surely causing the decomposition of the colours on the walls. This is particularly the case on the east wall, where the salt in the plaster has effloresced with the damp, and has in places given rise to a very remarkable double effect ; as, for instance, on the conventional borders of the panels, in which, by the salts and damp, the pattern has become raised so neatly- and clearly that the design is in relief, and also the original colouring has been reversed, so that the yellow has turned black and the black, white. If certain scattered fragments found down the slope below the house may be accepted as evidence, some portion of this building seems to have been provided with hollow walls for heating, for amongst the other debris were several fragments of tiles, with legs, such as those with which the hollow inner walls of sudatoria are formed. These tiles are not quite of the same pattern as those used in 76 HOUSE OF POLLIO the circular sudatorium, to be presently described, but have stouter and more mammillate legs. The legs are perforated for the iron nails or holdfasts by which they were fixed against the wall, and in this respect resemble the tiles fixed in the hot room of a private house in the Vicolo del Gallo in Pompeii, though they are of different manu- facture. The tiles used in the circular sudatorium of Posilipo exactly resemble those emploj'ed in the public Baths near the Forum of Pompeii. Some fragments of a large dolinm lay in a room not far off. Fig. 43. Upper E.nd of the Tunnel. THE TUNNEL ON THE VALLEY ROAD Below the House of Pollio the level of the vallej'-bed drops abruptly, and the torrent bed, along which surface-water runs in wet Fig. 44. South End of the Tunnel, with Vault A to the Right. weather, hitherto confined in a narrow channel with stonework sides, discharges into an old tunnel, the floor of which has been buried deep in accumulated silt. As every square yard of fertile soil in this sheltered valley is of value, the farmer has parti}' blocked the torrent bed by a wall of large grey tufa blocks, built across the upper end of the tunnel, thus holding back the loose soil above, with the result that the present arable area is at a height of several feet above the level of the valley-bed as it was known to the Romans. 78 TUNNEL The tunnel is 15 yards in length and 12 feet high bj' 11 feet wide, and therefore of much larger dimensions than would ever have been needed merely to allow storm waters to drain away to the sea : it is our belief that through it passed the old road from the Grotta di Sejano down to the harbour. The interior shows signs of having been plastered. A straight joint, crossing about 5 yards from the upper end, proves that the tunnel was at first only 5 yards long, but was subsequently lengthened by 10 yards. It is all built of concrete faced with opus reticulatum ; but whereas the' angles of the more ancient portion of the tunnel, t, were formed in tufa lateritium, worked in three 4-inch course ashlar, as may be seen to the north of the straight joint, the angles of the later work, t', as in the case of so man}' other buildings on this property, are formed in brick lateritium. On the roof of the tunnel is a pavement of crushed brick, which is now buried 3 feet below the vines growing above. The arch over the exit of the tunnel is faced with red bricks, the sides being faced with alternate courses of brick and tufa lateritium : two courses of brick to one of tufa (fig. 44). From below upwards there are six courses of brick, then for 5 ft. 2 in. five courses, each consisting of one tufa block and two bricks, then six courses of brick ; and then the spring of the arch, which rises about 6 feet in the middle. The Tunnel Vaults. On emerging from the tunnel the roadway was flanked bj' at least four large and deep vaults, of which two are still in a fair state of preservation. On the east an entrance of tufa and brick lateritium, of the same date as the exit from the tunnel, leads into vault a, 32 feet deep b}' 19 feet Avide, the most interesting constructional feature of which is the method of the intersection of the four somewhat flat barrel-vaults of the roof. The walls are reticulate, with a course of brick at the spring of the arch. A wide aperture opens into an adjoining and similar chamber b to the south, but this is so buried in earth, washed in b}'' the torrent when its bed was higher than at present, that the exact dimensions have not been determined. There is no indication of the purpose for which these vaults were built, but failing any other suggestion, and judging from their position on the roadwa}', they may have been used as stables or cart-houses, 8o TUNNEL perhaps as racdac rcccptacula, or at any rate as substructions for buildings above. On the west side two other chambers, c and d, are covered b}' simple barrel-vaults, which rise nearly 2 feet higher than those on the east side. The vault of c is perforated by a square man-hole or sky- light c that led up through a floor of opus signinum above, as is shown as if it were above d in the section in the inset in fig. 45. This floor extended over the neighbouring vaults and was approximately on the level of No. 3 tier of substructions (p. 83). These cave-like vaults on the west are partly excavated in the solid rock, partly built up with concrete and faced with opus reticulatum. The chamber d to the south appears to have extended further into the hill in its original state, but its length has been shortened by a massive wall, built across the vault, so as to divide off a small low end-chamber, d'. Chamber d' is partly hollowed out in the solid rock, partly roofed over by a roof of red tiles resting in lean-to fashion against the front face of a reticulated and plastered wall. In the rock at the back is a small oblong recess with three niches, one rectangular, one semicircular, and all plastered. It appears likeh' that this was a burial chamber partitioned off at the back of the large vault, but at the same time the thickness of the transverse wall would indicate an intention of supporting the added weight of later buildings above. All chambers on the level of the tunnel lay partly under a floor of opus signinum. Above this and under the western hill-side, two low vaulted chambers e and f of an upper store}' deserve further excavation. The crown of the vault over e is about 3 feet thick and is perforated by a rectangular ' skylight ' e, like the one over c. The purpose for which the tunnel was built would appear to have been to secure an adequate bridge of communication between buildings on the opposite hill-sides and thus to avoid the inconvenience, as well as the publicity, of a descent to a public thoroughfare below. The original bridge of the ' tufa period ' was only 5 yards wide : the later one was 15 yards wide. The large lateral chambers undoubtedly belong to the same period as that of the extension of the tunnel. A tomb, of which there is a further description on p. 197, was unearthed not far from the upper or north end of the tunnel. It was found very deep below the present arable surface, a further indication of the rise which has taken place all the way along the bed of the valley. From the tunnel the roadway descended with uniform gradient TUNNEL 8i to the harbour. On either side may be seen vaulted chambers and retaining walls to support steep and friable hill-sides, which could not otherwise have safely carried the great masses of building above without fear of landslip. At one place a subterranean passage leads right through the hill down into one of the quarry caves of Trentaremi.' Its exploration is not difficult, but should not be attempted in one's best clothes. ' The western exit has been indicated in the plan of the Caves in Earth Movements, pi. 48, and the passage is shown in the large map at the end of this volume. THE EASTERN SUBSTRUCTIONS AND TERRACES ' There are vaulted substructions everywhere along the western slope of the Gaiola valle}-. Thick retaining walls and vaults of concrete springing from walls faced with opus reticulatum extend for man}" j'ards, and even in their present ruined condition strongly uphold the hill-side. When complete the effect of this pile of building must have been imposing. It must have presented an appearance very like that of certain frescoes of large edifices built around and beneath with terraces rising step-like one above the other in a graduated series (cf. Arcliacologia, Iviii, p. 6, fig. i). And here we find no less than five or six such successive terraces carried on substructions, the massiveness of which reminds one of the concrete work of modern civil engineers. Of the lowest building of all we have no detailed information, but I have been informed that walls and pavements are buried some 12 feet deep below the surface of a garden. The accumulation of soil here is one of the consequences of the subsidence of the land, of which adjacent ruins buried in the sea-beach afford the clearest evidence. Owing to the rising of the beach, soil washed down the valle}' has not been able to reach the sea, and has accumulated in the valley bed. We may therefore look upon the walls at the garden level (fig. 46), in which the glass mosaic was found, as not belonging to the lowest or 'basement' floor (I), but as belonging to the 'first floor' (II) of the whole series of constructions. There is evidence that in front of this wall there existed a row of arched vaults carrying the terrace of the second floor (IIIi at a height of 19 feet above the first, but of these vaults onl}- parts of two piers remain, though outlined on the face of the back wall are the scars of arches and of several other piers. Six yards bej-ond the first pier and within a vaulted chamber that was next but one to it, is a rectangular niche containing a beautiful glass mosaic. The walls of the chamber ' Reprinted, in part, from Gunther, 'A Mural Glass Mosaic from the Imperial Roman \'illa near Naples,' Archaeologia, Ixiii, 1912. SUBSTRUCTIONS AND TERRACES 83 were plastered and painted in the manner to be presently described. Some 10 yards further north at the present ground level, and choked with rubbish, are three other niches of varying heights (3 ft. 6 in., 2 ft., 2 ft. 6 in.), but without decoration. Then comes the remaining pier, and at 1 1 j-ards bej'ond that, another small niche, and 6 yards further a modern retaining wall which holds back arable soil in the valley bed (fig. 46). The arches on this tier (II) were about 13 feet in height. They supported a terrace which has been raised to form a garden walk. On this terrace (III) and standing back is a second tier of vaulted substructions, of which the two southernmost chambers (fig. 47, A3, b^) are in a fair state of pre- servation. They now form useful cellars under a small garden house ; their floor is about 6 feet below the modern garden path. The interior was fine-plastered over a reticulated face and painted a rich yellow with the usual panelling of red lines. The chambers are nearly square, 12ft. Sin. deep by 12 feet wide, and are shown with the other structures on the hill above them, in plan and in section in figs. 47 and 48. The black represents Roman work ; the spotted, modern building. Further north are four more barrel-vaulted chambers (c^-f.,), of which the original floor is buried 3 feet deep under a modern floor : the eastern wall has fallen, leaving them gaping open. The vaults, still fairl}' complete, show impressions of the boards that were used as centering to hold up the concrete until it had set. The top of each vault rises about 4 ft. 6 in. above the spring of the arch. There is a small niche in C3. For quite 40 yards beyond the garden house are similar constructions all belonging to the same tier; but the last two chambers have a slightly different orientation from the rest. The walls of one of them are coloured red, and are provided with two niches, one at the back, shallow and G 2 •3 J ^ < in O S I- <: be O z < •A a. a z s o K O lb O U) u o < w H a < tf) Z o H U a Di i» a 3 UJ s H X o 3 O Bi X H Z o in O Id •Ji CO 86 SUBSTRUCTIONS AND TERRACES arched, the other in the north wall, rectangular and with grooves for two shelves. Of the building upon terrace IV only half vaults of two of the chambers, d^ and E4, remain, but traces of others, B4 and C4, are not hard to recognize. A flight of modern narrow steps leads up to the substructions of tier V, and another up to what may be considered as those of a sixth tier, which is on a level with a well-appointed hot bath, at about 50 feet above the present sea-level, and which would therefore have been some 70 feet above the Roman sea-level. Here and there are large blocks of concrete, apparently the foundations of higher parts of the building, and on terrace V we noted a bit of wall faced with tufa lateritium at a height of 10 feet above terrace IV. Above all these substructions were the Upper Baths. The Room zi'ith the Glass Mosaic. All that is left of the mural decoration of this chamber shows it to have been of a simple type, at an}' rate so far as the dado was concerned. White or cream-coloured walls were divided into panels by black bands flanked by narrower lines of Pompeian red. A second inner contour, in red, surrounded the lower panels. In the western wall of the chamber is the small niche lined and framed with the mosaic. The niche had apparently been hollowed out in the wall subsequent to the plastering, and indeed to the linear frescoed decoration, for it cuts into the red lines which frame the second panel. The opening of the niche measures 17 inches high by 20 inches broad, and it is about 8 inches deep. The sides as well as the back are encrusted with the mosaic. The Glass Mosaic. A chromolithograph of the mosaic prepared from an enlarged photograph forms the frontispiece to this book. On the back, against a background of bright cobalt blue repre- senting the sk}', are three plants, a reed growing between two green plants with simple ovate or oval leaves, over which is hovering a white dove with tail spread and wings outstretched, head and feathers being < m O X X 1- w u o o in u « -I < PS 3 s o 88 GLASS MOSAIC cleverly indicated bj' a linear arrangement of tesserae. The leaves of the plants are treated in a very realistic manner, mosaic cubes of two shades of green being employed in the shading, and between them bits of deep blue-black glass have been skilfully introduced to show up the leaves in greater relief Below are one or two flowers, simply sketched— a central blue tessera surrounded by four yellows. The bottom of the little picture is bordered by a trellis, represented by yellow lines crossing X-wise on a blue ground. Round the whole run lines of white and blue tesserae and then one of cockle-shells. The nature of the narrow mosaic decoration of the sides of the niche is apparent from a piece, measuring 13 inches by 5 inches, still in situ on the left. About ten small crosses placed vertically and in alternate colours are inlaid in a background of black glass, now iridescent. Each cross is formed of five tesserae, the alternate crosses being formed by tesserae surrounding one blue b}' four j'ellows, and four whites surrounding one j'ellow. The most interesting feature of this side panel is, however, its inlaid rectangular border of spirally twisted glass, black with a thread of yellow, pieces of which were found m situ. A cockle-shell border surrounds this, and round the inner angles of the niche the rope-like glass ornament is repeated. The shelf of the niche appears to have been coated with plain hard plaster. The niche was framed in a mosaic border mlaid in the plaster of the wall and surrounded bj' a moulded marble frame in relief The pattern of the mosaic border was formed by interlacing bands of yellow mosaics on a background of cobalt blue, and flanked on either side bj- three parallel lines, each a tessera wide, in order from within outwards, j-ellow, green, blue. The outer marble frame was fixed in the usual way hy iron holdfasts, rusted traces of some of which still remain. Its size over all may be estimated from a scar left on the plaster, at about 2 ft. 5 in. wide and 3 feet in height. In an attempt to ascertain the age of this little glass picture, at least three factors maj^ profitabl}- be taken into consideration. First, the glass mosaics. We find essentially similar ones upon the handsome encrusted columns from a house in the Street of Tombs at Pompeii, now in the Naples Museum, and in several fountain-niches of the same period in Houses, nos. 22 and 23 in Regio VI, Ins. viii. Secondly, the cockle-shells, which are employed in exactly the same manner and affixed by the same reddish-brown cement as in the ornamentation of the above-mentioned fountain-niche. GLASS MOSAIC 89 Thirdl}-, the twisted glass rods, which, so far as we have been able to ascertain, have not hitherto been described as forming part of the decoration of a mural mosaic, although similar detached sticks of twisted glass have not unfrequently been found on other sites ; and there is a fragment of plaster measuring 4 inches by 2 inches in the Capitoline Museum with three glass rods embedded. My friend Sir Hercules Read has drawn my attention to similar spiral rods in the national collection in the British Museum, which have one end thickened and flattened, very like the implements that were formed}' in general use in this country for crushing sugar in a glass of warm toddy. In style as well as in the materials with which it is decorated, this mosaic reminds us of those of the charming fountain-niches of which some half-dozen have been found in Pompeii. In several of them we see the cockle-shell border and birds flj'ing over plants ; cf. Presuhn's illustration of a mosaic found in Reg. VI, Ins. xiv, no. 43. And the same details may be seen in a fountain-niche found at Baiae, and now deposited in the South Kensington Museum. In the absence of anj' evidence to the contrar}', it is not unreason- able to conclude that the period to which the construction of this glass mosaic is to be referred, cannot have been very long before or long after a.d. 79. We see no good reason for believing the niche to have been intentionally decorated with the Christian emblem of the Dove, although that idea was the one which first occurred to us. But in spite of the great resemblance of this little mosaic to the fountain pictures, it is not impossible but that it maj' have formed the background of a small lararium or shrine sacred to a deit}', in which case a religious significance may be attributed to it. It will be remembered that the dove was very closely connected with the founding of the first Greek colony upon this site, and that that fact was never forgotten in ancient Naples. The ships of the original Chalcidian colony had been piloted from Euboea to the Campanian shore by a dove, and in pious memory of that event, the cult image of Apollo in the Temple of Apollo at Naples is said to have had a 'columba ' perched upon his left shoulder. 'Ipse Dionaea monstravit Apollo columba.' Stat. Sih'. iii. 5. 80. 'Huius classis cursum esse directum alii columbae antecedentis volatu ferunt.' Velleius, i. 4. 'Tu ductor populi longe migrantis Apollo Cuius adhuc volucrem laeva cervice sedentem Respiciens blande felix Eumelis adorat.' Stat. Sifv. iv. 8. 47-9. ■a a o a o < H a H o -J o lO THE GAIOLA BEACH AND THE LOWER BATHS The rough mule-track that leads down the Gaiola valley follows the Roman Valley Road that went down to the ancient harbour from the tunnel, but the modern track lies several feet above the level of the old road. As it nears the sea the path descends in much-worn steps cut in the rock, and after passing outside a modern retaining wall traverses some massive ruins that rise from the little beach; rough, uncomfortable steps worn in these ruins lead down to the sand. The rums are for the most part those of baths which stood on the right-hand or western side of the aid road ; and the old walls pierced by the tunnel-like entrance to Mr. Foley's house clearly belonged to the same structure. In any intelligent examination of the site, it must alwaj's be remembered that at the time when the buildings were erected, the land stood some 20 feet higher above the sea than it does now, and that floors over which the sea now washes, had a whole ground- floor suite of apartments under them, but these submerged chambers are now filled with sand and water and have not been seen by any living man. If, then, we describe the various parts of the buildings on the Gaiola Beach according to the storey to which they belonged, the uppermost now standing, although now not ver}' high above the sea, formed the third storey in Roman times. Of the third storey we have a mass of building near the entrance to the Proprieta Foley. Just within a modern gate is a narrow chamber, fig. 51, A, floored with opus signinum, with reticulate walls plastered and painted with yellow stripes and pattern. An oblong niche a, 4^ feet in length and raised 3 feet from the floor, was built in the west wall, but it has been partly filled up b}^ a more recent wall of reticulated work. The present pathway, 18 inches lower, cuts through the signinum floor, a continuation of which, b, fig. 51, maj' be seen outside the modern boundary wall. This pavement is about 22 feet above sea-level. J ^ i«^!^^y^^ '^' 0. ^^ 'y> *f ^ ■^z^. = ^-^z^Z ^ ^-- ^ '% ~ f?^-^ >"-^ Fig. 51. Plan of Walls on the Gaiola Beach. Heated rooms D- C are shown on a larger scale in fig. 52. LOWER BATHS 93 To the south of the modern gateway is more wall-surface, 6 feet high, reticulated, plastered, coloured yellow and decorated with a red border i^ inches wide ; and next it, is a similar blue-painted wall. These two walls were the western ends of two small chambers, a yellow room b and a blue room c, which extended eastward above, and at the back of the hot rooms on the original second storey, and which, as may be seen on the plan, have a different orientation. The second store}' has been overlaid by a recent floor at a height of about 10 feet above the present sea-level, but its structure is partly revealed at the lower end of the small path which leads across the ruins down to the beach. Here certain small apartments are provided with so complete an investment of hollow floors and walls, that not only is the purpose of the building as a thermal establishment clearly indicated, but also that here must have been the warmest chambers of the baths. Unfortunately the site is traversed by a public footpath, which we could not remove without getting into trouble. Nevertheless, by scraping away here and there we were able to make out several details clearly. One chamber, e, fig. 52, measuring 9 feet by io| feet, was fitted with hollow walls and with a suspensura floor, carried partly on oblong stacks of bricks, c , e", 3 feet in height. The construction of the west wall is shown in elevation at e. Twenty courses of brick, rising to the same height as the two piers e', e", form a fire-proof facing to the hot-air chamber below the suspensura; thereon are laid nine courses of tufa lateritium (2 ft. 8 in.), in which are several iron holdfasts; then four or five courses of brick lateritium, above which, for a height of 3-4 yards, the wall is faced with reticulate work banded with courses of brick, in which more iron holdfasts afford an indication of an inner lining either of flue tiles or of marble veneer. The suspensura floor of this hot chamber is 5 feet, and the under floor only 2 feet, above the present sea-level. On the north side was a second, apparently similar though smaller, chamber d, also provided with a complete lining of rectangular flue tiles, in section 6 mches by 4 inches, under plaster. There are flue apertures near the south-east corner of e (marked ' passage ' in the plani and through the east wall of d. To the south was a lower vaulted chamber r, with two plastered niches in the west wall, but the south wall has fallen, so that the chamber is open to the wash of the sea in rough weather. The eastern side has been hidden under the concrete mass of a new breakwater. To the east was a long apartment g, 8 feet wide, of which the Fig. 52. Plan of the Lower Baths. LOWER BATHS 95 east wall is now level with and parti}- hidden under the sand of the beach. The vestiges of a suspensura and of hollow walls still remain at the north-west end, but as the floor is 3 feet below the floor of the other rooms e and f, and is onh' 2 feet above sea-level it is surprising that any of it should have survived destruction. The suspensura was not very thick. The pilae were formed of long tubular tiles, like those of the upper baths, but with squarer ends, measuring- 6 inches across. They were further heightened by one Fig. 53. Suspensura in the Lower Baths. small square tile laid upon each, and on these were laid two courses of large tiles, each about i;J inches thick by 18 inches square, which bridged the intervals between pile and pile; the three courses of tiles occupying 6 inches. Upon this 4 inches of opus signinum finished the hot floor, as shown in elevation in the inset at the top of fig- 52 and at h' in figs. 50 and 53. TJ-ie whole construction was very similar to that of the hypocaust and hollow walls of the Baths in the Thermae of Caracalla as described by iVIiddleton, whose drawing, by the permission of Mr. John Murray, we reproduce on p. 97. In the wider northern part of room g the wall is partly reticulate, partly brick, the higher parts standing being overhung with 96 LOWER BATHS a luxuriant growth of Mesembryanthemum, but the eastern wall shows tufa lateritium work in places : in the west wall is a shallow rounded niche, ^^, 2 ft. 8 in. in width. A doorwaj^ led northwards to an apartment with iron fixings plugged with marble chips in the walls, from which we may infer a former lining of marble slabs. Another wall K, fig. 51, faced with reticulate tufa, lies at a distance of 21 feet to the north. It is traversed by a horizontal band of six courses of brick lateritium, and is channelled with a vertical groove 3 feet in length, evidently for a water-pipe. This part of the building is obviously of the same period as the back wall of chambers d and e. At the head of the beach a modern retaining wall, 12 feet high, has been built across some ancient and more massive constructions of concrete, l, m, n. A fine thicket of canes flourishes above it. In Roman times, we imagine, there was no cross wall, but the way would have been open, giving space for a 12-foot roadway between the piers. The purpose of the piers was, we believe, to support a large arch spanning the roadway and a narrower arch beside it over the foot-way M, the whole forming an entrance, perhaps with sculptured reliefs and inscriptions worthy of the owner of the Villa. No better place for such a monumental gateway could be found than this, near the head of the harbour and at the lower end of the roadway. On the western side of L is a wall faced with alternating courses of tufa and brick. On the east side a vaulted chamber n measures 11 feet by 9 ft. 8 in. Beyond this again, on the beach at the foot of a stairway leading up to the squalid quarters of the Casa Marotta, and separated from the lower baths by a space of 6 yards, another building, fig. 51, p, Q, R, s, the purpose of which is not obvious. Firstly there is a reticulated wall under the steps, which appears to have been part of the same structure as a solid brick pier, o, which is just showing above the beach. Owing to the subsidence of the land, the top of the concrete floor of p, the principal room now visible, is barely 2 feet above the sea, and the strong walls have been worn down level with the sand in which they are buried. The pavement was of white marble mosaic with a few black tesserae spotted about at regular intervals. Near the eastern end of the room is the base of a column, 2 feet in diameter, built of fragments of brick and plastered. The north wall is reticulated, but a vestige of tufa lateritium may be seen near the angle inside a small modern enclosure indicated b}^ dotted lines in the plan. There is good reason to believe that another chamber of similar dimensions exists beneath this floor, but that owing to subsidence it is now below sea-level and filled with sand. The adjacent apartment, LOWER BATHS 97 Q, is also flooded, so that the sea is now actual!}- washing against the plaster at the top of the back wall and against what remains of the springing of the vaulted ceiling, q, in the north-east corner of the room. Few instances of the effects of land-movements are calculated to make a more vivid impression than the sunken ceiling of this room, engulfed like a wrecked ship in the Gaiola beach. Vestiges of plaster remain on the walls of the adjoining rooms, r and s. The latter, measuring 13 feet by 11 feet, is entirely under water. The e.xcavation of these buried submarine chambers would be a matter of great interest. Who can tell what works of art might not be discovered? A final answer might also be found to a question which has often been asked. Was the subsidence a sudden one, or was it gradual enough for the buildings to be dismantled b}- their owners? But for such an excavation it would be necessary to build a coffer-dam and to pump out sand and water. For the present any treasures these submerged rooms maj" contain are nearl}' as safe as ingots in a sunken treasure-ship. On the beach rolled fragments of granite, of Car^'stian and Porta Santa marbles, together with small bits of red porph3-ry and green serpentine, bear witness to the departed luxury of the baths and to the wealth of the builders. Fig. 54. Wktuud of hk.ming hie B.vths ix the Theu.m.vi; of C.vkacalla. After Middleton from Smith's Did. of Antiquities. AA. Concrete wall faced with brick. B. Lower part of wall with no brick facing. CC. SiispciinKrij, or upper floor of hypocaust, supported by pillars. DD. Another floor, with support only at the edges. EE. Marble flooring. FF. Marble plinth and wall lining. GG. Under-tloor of hypocaust, paved with large files. HH. Horizontal and vertical sections of the flue tiles, which line the walls of the Caldarium. an. Iron-holdfasts. J J. Socket-jointed flue-pipe of Tepidarium. K. Rain-water pipe. LL. Vaults of crypt, made of pumice-stone concrete. H o > u < a. o < THE UPPER BATHS OR BATHS OF HADRIAN In Neapolis, there were hot baths near the theatre, hnpatiens secreh a balneis in theatrum fransiit (Suetonius, Nero, 20). At Pausilj^pon it would have been scarce four minutes' walk from the one to the other, and the distance to the principal dwelling-house which we have named the House of Pollio was even less. Plans of the baths are given in figs. 57 and 60, but the buildings are not by anj^ means all of the same age. Both the furnaces and the caldarium present features of exceptional interest. Neither formed part of the original building, nor, for the matter of that, of the later one. They are survivals of a transition stage of unknown duration which have been built over during structural alterations in the baths, thereby being preserved to us. Our experience of the instability even of modern systems of heating, of hot-water, warming, and ventilating appliances helps us to realize that improvements were constantly being demanded by wealthy patrons of the architects and plumbers of Roman baths. At Silchester in our own country we have an excellent instance of this. Mr. St. John Hope has been able to trace no less than six distinct periods in the history of the baths, each marked by some alteration or improvement : and there was hardly a bathing establishment of any repute between Calleva and Pompeii which does not show signs of repeated readjustments to the tastes and new inventions of successive generations. In the present case we were fortunate enough to find irrefutable evidence that alterations in the Upper Baths of Pausilypon were made at the time of Hadrian and probably by his orders. The rooms in this block of building were disposed in at least three storeys, but successive owners of the property have thrown down such large portions of them that now it is difficult to arrive at a clear comprehension of the plan as a whole. The remnants which are still preserved to us with any approach to completeness are of excep- H 2 loo THE UPPER BATHS tional interest, and the circular hot bath basin especially so. The rooms of the middle storey were used as bath-rooms : the purposes to which the rooms of the lower and upper storeys were put are unknown. A general view of the ruins is shown in figs. 55 and 56. No trace of any stairwaj- has been found, so that it may have fallen with the cliff-side, or it may be that there was no near waj' of getting from the rooms on the lower floor to the baths above. We will describe the lower rooms first. The Rooms on the Loiver Storey. {Storey V.) Near the end of a modern wall built to shelter a garden behind Mr. Foley's house (=Casa Bechi, fig. 56) from the blighting effects of the salt-laden Scirocco, and close to where it abuts against the older walls under the hill-side, we find a remnant of building, shown in ground-plan in fig. 57, which is interesting enough to make one wish that a larger portion of it had been preserAJ^ed. Here in the corner of room a are the remains of a small rectangular tank or basin, /, formed in white cement, measuring 27^ inches by 22 J inches, and 21^ inches deep, and two holes for the suppl}- and overflow pipes are to be seen in the walls. When excavated, the overflow, in the middle of the longer side of the basin, was found to be choked by two pieces of pumice-stone, which had evidently been in use, for the)' were rubbed flat on one side, and we wondered whether the}' had eifectually blocked tlie drain, thereb}' causing a flood, and whether the careless domestic, from whose hands the}' had slipped, had entirel}- disclaimed all blame for the stoppage. The pavement of the room upon which this basin was built, was of opus signinum and rested upon the untouched volcanic soil of the hill. Below the plastered basin, but quite near it, there existed until recentl}' a rather remarkable structure, s, removed by Mr. Foley to make way for a new pathway below. It was a vertical tubular shaft, lined internall}' with smooth white cement and extending downwards for some 15 feet below the signinum floor. Near the bottom, the shaft became square in section, with an opening towards the east. Its relations to the channel c shows it to have served as a water-conduit or drain. Near the basin was unearthed a piece of lead pipe which was a find of the greatest consequence, for it has enabled us to give a name and date to this portion of the Baths. It bore the name of Hadrian and will ^ wi^ (n o < 2; [02 THE UPPER BATHS be described in detail in the section dealing with the Water Supply, p. 129. Close to the little basin a doorway opens into a passage, b, also floored with opus signinum, beneath which ran obliquely a water- channel, cc, formerly opening into the vertical shaft alread}' described. The passage, 6 feet in width, was plastered, the walls showing a pattern of red lines on a white ground. Fragments of other plaster with j-ellow and brown floral decoration (p. 256) and bricks with pentagon stamps (p. 216) were found here. Turning to the left we found a doorwaj' with a threshold of two blocks of lava, //, rebated and recessed in the middle for the doors and showing three small round cavities at the ends, the pivot cavities of successive doors hung in different positions. The inner corners of the walls of the room c are strengthened with brick lateritium. This room has not yet been excavated, but a reticulated wall, c', some 12 feet distant and about 8 feet high, probabh' indicates its western limit. The floor is of opus signinum. Among the debris were iound pieces of rectangular flue-tile which were more than 8 inches long when whole, and measuring 5^ inches across the wider side. The sides are thin, being scant jj in. thick ; they are well baked and ring when struck. From these we ma}' infer the former existence, on this storej' or on the store}' above, of a room heated by hollow walls constructed of similar tiles ; while numerous fragments of veneer of cipolline and Porta Santa marbles may indicate the nature of the mternal decorations. At the back of these constructions at n and above the garden behind the house, the walls shown in fig. 58 project from the hill-side. The rectangular chamber now partly occupied by a semicircular garden-house, was once covered by a floor crossed by a small channel of rectangular cross-section. A section across the adjacent walls is shown as an inset at the top of the figure. Further west a range of building with a different orientation is crossed obliquely by the reticulate wall c' already mentioned. A large vault E, 10 feet wide by 12 feet to the spring of the vault, or 16 feet to the crown of the vault, extends for near 30 feet in a north and south direction. The vault is perforated by a round hole e, plastered and 16 inches in diameter. Near the middle of the east side a doorway, 6 feet high, 4 ft. 3 in. wide, spanned by a horizontal lintel, leads into a small oblique passage-chamber d, now blocked, but which appears formerly to have led further in two directions ; it probably communicated with room c and so with the passage OR BATHS OF HADRIAN 103 behind the white cement basin already described, and it may have led north to a stairway to the floor above. The ceiling of d is flat and about 8 feet above the floor. The interior of the large vault was plastered and marked with a dado line 5 ft. 3 in. from the floor. Near the south-east corner at 28 inches above the floor a groove in the plaster denotes a former shelf or long bracket. On the west two other chambers, 16 feet by 11 feet, are situated side by side. The inner one, g, is still perfect, but the greater part of the other, f, to the south has fallen away ; a small fragment of wall shows it to have been ceiled with a barrel-shaped vault with Fig. s8. Plan of Walls, to the East of the Upper Baths. the springing from north to south. The floor rests directly on the undisturbed pozzolana. The decoration (see inset to fig. 57) consisted of a high red dado up to 7 ft. 4 in. above the floor with long oblong pictures, at intervals which have been carefullj' e.xcised by the hand of the spoiler. The upper part of the wall was divided from the dado by an incised line and was coloured cream. The fact that the whole of this building is of tufa and that the pier between the two vaulted chambers is constructed of lateritium tufa is an indication that the structure is to be referred to the earlier architectural period of the Villa. The orientation of room e ' is due north and south ; that of the I04 THE UPPER BATHS later brick and reticulated building to the east being turned about 45' to the east. If we were called on to advance an}' theory with regard to their respective ages we should sa}' that c might be Hadrianic but that E, F, G, are much older. The inner vaulted chamber g was also decorated with a reddish dado, the walls being white above. The crown of the vault was also about 16 feet above the floor, but has now flattened to a dangerous e.xtent, and is threatening to fall. There may be a door into this chamber from the cross chamber e, but as the removal of the debris with which both are encumbered could not be undertaken without danger, we did not verifx' this point. The adjoining rooms on this storey, V, further west were demolished during the quarrying operations of the Marchese del Tufo, by which the hill-side, which had been level with the Vineyard Site, was reduced to the condition sliown in figs. 55 and 56. Upon the floor above are the most interesting, perhaps because best preserved, parts of the baths. 1 ■ M i^ , 1 r^ t^*T^^^k lifl I ■ ^B bA Wk llibetti ^ti J --«*•- P 1 ^^~H ^fl^^B . ■*. -^ -^■CMMH ^ ^ JBH Fig. 59. The Corridor with the Marble Pavement, J. OR BATHS OF HADRIAN 105 THE MIDDLE OR BATH STOREY [No. VI.) The Corridor zvith the Marble Pavement. Behind the Hne of the rooms e, f, g of the lower storey and one storey higher, is a long corridor, j, with a marble floor. It was formerly covered by a barrel-vault which carried a floor of opus signinum belonging to the third or upper storey (VII), fragments of which may be seen on the fallen blocks of concrete, jj. Owing, however, to the removal of pozzolana from beneath the foundations, the south wall, excepting near the two ends, has entirely collapsed, and the floor of mixed coloured marbles extends for 19^ yards along the face of the cliff like a terrace, 7 ft. 2 in. wide, without a parapet. One entrance to the corridor was from the east, but other doorways no doubt opened from it into rooms to the south, which have fallen with the cliff beneath them. Another doorway, d, 4 ft. 10 in. wide, opened through the north wall between jambs of tufa lateritium, laid in 3|-inch courses ; but this had already been blocked up in Roman times, probably when the marble stairway, s, to the caldarium was cut through a wall with a previously continuous facing of opus reticulatum. The west end is terminated obliquely by a wall of masonry, a, of a type unusual in the Pausilypon Region, namely, of large blocks of tufa, in 8-inch courses, and corresponding in orientation with an older vaulted building v, whose walls protrude 6 feet below the marble pavement. Two features in this corridor are of especial interest. The floor has been constructed to throw off water easily, indeed it slopes with so steep a gradient that the east end is 2 ft. 10 in. lower than the west end ; and secondly, it had been crossed by two water-channels or drains fig. 59, x^, x-, each 16 inches wide, and plastered, the bottom of each channel being 2 feet below the top of the pavement. But the interior of the channels had already been filled with rubble masonry before the laying of the pavement, they must therefore have belonged to some earlier construction. A connecting channel runs close under the north wall of the corridor ; two manholes, bb', quite close together, which open into it, are closed by large bricks ; one of them, b', cannot have been used since the building of the stairway, for its cover is tighdy fixed under the stone and mortar of the lower step. This manhole was probably the older of the THE UPPER BATHS 107 two, the building of the stairway necessitating the construction of the second opening, b, within 5 feet of the first. Within the channel, which had apparently escaped the notice of previous excavators we found a small bronze hinge (p. 291), fragments of glass and pottery (pp. 283, 289), and pieces of fine white painted plaster (No. 7, p. 256), indicating that the earlier colour scheme of this room was green and yellow, and finally a base denarius of Claudius Gothicus, showing that the channel had remained in use until after a. d. 268. The floor over the long channel is covered with plain cement, which now appears as a border, 32 inches wide, to the polished floor of marble concrete. It is possible that this plain border was covered by an edging of flat marble slabs. The walls still show four coats of plaster with traces of colour here and there, the decoration being rather that of a frigidarium than that usual in the warmer portions of baths. The dado was cream, with a black and blue trellis, in which was intertwined blue, grey, and slate-coloured foliage. Higher up were floral wreaths, blue festoons with red flowers ; and a brown and yellow band bearing clustered flowers, in order blue, yellow, black, and blue (fig. 157, c). Another fragment of plaster was painted with red and yellow flowers among green leaves. The earlier style of decoration, as indicated by the fragments found in the flue, seems to have been much lighter in effect. The plaster was of whiter and finer qualit}^ and was laid in a thinner coat. In the decoration yellow flowers among olive-green foliage, painted in a somewhat conventional manner, predominated (fig. 157, g). Near the eastern end of the corridor there is a small trapezoidal mark / on the floor, measuring 28 by 16 inches, with a small circular pit in one angle. It is evidentl}' the impression of an object which formed a permanent part of the furniture of the corridor, since the marble pavement is not continued over it. The pit is not unlike the hole for a door hinge in a threshold stone. Eastward the marble-paved corridor opened into a wider gallery, 115 feet in width, which ran east and west at the back of the high- vaulted rooms G and e of the lower storey V. Immediately behind this gallery are some massive blocks of masonry, more than a yard thick, and faced with reticulate tufa, and 5 yards beyond is the furnace-room m. io8 THE UPPER BATHS 7n ? t. ■m//if/mta///mm ^\ ':'((iim/mm,'iUiMi.:i.i/.'n The S/ainuay. s. A flight of six marble steps led up from the corridor to the caldarium k. Five of the steps, 4 ft. 7 in. in width, have been somewhat rudcl}' hewn out of the thickness of the wall, showing the whole stairway to be a later addition. The lowest step projects into the corridor, covering the opening b' of the water-channel already mentioned. All the steps were covered with marble slabs, bedded on plaster, spread i^ inch thick, on a foundation of large bricks. The slabs have been stripped long since, but their posi- tions are indicated b}- marks on the cement and by frag- ments of marble, bedded here and there in the cement bj- the masons engaged in levelling the slabs. The treads of the steps were 15 inches wide ; the risers being 8 inches high, and made of marble about i inch thick. The method of construc- tion is shown in fig. 61, in which in = marble, c = cement, and / =-- brick. 10 M ,♦* SO C^ Fig. 61. Sectional Elevation of Steps. The Old Caldarium. K, fig. 60. The steps led into a round chamber of about 21 feet in diameter which we will call the Old Caldarium, because there is evidence to show that it was used as such during the earlier but not during the later period of the Baths. The surrounding walls, which do not appear to be all of the same age, are parti}' curved, partly straight. On the east is a semicircular apse, with a small niche k on the north side ; the western wall, of opus incertum, runs straight with an orientation of about 118°. When we first examined the site, there was evidence that an ancient continuous floor of concrete had extended right across the chamber. Close examination of the surface of this floor showed that there was a circular crack y, in the floor, 15 feet in diameter, and at about 3 feet distant from the walls ; and excavation revealed a very remarkable heated bath of an hitherto undescribed type. OR BATHS OF HADRIAN 109 The earlier Roman heated bath-basins were built upon the suspensurae or hollow floors of heated rooms, as may be seen in the men's caldarium of the Forum Baths at Pompeii. There and in man}'- similar cases a low wall was built across the h3'pocaust floor, so as to enclose a tank which was made waterproof with cement and marble veneer. In some cases the side and end walls of the tank contained hollow flue-bricks, thus materially increasing the heated surface in contact with the water. The plan was soon found to be uneconomical and not very efficient. The great thickness of tile, cement, and marble necessary to make the structure watertight, often not less than one foot between the furnace flues and the water, as well as the very bad conductivity of these materials, must have made the heating of a considerable volume of water an extremely slow process. In all probabilit}' it was the usual practice to heat the water in special ' coppers ' before running it into such a warm bath-basin. Alvci or bath-basins constructed at a later period were generall}^ provided with a testudo alvei} an invention consisting of a bronze circulating boiler built into the furnace flue, warmed directly by the hot air, and opening b}' a wide mouth near the bottom of the alveus. In high-class bathing establishments the cold-water bath in the frigidarium was frequently' of circular shape, and being sunk below the level of the floor, was provided with concentric steps for the convenience of bathers. In the structure of the Pausilj-pon bath we have a unique solution of a ver}' neat problem in construction. It was an attempt to construct a circular bath with steps so that it could be heated by a complete investment of hot-air flues, running immediately beneath each one of the steps and under the bottom of the bath. It reveals in its construction a higher degree of skill in design and bricklaying than has yet been met with in similar structures ; and as such it is to be regarded as a unique specimen of the heated alvetts, as the last word in balncae pensilcs, the valued invention of Orata. So unlike was the construction to that of any of the other bath- basins with which we are familiar, that this explanation of its real purpose did not occur to an}' of the persons to whom the antiquity was shown. Among theories put forward one was that it had been a columbarium, another, an oven, 3'et another, a warm hollow in the floor of a sudatorium, but the finding of an overflow pipe showed that the basin had been intended to hold water. ' Roman Testudo Baths and modern Tortoise Stoves received their names tor different reasons ! II ■J) u d H 6 THE UPPER BATHS III Strttcinre of the Warm Bath-Basin. The circular basin, surrounded by three concentric steps upon which the bathers rechned, occupied the centre of the floor of the chamber. The basin itself measures 15 feet in diameter. Hot air circulated from a praefurnium, 4 yards distant to the east, behind the risers and under the treads of the steps, as well as under the floor, Fig. 63. Frigidarium of the Old Baths at Pompeh. After Gell, from Smith's Did. of Antiquities. SO that the water got well warmed all round. An overflow pipe d, about 3 feet above the bottom of the basin, was found in the north- east side ; and the finding of this conclusivel}' proved the truth that the basin was intended to hold water and had not been used merely as a hot-air bath. The steps were formed in smooth cement of good quality, but the bottom had been overlaid with white marble, of which we found a few fragments. The bottom of the basin was a suspensura, carried by pilae, each formed of a single tubular brick (fig. 151) 17J inches high. The builder must have had a lot of trouble to get these suspensura floors water- tight. In this case the suspensura was about one foot in thickness, and was formed of two or three separate layers of a concrete of broken tufa spread upon the large tiles, tcgitlac bipedales, which spanned the distance from pila to pila. And over the tufa concrete was spread a layer of crushed brick concrete, and finally' the layer of 'n=^ m W^ I #^1 % m ^■>- 2 ^ o z m it ^~^-^s %« Em ,•.;,^^:.v>:.x^V.^.^^^>>^>^X<;XX^^ o u > o Q Z < z X i- ca Q u H Id s H b. O u a Ji Id X (- b. O Z o p 6 iiili i?i^ i ■''X (^ U< 2: « a- i S ca O (0 ^ Si t/} ^ THE UPPER BATHS 113 cement on which the marble was bedded. The hollow walls were formed of large bricks held in position by iron holdfasts fixed with plugs of marble : a method of hollow-wall formation which was certainly earlier than the use of the rectangular flue-bricks employed in the Lower Baths. The debris of a domed vault lie on the floor. The bath was heated by a simple furnace, hypocausis, on the east side, from which a flue 5 feet high, by 2 feet wide and 10 feet long led hot gases to a large vaulted flue, 2, embracing half the basin. From this the hot gases passed through six intake openings with brick arches, into the space 3 under the hollow floor, whence it appears to have been drawn into a second large flue, 4, embracing the other half of the basin, through six other apertures, five of which are square-headed and much narrower than the intake apertures on the other side. One, larger than the others, was no doubt used as a manhole by the builder. The two intake and exhaust flues form, as it were, two large arcs of the same circular tunnel, but are separated by two brick piers w, w, which are needlessly massive for the purpose of merely dividing the flues from one another, and obviously belonged to some older structure. The floor of the flues is sunk 18 inches below the floor upon which the hypocaust pile-bricks stand. A series of vertical flues conduct hot air from the basement flues to the spaces under the concentric steps, as shown on the left of fig. 65. The lower horizontal channels /\ fig. 64, are spanned with arches of brick, but the smaller upper horizontal channels /^ are covered by imbrices. It stands to reason that the lighting of the furnace fire would be very unlikely to start a draught through so complicated a system of air-spaces. In the case of the Pompeian baths, provision was made for the lighting of special fires near the upcast flues for the purpose of starting the draught in the desired direction, and so it is fairly certain that there must have been provision for such draught fires here too. There is evidence that the caldarium with its raised floor and the steps leading up to it from the passage were later additions to the establishment. The thick brick walls, w, in the middle of the circular bath certainly belong to an earlier arrangement of things. At a later date still, the marble investment was removed, the cavity filled with debris, and the space within the topmost step overlaid with the cement floor, the circular junction of which is still to be seen. It is owing to the concealment and protection afforded by this upper floor that the structural details of the bath below have been as well preserved as they are. I Fig. 66. Wall of Heated Bath. East Side. Fig. 67. Wail of Heated Bath. West Side. Siiowing the square-headed exhaust flues. BATHS OF HADRIAN 115 The Furnace. A second furnace, other than that used to heat the circular hot bath-basin was found in the adjacent chamber m. The early Pausi- lypon Baths, Hke those at Pompeii, were probably solelj' heated by brasiers, and were built without hollow-walls, flues, or hypocausts. But whether this was so or not, it is certain that this furnace formed no part of the original building, for it stands isolated in the middle of a small plastered room, m, paved with a white marble mosaic with a border of two black bands. (Mosaic, fig. J53, no. i.) The builders did not even trouble to take up the mosaic, but built the Fig. 68. Fuknace Room RI. furnace directl^^ upon it, thereby concealing all but a small part of the original floor space. The new praeiurnium f is raised 2 ft. 6 in. above the floor upon a concrete foundation ff (fig. 69) 7 feet square, and is carefully built in reticulate tufa (blocks 4 to 5 inches square), and with brick-shaped blocks of tufa at the angles (fig. 71). In ground- plan it is horse-shoe shaped : a brick arch is neatly turned over the stoke-hole, and the flue is circular. In the convex back of the furnace an ornamental horizontal row of si.x brick cubes, as shown in fig. 6g, is introduced across the diagonal rows of tufa blocks of the opus reticulatum. I 2 i Fig. 69. Conjectural Restoration of Building over THE New Furnace Room. Fig. 70. RuscoNr.s Picture of Roman Baths. From Sitii//i's Diet, of Attliquilies. BATHS OF HADRIAN 117 On either side of the furnace newer walls (Hne-shaded in fig. 69) were built on the mosaic pavement and against the plastered surface of the older walls (dotted in the figure) : thej' were arched against the furnace-building in the middle, so as to support an upper floor. The lower courses of the wall round this floor were built of brick lateritium. The chamber was circular in shape, and judging from the fact that the centre of the floor is pierced by the flue of the praefurnium, it may very likely have been a sudatorium with a pillared hypocaust heated by the flue, or a room for heating hot water in a 'copper' for the use of the bathers, as shown in our conjectural restoration (fig. 69I. Fig. 71. Stokehole of the New Furnacf. BUILDING TO THE NORTH OF THE UPPER BATHS Some 50 feet to the north (bearing no) of the hot bath-basin, surrounded by a raised bank of rubbish, are some well-built walls I'll.. 72. I'l-.x.N ui Bru.iJi.NG. enclosing a chamber measuring 10 feet by 18 feet (fig. 72). These walls extended further, but whether they were connected with the Upper Baths or with a mass of brickwork 25 feet off to the north-east, is not certain. The walls on two sides of this chamber arc faced with brick, but the other two sides are reticulated. Their orientation is 312". BUILDINGS TO THE WEST OF THE UPPER BATHS Ruins of several constructions are conspicuous on the slope of the hill to the west of the Upper Baths. In order, from east to west, we see first the large chamber v, on the Vth storey, with part of its barrel-vault still standing (fig. 55). Chamber ys. . '( Fig. 73. CiiAMBLn X. It is noteworthy on account of the orientation ot the walls being nearly at half a right angle to the direction of those of the Upper Baths. Internally the walls are plastered and the floor was of white mosaic. Next we noted a strongly built wall, of opus incertum, which appears to have supported the south-west angle of some building on the Vlth storey above chamber v. From it another wall extended southwards. Then at a distance of a few feet and six feet lower down is a most interesting little chamber, x, with a white mosaic floor and a low white plastered vault, shown in sectional elevation in fig. 73. The eastern wall is partly of brick. In this chamber we found the upper drum of a fluted column (no. 10, p. 221), the fragment of a full-sized sculpture I20 BUILDINGS TO THE WEST of a Bull (p. 272) ; and on a piece of detached plaster the painting of a male figure, perhaps a bath attendant, holding out a towel or garment (p. 254). Further westward and on the floor below is the larger ruin, y, z, with the Room with the Four Niches (figs. 56 and 74I. Tlie Room zcit/i the Four Xiclies. The floor of this room is 3 feet below the level of the X'ineyard. The southern wall, undermined by much pozzolana digging, has unfortunatelv fallen : the north wall, built against the hill-side, shows that there were three vaulted chambers in the block. The building is clearly of two different dates, the north and partition walls having AV^ l\f ^ ■-x^ F"iG. 74. Room with Four Niches. been built later than the western wall, as may be seen in the north- west corner, where the plaster of the latter has been hacked awaj- to accommodate the north wall. Moreover, the walls are of different construction, the older west wall being of opus incertum, the later wall being of isodomous tufa. The footings of the newer walls are carried up straight to the floor- level, above which there is a set-back of four inches. The rougher and wider-jointed work below the set-back is covered by one course of 20-inch bricks, above which course are si.\ courses of ii-inch bricks and then the tufa wall built in 3-inch courses in neat isodomous ashlar, all as shown in the inset to fig. 74. The spring of the barrel-shaped vault was 6^ feet above the floor, and the chambers ma}- each have been 12 feet long. OF THE UPPER BATHS 121 The westernmost chamber, z, is the most perfectly presented of the three. The west wall, of opus incertum, is curved. In it at a height of 4 feet above the floor are 3 rectangular niches, each about 2 feet across by i foot deep, and there is a fourth niche, rectangular below but rounded above, in the middle of the newer back-wall. Decoration. The west wall was plastered with hard plaster mixed with crushed tile (opus signinum), and was decorated with a Pompeian- red dado extending a yard above the floor. It was embellished with representations of green reeds and other plants, which were painted in a very natural manner. Above was a slight ornamental bead, marking the upper limit of the dado. The ground colour of the walls above was white or cream. Further west are two walls close together which appear to have belonged to a passage or stairway leading northward, which, has yet to be cleared. If we may judge by the general level and direction we should expect this passage to lead up to a space paved with a mosaic floor near the Water Reservoir described on page 127. Id a o fa THE 'BELVEDERE' We now pass in our sun^ey of the Pausilipon antiquities to the western boundar}' of the property. There, high up on the top of the sheer cliff over the Grotta dei Tuoni, are the massive walls of an extensive building, x in fig. 76, one part of which we have named the Belvedere on account of the beautiful view of Capo Coroglio, and westward of Nisida, Cape Misenum and Ischia, which may be obtained from within its walls. The building ends in a low curved wall on a tufa foundation about 4 feet in thickness, upon which wall are built seven brick-faced piers, which in the original plan may have been symmetrically disposed along the quadrant of a circle, but are not so now. The chamber within the wall measures 31 feet wide by 21 feet across the middle. The side walls and part of the back wall are faced with opus reticulatum, but the greater part of the back wall is of opus incertum and gives the impression of having been of more recent construction : it is channelled with two grooves, possibly for water-pipes. They are marked p in the plan in fig. 77. A noteworthy architectural feature of the building is that the side walls are not in line with, or in any way bonded to, the curved wall, the corners of which are faced with carefully laid brickwork. Possibly in the original state columns stood against and within the piers. Columns in such a position would appear to continue the interior lines of the building, and would have carried an architrave with mouldings continuous with those of the cornice of the side walls. It is certain that the piers were covered with marble, for here and there are iron holdfasts firmly plugged in the wall with chips of marble, in the manner usually followed by marble masons in this region. The brick facing to the corners is ashlared into the reticu- lated work in six-course ashlar. The side walls were frescoed — a red dado up to 3 feet from the ground, the main wall being decorated with paintings in yellow panels with red borders. The back wall of opus incertum is 124 THE 'BELVEDERE' covered with very thick plaster of poorer quahty, which may well have been a later addition. The way into this semicircular room is by a rather awkwardly arranged doorway through the north-east end of the quadrant wall, with some tufa-lateritium work at one side. At the south-east some walls emerge here and there from the scrub and debris which require fuller investigation. They appear to Fig. 76. Hill-side betwee.n the Belvedere and the Vi.nevard Site. extend about two-thirds of the length of two great water reservoirs, which lie to one side at a lower level. In our opinion, they belonged to a house of which some rooms were built over the vaults of the reservoirs themselves. The staircase at the south-east end of the reservoirs probably also belonged to this building and led down to a mosaic pavement described on p. 128. A ver}- massive wall, orientation 311", marks the southern angle of the brick-faced building. A view of the south-western wall-face THE 'BELVEDERE' 125 from the old vineyard below is shown in fig. 76. From that point of view, under a on the left, may be seen one end of the quadrant wall and also a straight joint in the wall between two reticulated wall areas — the older work (c), like the newer work {b}, having the wall-end faced with brick lateritium. The position of this straight joint is also shown near the left-hand p in fig. 77. The outer surfaces of b and c were plastered. The wall d of which an angle projects on the south side, may have been continued over the ruins which crop out all the way downhill to the buildings on a flat terrace, which will be described as the Vine3'ard Site in a later chapter. THE WATER SUPPLY The principal facts as at present known about the plan of the water conduits by which the Pausilj-pon Baths were supplied are as follows. One underground channel, with wells at short intervals, came from below the nymphaeum, and led between the theatre and the odeon. It was opened by Mendia and was found to be built in neat isodomic masonrj', and was sufficiently large for a man to walk along. According to Signor Acampora's account, the water came from an aqueduct which ran under the theatre ; there was a well to the south of the temple, and a branch went further skirting the east end of the odeon. To the south of the odeon some 8 yards of cuni- culus of about the height of a man were opened in 1842, and a conduit from it no doubt discharged into the big reser\'oir. The course of the aqueduct is indicated in the large map from hearsaj' and must therefore not be taken as authoritive, for we have no first-hand information concerning it. A low-level branch of the aqueduct is believed to have run down the bottom of the valley near the road, presumably to supply the Lower Baths on the beach. In it a peasant found the two inscribed lamps/and^ described on p. 286. The Water Reso-ooir. This large structure consists of a pair of tanks, parti)' excavated in the hill and partly enclosed by a strong retaining wall measuring 4 ft. 4 in. thick, at a height of about 3 feet above the floor of the tank. They are 64 feet in length b}' 19 feet in width, and are completely vaulted over. The side walls measure 6 feet to the spring of the vaulting, the total height to the crown of the vaulting being 13 ft. 4 in.— a height which, curiouslj' enough, is identical with that of the somewhat smaller reservoir near the Forum Baths in Pompeii. The two tanks are separated by a 3-foot wall, which is pierced by four openings with pointed arches 5 feet high by 4 feet broad, one of which is sketched in elevation in fig. 77. The onl)- openings into the two tanks were three round shafts or well-holes perforating the crown of the vault of the eastern tank and a small oval window in the south- WATER SUPPLY 127 east wall, through which a dim light enters and which may have served as an entrance. The capacity of the tanks would not have been much short of 100,000 gallons. The outlets are arranged as in the Pompeian tank. One in the ^\\H\ll/A\l|l/.,\\|//,//^^ Fig. 77. Plan of the Belvedere and the Covered Water Reservohj. eastern wall, raised about 2 feet from the floor, permitted water to be drawn off without disturbing sediment, and there may have been another on the bottom, for emptying out the tank for cleaning. The pipes would probably have been closed by valves of bronze. There appears to have been another opening in the south corner of the western tank. 128 WATER SUPPLY The accumulation of building-debris over and round the tanks indicates a terrace or rooms, as we find them in Pompeii. And we discovered a staircase at the south end which would have afforded access from a peristj-le (?) below. The steps of lava or piperno rock, each 3 ft. 7 in. wide bj' 14 inches broad, with 8-inch risers, ha\e been built between the end wall of the tank and a less massive wall, 2 feet thick ; the interior surfaces of both walls were plastered. The land to the east of the tanks was under cultivation, so that we were not able to uncover much of the site. However, by digging a trench 8 feet deep we found m situ at the lower end of the stairway a marble threshold stone, /, with sockets for door-hinges and a bolt- hole, ihe foramen in quo ianuae pessuli descouiunt (Marcell. Empir. 17), on the inner or staircase side. Bej'ond the threshold is a white marble mosaic pavement with a border formed of a broad black band between two narrow lines of white tesserae. We found the northern corner of the pavement at about 6 feet north, where there was another doorwa)-, but we were not able to continue the excavation partly on account of the great mass of soil to be moved, and parti}' because it was not permissible to move the vines above, and rarel)- have we more regretted the cessation of work. This pavement is about 8 feet below the level of the vinej'ard, and is about the same level as the hot bath-basin of the Upper Baths (Storey VI). Its complete exploration is greatly to be desired. So far as we were able to make out, the buildings which stood above the tanks were on the same level and probablj^ formed a part of the building near the edge of the cliff which we have already referred to as the 'Belvedere'. Doubtless the connexion between the two sets of constructions might be easilj' demonstrated b)- clearing the inter\'ening piece of rough ground. In excavating the staircase we found several pieces of cur\'ed mouldings of a beautiful marble like cipollino, but of a delicate mauve tint. We figure the section in fig. 144, />. The north-eastern tank in recent times has been used partly as a garden store-room, partly for rain-water storage, and its walls have been mutilated for the purpose. Two holes have been hacked through the side and south-east walls. That through the south-east end wall is reached by a flight of eight steps and has been cut right through two walls and through the ancient stairway between them ; a small modern cistern has been constructed within the eastern tank by building a new cross wall and b}" closing up one of the pointed archways in the central longitudinal wall. WATER SUPPLY 129 Although houses in this region in Roman times were probably better supplied with water from Serino in the Apennines than they are now, yet in the case of almost every villa of any pretensions a large cistern was also thought necessary for such emergencies, as a failure in the supply, or the need of a large flow of water for a short period. For example, there are ruins of twin cisterns, almost identical in dimensions and in construction, and therefore probably contemporaneous, with this Reservoir on the adjoinmg hill above the Vallone di Lampi (p. 172). There, too, the cisterns communi- cated by openings spanned by pointed arches, constructed of similar brickwork. Another ver}' large Roman water-tank, now used as a barn, is one of the ' sights ' of the neighbouring village of Mare- chiano. It is under the principal farm-house. It measures 19^ feet wide by 192 feet deep by 72 feet long, and its capacity is estimated at 175,000 gallons. Fig. 78. Lead \V.\ti:r-Piim;. The Upper Baths were doubtless supplied direct from the reser- voir. Large quantities of the lead conduit pipes were removed by the Marchese del Tufo, who, when in possession of the estates appears to have added considerably to an income derived from excavating pozzolana, by mining lead pipe! Fortunately, a short length hidden amongst the debris encumbering one of the rooms ot the Thermae, escaped Jiim. It is about a yard in length and of the characteristic pear-shaped section with a pinched upper margin, measuring i| inches by i mch inside, and a little over 2 inches across outside. The most important thing about it, however, is the short enigmatical inscription in relief: [GV]A • INAIRDAH • 1NAI[ART ■ SEAC] but which on beingreversed reads: CAES -TRAIANI • HADRI ANI • AVC, and means that alterations or additional plumbing were carried out in his reign — a fact of the greatest importance for the chronology of our site. K I30 WATER SUPPLY A local builder supplied information of a jet larger pipe, 2^ inches in diameter, which, according to his account, formerly ran from east to west across the pozzolana quarry: its continuation is probablj- still lying beneath the rectangular vineyard above the Scoglio di Virgilio. It would have probabl}' supplied the buildings on and under the Scoglio. Various ends of underground water-channels may be seen under the hill above the Vineyard Site, on the cliffs of Trontaremi Ba}', and beneath the Scuola di Virgilio. '&' The Aqueducts. The aqueducts and conduits of Pausilj'pon were plentifully sup- plied with water bj' a branch from the main aqueduct which brought water from above Serino to Naples and to the Phlegraean towns, including Misenum. The bringing of pure, cold water from a distant valley of the Apennines, more than 20 miles awaj-, and the conducting it to so many reservoirs scattered over an arid volcanic region extend- ing along a line of 8 miles or more, was an engineering feat of which an)' civilization niiglit have been proud. The water which was laid on at the Pausilypon \'illa came from the \'allone del Sabato, not far from Abellinum in the Country of the Hirpini. The course of the ancient aqueduct, carefully explored at the command of the X'icero}' Don Pietro di Toledo in the sixteenth century, was described b}' Lettieri, and was found to run underground as far as the Mercato di Serino, thence across two bridges it passed Ajello, Cesenale, Bellezza, and the Piano di Forino. An underground cutting in the rock, three miles in length, brought it to Perduro and Pandula. It passed below Tor di Marcello and Castel S. Giorgio to Taverna di Lazzaro and the Serra di Paterno. Then following the slopes of the hills above Sarno it reached Episcopia, whence it was carried on arches, of which ruins are still visible, to Palma. Near Palma, side branches led to Pompeii and Nola. Another underground channel passed Santa Maria del Pozzo to the Masseria La Preciosa ; and then began a series of high arches, hence the name Pomigliano d'Arco, whicli led on to Casalnovo, Afragola, S. Pietro a Patierno, and S. Giuliano. In tlie Cupa di Miano still stand the high arches, the Ponti Rossi, over which the water flowed and entered a tunnel in the hill which convej'ed it to S. Efremo, to S. Maria delle Vergini, and over other arches near the Museum to S. Agnello. At S. Agnello the aqueduct divided into two branches, one of which WATER SUPPLY 131 proceeded direct to Neapolis, and was the means b\' which the troops of Belisarius entered the town in a. d. 537 ; the other branch skirted the lower slopes of the hill of S. Elmo, and after giving off a side branch to Monte Echia (Pizzofalcone), where the Villa of Lucullus was situated, passed right through the ridge of Posilipo and then right across the Phlegrean Fields to Puteoli, Baiae, and to the Piscina mirabile at Misenum. Pausilypon, according to one account, is stated to have been supplied by a conduit which left the main near its entrance into the hill and ran along the eastern face of the hill, a second conduit along the western face being devoted to the supply of Nisida; but it not only appears improbable that two separate channels should have been constructed to supply two places so near together, but the channel which has actually been discovered does not correspond with either description, being over a third of a mile inside the mountain. When the Grotta Nuova di Posilipo was made for the tramway through the hill in December, 1882, the engineers were surprised to find that the Romans had been in the heart of the mountain before them, for their excavation cut obliquely into a Roman aqueduct running in the direction of the Villa Giulia. For the exact measure- ments and position of the intersection of the Roman aqueduct and the tramway tunnel, I am indebted to Signor Panunzi. From its eastern entrance, fcsiaia (60 feet above sea-level), the tunnel rises with a regular gradient of i in 38. At a distance of 690 3'ards the southern wall intersects the aqueduct at a height of 15^ feet above the tunnel floor, which is there about 115 feet above sea-level. The floor of the aqueduct at the point of intersection is therefore at an altitude of 132 feet above sea-level. In section, the cuniculus is oval, 5^ feet high by 2| feet wide. The upper part is covered with plaster mixed with rough sand, but the lower part forming the water-channel is of opus signinum and is encrusted with a stalactitic deposit from the hard water. The direction is from north to south, though the line is anything but straight. Inscriptions of great importance occur upon the plaster, apparentl}' made with a pointed stick before it had set. At intervals of 96' feet a consecutive series of numerals indicates the distance presumabl}' from the southern end of the aqueduct. They run : DCCC-CCIO • CID (here is the tramway tunnel) C • CC • CCC • CCCC • D, or from 800 to 1000, and presumably from iioo to 1500, and measure Roman feet (taken as 0-97 foot each) with an accuiac}' to within three inches. K2 132 WATER SUPPLY At one spot some cheery person has scribbled 'LIBERI VIVAS (' Long live Liberius'l ; but more important still are inscriptions thrice repeated which enable us to fix the date of this work as a. d. 65 and its destination Limon, the villa of PoUio Felix, the friend of Statius. On the side of the wall was thrice written, though with slight variations * : Macrinus Diadumeni Aug. 1. proc. Antoniani disp. hie ambulavit a villa Poll! Felicis quae est epi Limones usque ad emissarium Paconianum Nerva et Vestino cos (65 a.d.). Hence it follows that the villa of Felix PoUio, mentioned as ' Limon ' by Statius, Silv. ii. 2, was locally known as e-l Adfxo'jjjy ; and though it has been suggested that Limon was the old name for that part of the shore of Posilipo which is now known as Mergellina, j-et it may equallj' well have been a tract nearer the Capo di Posilipo, and so would have been nearer to and more conspicuous as viewed from Sorrento, where Statius composed his poem. At the Capo di Posilipo, the imposing group of buildings on the now sub- merged site of the Roseber}' Region would have been very likel}' to have been those of the extensive villa of the wealth}' Roman, whereas at Mergellina there are absolutel}^ no remains which can be pointed to as those of Felix Pollio's villa. The exact period when Serino water was first brought to Neapolis and on to Limon and to Pausilypon is uncertain. The entire aqueduct has been associated b}^ some archaeologists with the need of fresh water for the naval port of Misenum ; but although it maj' have been prolonged for that purpose )-et the greater length was certainl}' much older. We have good reasons for believing it to have been anterior to the age of Augustus. It is far from probable that a Roman of the class of Vedius PoUio would have chosen a site for his villa which was on the point of an arid volcanic promontorj', destitute of a good water suppl}-. Baths such as he would have considered a necessity require far more water than could have been provided by rain-water cisterns. A more convincing argument may be drawn from the water- works of Pompeii itself. There Oscan letters, used as masons' marks on fountain-basins in certain pre-Roman buildings, point to the exis- tence of a continuous water supply during an epocli anterior to the founding of the Roman Colony, and it has been suggested by Mau that ' Notizie deffli Scnvi, 1883, p. 21 ; Mincrvini, Nuove Scoperk in Napoli, Roma, 1883 ; Mommsen, Hermei, 1883, p. 158. See Insaiptions, p. 213. WATER SUPPLY 133 the great aqueduct was built during the time of peace and prosperity in Campania between the Second Punic war and the Social war. It is unlikely that so great a work could have been undertaken for Pompeii alone ; Neapolis and the towns beyond must have contributed to the cost of construction. Numerous plumbers' inscriptions testify to the continuance of the supply in later years. Pontanus, Dc Magnific. c. 9, says that within his own recollection in the ruins between Baiae and Puteoli, there were found m3.ny fistulae plumbeae niirae crassitudinis which bore the name of Claudius Augustus. In Naples a lead pipe has been found similarly inscribed to the pipe found in the Pausilypon Baths, except that in this case the inscription was not reversed. The general failure in the suppl_v may have been due to the cutting of the conduit by Belisarius in the sixth centur}'. '■7i o 2 n o THE VINEYARD SITE Near the south-west corner of the estate there is a rectangular area which at some time or other has been artificially levelled, and which may be conveniently referred to as the Vineyard Site. A few straggling vines bear witness to a former attempt at culture and to its failure, for although the land lies some 85 feet above the sea, its arid surface is much exposed to salt-laden winds. Now, within a waving fringe of canes, the ground is covered with Achillea and Ferula communis, and with many other plants ; it is a charming field of wild flowers, a pleasant sight in this land of so much forced cultivation. The many ruined walls which peep out between patches of Lentiscus and tussocks of grass on the hill-side above, and jut out over the crumbling cliffs to west and south, show that we have here the last vestiges of a terrace with buildings on both sides, in all probability continuous with the terrace in front of Mr. Foley's house, which according to our enumeration is approximately on a level with Storey IV of the substructions. North Side of Vineyard Site. The ruins in the best state of preservation are those on the north side of the vineyard, where they have been protected by an accumu- lation of debris rolled down from the steep slope of the hill. Here we found a vaulted passage 5^ feet wide and 9 feet high : it extends for about 17 yards and then ends abruptly at the edge of the cliff, where a fragment of its rough concrete vault, a a' (figs. 76 and 79), stands as a weatherworn arch at the height of a sheer 85 feet above the sea. Through the aperture an entrancing view is obtained of the cliff-girt bay of Trentaremi, of pine-clad Nisida, of Misenum, of Procida, and of Ischia in the misty distance. The side walls of the passage, which, before the cliff fell, un- doubtedly extended much further westwards, are faced with panels of reticulate tufa between angles of brick lateritium, disposed as shown on the plan. On the south, 20 feet of wall and a buttress are still standing. The north wall is in a better state of preservation : 136 VINEYARD SITE and patches of the original plaster adhere here and there. It is pierced bj' a door\va3' opening into a small, square, vaulted room b, without windows, which gives the impression of having been used as a store-room. The walls show no traces of colour, but are coated with fine, white plaster of the verj- hardest quality. The corners too, have been eased off with a small oblique filling, so as to allow no nook for dust. Like most of the larger Pausil3'pon buildings, this one is not all of the same date, for the two sides of the entrance to the first ' store-room ' are clearly of different periods : being brick lateritium on one side, two-coursed tufa on the other (fig. 79). Of the westernmost chamber, h, fig. 80, onl}' the near corner is left, the rest having fallen with the cliff. A cunicular passage, 2 ft. 6 in. high, which may have been an aqueduct, opens on the level of the vineyard into the room c. It apparently' comes straight from the middle of the hill and passes through the north wall of the passage obliquely, but we had no time to remove the rubble which blocked it and explore it fiirther. Within a few feet are the remains of another vault with a span of some 16 feet. As seen from the vine^-ard, the northern slope appears to have been laid out in great steps or terraces ; the appearance is due to the persistence of the north walls, d, d', of chambers which were built against the hill-side and have acted as retaining walls to the rubble for twent}' centuries. Parts of the reticulated surfaces of two of these walls are shown in the photograph. Near the eastern end of the slope, in particular, there are about twent}' yards of well-preser\'ed reticulate wall supporting a fragment of vaulting. Near by lies a large block of white marble (fig. 146, b), a piece of the architrave of some large and important building that is unlikely to have stood very far away. Vineyard Site {IVest Side). When the ground was cleared for the vines, the loose stones and blocks of concrete were piled up along the edge of the cliff so as to form a dike, in which canes grow luxuriantly. Outside this dike protrude the jagged ends of broken walls, the partition walls of rooms that have fallen into the sea. Two of these rooms, g, h, measure 37 ft. 6 in. and 23 feet wide respectively — the plaster in the latter was coloured blue. A few fragments of coloured marble pavements and facing, li'ing in niches in the cliff-side, indicate the wealth of finely decorated building that has fallen into the sea below. '^1 <•%' ■-1 ■■ '^i i «l ^^ CLIFFS \ w ^' ■;^' ^S.V / §>- ^v OLD VINEYARD '" ' QUARRY Fig. 8o. Plan of Western Side of the Vineyard Site. 138 VINEYARD SITE Vineyard Site {South Side). The rock under the south side of the old vineyard has been cut straight down almost to the sea-level b^' post- Roman quarriers of stone.' Flush with this rock-face, and threatening to fall, are two ruined walls with a passage between them, evidentlj- once a part of a considerable building (fig. 8i). The passage, 32 inches wide, was vaulted and formerly ended in a curved flight of steps, 5, which turned round Fig. 81. Vi.NEYAKD SiTK. South Side. to the south and gave access to a large room, k. The southern walls and floor of this room k have long since been undermined by the quarrying in the semicircular stone-pit below, but from traces of plaster which still adheres to the face of the north wall, it appears ' The tufa chff near the western end of these ruins exhibits a geologically most interesting section. The rock has been hewn across the course of a small valley in the yellow tufa, which has become filled with volcanic mud which has consolidated as grey tufa, conspicuous on account of the contrast of colour, and clearly shown in our illustration figure 8i. VINEYARD SITE 139 that the dado was coloured red on an undercoating of yellow. The width of the room, if we may judge from a projecting end of wall- FiG. 82. Ground Plan of Walls K shown on the Brink of the Quarry IN FIG. 81. foundation, k\ to the east, was 12 or 13 feet, or if we measure to another wall, P, parallel to these and standing 6 feet further south, would have been about 18 feet. The foundations of these two walls are shown in figure 81 below a group of aloes which are growing in debris between the walls. A continuation of the first- mentioned passage ma}' be traced at a distance of 15 yards further east, at a point where it opens into a somewhat wider passage, j, of which only a short length leading in a north and south direction now remains. To the east of this was a small chamber, l, vaulted from east to west, and ornamented with panelled decorations in red. The floor of this chamber was not very thick when first laid, but in the course of time it had been increased in thickness to 12 inches, the top layer of opus signinum having been added long after the walls were . ....... ... .... .