^{page 3 note 1} In the Addenda I confine myself to the more important points; bare references to the literature of the subject, whether too recent to be included in the original work, or not collected in time for it, are not as a rule given.

^{page 3 note 2} I have not aimed at an exhaustive or complete description of the immediate suburbs of Rome: the great necropolis on the W. of the Via Salaria, for instance (infra, II), would require a volume to itself.

^{page 4 note 1} The passage of Livy (vii. 9. 6) in which he says that the Gauls in 361 B.C. ‘ad tertium. lapidem Salaria via trans pontem Anienis castra habuere’ proves the antiquity of the road and the bridge, the latter playing an important part in the episode of Manlius Torquatus. As to the questions connected with the battle of the Allia, see infra, 23 sqq.

^{page 5 note 1} Persichetti's Viaggio Archeologico sulla Via Salaria nel Circondario di Cittaducale takes up the course of the road from Rieti onwards.

^{page 5 note 2} Both the curatores of whom we have inscriptions are men of equestrian rank (Bull. Com. 1891, 129).

^{page 5 note 3} The question as to the exact point of junction is a difficult one (infra, 27 sqq.).

^{page 5 note 4} We have inscriptions of eight of its curatores, all men of senatorial rank (Bull. Com. 1891, 124) and distinguished career.

^{page 5 note 5} It is an open question whether the road ran beyond Cerfennia before the time of Claudius, (C.I.L. ix. 5973Google Scholar; Besnier, , De Regione Paelignorum, 108 n. 2Google Scholar).

^{page 7 note 1} The expression is somewhat strange and its meaning hard to see.

^{page 8 note 1} It will be seen that the catacombs of S. Anthimus are, according to them, about 22½ miles from Rome, whereas our ancient authorities vary, some iridicating them as 22, others as 28 miles from Rome (infra, 31).

^{page 9 note 1} The word Nova is not actually used in the catalogues, which speak of it simply as Via Salaria; but the addition is convenient as serving to distinguish the two roads, and is generally made.

^{page 9 note 2} By an unfortunate error for which I am responsible the district to the W. of the Salaria Nova has not been included in my map: though the smallness of its scale would hardly have admitted of the necessary clearness in indicating the topographical details of this district.

^{page 9 note 3} Its pavement was found in 1891 at 3 mètres below the modern level, just outside the Aurelian wall, and 3 mètres further down was found another pavement of gravel, pointing to its being a road of considerable antiquity (Bull. Com. 1891, 290; Not. Scav. 1891, 132).

^{page 9 note 4} A recently discovered catalogue is given by Stevenson (Bull. Crist. 1897, 255), but it does not add to our knowledge of this district.

^{page 9 note 5} The Vicolo delle Tre Madonne, the Vicolo dell' Arco Oscuro (both of which diverge S.S.W. from this road) and the cross road connecting them N.E. of the Villa di Papa Giulio are all, probably, of ancient origin, as are, indeed, all the lanes in this district (Bull. Com. 1891, 144). The discovery of pavement in situ in the Via dei Parioli, and of a fragment of a sepulchral inscription, is described in Bull. Com. 1892, 292. I copied there, at the beginning of the descent, a tufa cippus still in situ on the S. W. edge of the road, bearing the following inscription in letters of the last century of the Republic. The letters are 8 cm. high.

In the Vicolo di S. Filippo a cippus of the Aqua Virgo may still be seen in situ (C.I.L. vi. 31565d).

^{page 10 note 1} This is, probably, hardly necessary. Aringhi, , Roma Subterranea, ii. 94Google Scholar, quotes an instrument of Charlemagne, preserved in the Archives of S. Peter's, which speaks of the ‘Salaria vetus quae dividitur ad pontem Molui.’

^{page 10 note 2} Jordan, , Topogr. i. 1. 354Google Scholar; Lanciani, Ruins and Excavations, 75.

^{page 10 note 3} It became known as Via Pinciana in the early middle ages, and William of Malmesbury say of it ‘cum pervenit ad Salariam nomen perdit’ (Urlichs, Cod. Urb. Rom. Top. 87).

^{page 11 note 1} A view of the gate, which had two round towers and three windows above the arch, may be found in Nibby and Gell, Mura di Roma, tav. viii.

^{page 11 note 2} This, the concluding volume of the work, is from the pen of Prof. Hülsen, who has kindly allowed me to see the work in proof.

^{page 13 note 1} Tomassetti, , op. cit. 25 n. 1Google Scholar, erroneously refers some of the discoveries of tombs made immediately outside the Aurelian walls to this portion of the road.

^{page 13 note 2} For S. Chrysanthus and S. Daria see Marucchi, op. cit. 404. The S. Alexander mentioned is a son of S. Felicitas (ibid. 400).

^{page 13 note 3} Nos. 530–532 also belong to figlinae of the Via Salaria, but probably (not certainly) to kilns situated further along the road, in the Sabine territory.

^{page 14 note 1} The ‘Ruderi’ marked in the map are not ancient.

^{page 14 note 2} Hülsen in Pauly-Wissowa, , R.E. i. 2350Google Scholar.

^{page 14 note 3} The inclusion of Labici is a piece of careless writing, for as Strabo himself well knew (v. 3. 9, p. 237) it was fifteen or more miles from Rome.

^{page 15 note 1} He also mentions excavations made in the tenuta of Ponte Salario in 1821, the result of which is unknown.

^{page 16 note 1} Nibby, , op. cit. ii. 594Google Scholar, cites Procopius, , Bell. Goth. iii. 24Google Scholar, fin., as staling that Narses destroyed all the bridges over the Anio; but the passage runs . It certainly looks, however, as if Procopius had here, as in iii. 10 (where he says that Tiber lay on the Tiber about 120 stadia (15 miles)—a rough measurement—from Home, so that Totila's occupation of it prevented the Romans from bringing provisions down by river from Tuscany !), confused the Anio with the Tiber. The Pons Milvius is of course the bridge by which the Via Flaminia crosses the Tiber, and there was no bridge across the Tiber above it until the Via Flaminia recrossed it near Otricoli, nor any bridge below it, except those actually within the city of Rome. Besides, it would have been the bridges over the Anio which it was important to destroy.

Bartoli (Mem. 135, in Fea, , Misc. i. 260Google Scholar) notices that, during winter flood in the time of Innocent XI, one of the banks of the river fell in, and a large marble sarcophagus was found by some boatmen, who broke it to pieces, thinking that treasure was concealed in it. He does not give the exact locality of the discovery.

^{page 16 note 2} Guattani's view (Mon. Sabini, i. 40, cf. 147 n.—the book which he there cites is unknown to me, and the reference cannot be to a main road, but to a mere lane) that the road may have taken to the hills directly after the bridge, is incorrect. Holste (ad Cluv. p. 709, l. 22) is referring to the divergence after Malpasso (infra, 24).

^{page 17 note 1} For its history in early times cf. C.I.L. xiv. p. 453.

^{page 18 note 1} The brickstamp 931b, of the period of Hadrian, copied at Villa Spada in 1741, may belong to this building.

^{page 18 note 2} A little to the E. of this reservoir a round shaft 68 cm. in diameter with footholes (descending probably to a subterranean cistern) has recently been found; and some caves further E., though now much altered, may have served for the same purpose, as a round shaft communicating with them from above seems to be of Roman origin (Gori, Dal Ponte Salario a Fidene Crustumerio ed Ereto—reprinted from Giorn. Arcad. clxxiv. (1863)—9Google Scholar). It may be noted that the contention of this author, that the Via Salaria came up to the Villa Spada itself, is quite unwarranted (supra, 17).

^{page 19 note 1} The quarries of Fidenae are mentioned by Vitruvius (ii. 7. 1) and Pliny, (H.N. xxxvi. 167Google Scholar) as producing soft stone. The tufa here is, as a fact, not of a very good quality.

^{page 19 note 2} Its site is indicated on the map a little to the S. W. of the F of Fidenae.

^{page 19 note 3} Nibby, (Schede, iv. I^vGoogle Scholar) describes it as follows: ‘ruderi di una fabbrica di opera laterizia composta principalmente del cryptoportico (a) con fenestre e feritoie nell' alto. La costruzione è di mattoni sottili con inter … e somiglia a quella delle terme Antoniane; forse è un avanzo di villa o di una parte della Fidene romana’ (cf. Viaggio, i. 76). He then passes to the reservoir at the Villa ‘Edificio quadrato ad emplecton di scaglie di selce forse avanzo di conserva: la larghezza è di passi ord. 25 la lunghezza di 6. 8. [?] il lato settentrionale è in parte rovinato; il meridionale ed occid. hanno ristaur di opera mista dei tempi bassi.’

The reference is to some volumes of Nibby's MS. notes now in my possession (Papers, i. 177 n.).

^{page 21 note 1} In his earlier work, the Viaggio (i. 85) published in 1819, he states that squared blocks of stone (not in situ) were to be seen, which must have belonged to the walls. At that time he placed the arx at the highest point of the hill over the modern road, excluding Castel Giubileo from the circuit of the town.

^{page 22 note 1} It may be noticed in passing that the tombs he indicates above the Casale di Villa Spada are no longer visible—perhaps owing to the fall of the rock. Some damage has very likely occurred to the tombs—though not at all recently as far as one can tell—from quarrying. Lanciani, (Storia degli Scavi, i. 205Google Scholar) mentions the letting of a quarry near Castel Giubileo in 1521.

^{page 23 note 1} C.I.L. xv. 506b, 507b, 702. From Schede, iv. I^v it appears that ibid. 506a was also found.

^{page 23 note 2} C.I.L. xiv. 4064 (a fragment apparently of a sepulchral inscription) was copied ‘in agro Crustumino ad Alliam’ by Detlefsen.

^{page 24 note 1} Op. cit. 44; op. cit. iii. 634. The view is not due to them, but may be found in the older maps of the Campagna, and in Holste (ad Cluv. p. 709, l. 22). Desjardins (Topographie du Latium, 22) propounds the rather strange theory that this road was a somewhat late correction of the original road—which, running along the river valley, would be liable to be interfered with by floods—and that it did not join the Via Nomentana, but returned to the original road after Monte Rotondo.

^{page 24 note 2} This is the site selected by Cluver, (Italia antiqua, 658, l. 45Google Scholar) for Crustumerium—but wrongly (infra, 50, 51).

^{page 24 note 3} Cf. also Hülsen, in Pauly-Wissowa, R.E. i. 1585Google Scholar. This view seems, however, to appear first in Holste (ad Cluv. p. 709, l. 23).

^{page 25 note 1} There seems to be some inconsistency in saying (7 fin.) in connexion with this argument ‘die Befestigungen des republikanischen Roms sind durch die zahlreichen Reste und durch Beschreibungen bis in die Einzelheiten gut bekannt; die so gut wie uneinnehmbare Befestigung war am festesten an der Tiberseite’; and in then (17 ad fin.) placing the present walls after the Gaulish invasion—though it is not denied that Rome had walls of some kind—not, however, ‘so gut wie uneinnehmbare,’ for ‘jedenfalls hat die Gallierkatastrophe die Römer darüber aufgeklärt, dass ihre Stadtbefestigung … dem Ansturm eines grossen Heeres nicht Stand zu halten vermochte’ (18 fin.).

^{page 26 note 1} Gori (op. cit. 31) thought it to be the Via Salaria, and saw near the first chapel what he believed to be a milestone of it !

^{page 27 note 1} 3937 was copied at the Osteria delle Capannuole 1½ mile off on the high road.

^{page 27 note 2} See C.I.L. xiv. p. 439.

^{page 28 note 1} Cod. Urbinas 105 has ἑξήκοντα, which would make the distance 20 miles.

^{page 28 note 2} Cf. Holste ad Cluv. p. 668, l. 25, who remarks that there is no contradiction between the passages, as the site of the battle may well be described as near the town when it was only 33 stades (4⅛ miles) away.

^{page 29 note 1} Guattani, (Mon. Sabini, i. 47Google Scholar) speaks of traces of it seen by Prosseda a little while before he wrote, but says that they had been destroyed: ‘A voler riconoscere da questa parte il confluente delle due vie per mezzo di selci rimasti in opera e fuor d'opera, è duopo volgere a sinistra (dalla Nomentana) per la via che conduce a M. Libretti, e deviando a sinistra ancora giungere alia Collina di Rimane; ove per quei campi non è difficile rinvenirne. Lo Chaupy ve gli ha visti, ed anche ultimamente il nostro Signor Prosseda. Ma come i campagnoli Sabini hanno spianato affatto gli avanzi di Ereto cosi i selciaroli negozianti di vie hanno fatto man bassa sulle consolari antiche massime sulla Nomentana che era forse la piii conservata.’

^{page 30 note 1} We shall see (infra, 74) that this is merely the continuation of the Via Nomentana. Canina falls into the same error.

^{page 30 note 2} He says above (p. 76), ‘on en voit les grands pavés dispersés dans la montée qui la précède immédiatement.’

^{page 31 note 1} I should imagine that it was to this cutting that the following note of Stevenson's (Vat. Lat. 10551, f. 33) refers: ‘1896. D. Carlo Villari mi dice che dopo Monterotondo sulla via ferrata si vede come una strada antica che attraversa la tenuta di Montemaggiore.’ I do not think the course of the Via Salaria itself can be so easily detected that it could be seen from the railway.

^{page 31 note 2} For all this district much valuable information is contained in a volume of Stevenson's MS. notes, now in the Vatican, (Vat. Lat. 10551, 55Google Scholarsqq.).

^{page 36 note 1} Galletti mentions the discovery of a large dolium there in his time (1757).

^{page 37 note 1} Vespignani makes it only 4 metres, but I quote my own measurement. It will be seen, too, that the measurements of his plan do not agree with those of his elevation, the latter being, it would seem, correct. Apparently the scale of the former is about one half too small, which would make the total length about 40 mètres.

^{page 38 note 1} It may be well to remark that Ovid, when he tells us (Fasti, iv. 905 sqq.) that, on his return from Nomentum to Rome, he met the procession going to the grove of Robigus, which was situated at the 5th mile of the Via Claudia, was not returning to the city itself, but to his gardens, which were an the right bank of the Tiber, near the bifurcation of the Via Flaminia and the Via Claudia (Mommsen, in C.I.L. i ². 316Google Scholar).

^{page 39 note 1} For the various discoveries made in the Villa Patrizi see Not. Scav. and Bull. Com. passim (from 1885 onwards).

^{page 39 note 2} The first of these pipes is attributed by Dressel to Augustus.

^{page 39 note 3} About the same place another road ran southwards to the Praetorian Camp: it is described as having been found about 500 metres from the gate, on the right of the modern road, and as running from N. to S. It lay 1·70 mètre below the modern level and was 2·50 mètres in width (Not. Scav. 1903, 93; Bull. Com. 1903, 290).

The tomb inscription, Kaibel, I.G.I. 1444, is given as having been found ‘in vinea viae Nomentanae’: while 2069 was found on the same road in 1601. Both are recorded by Sirmond.

^{page 39 note 4} Some remains of a Christian cemetery were also discovered (Bull. Com. 1888, 148, 174Google Scholar).

^{page 40 note 1} A fragment of a Greek inscription (no doubt from a tomb) was found.

^{page 41 note 1} Bartolini (S. Agnese, 118) is wrong in stating that the ancient road crossed the modern after the Villa Torlonia, passing N. of S. Agnese and through the valley by the Sedia del Diavolo (infra, 45).

^{page 41 note 2} A comparison of these two accounts, which are both from the same pen, will show an extreme case of the difficulty to which I have alluded above (3, 4). In the former the site of the discovery is described as ‘presso il muro di recinto dell' Istituto delle Snore della Provvidenza,’ in the latter as ‘dall’ altro lato della via Nomentana (from the Villa Torlonia) presso l'imbocco della via Pasqualina.’ To one unfamiliar with the nomenclature of the streets of the newest quarters of Rome, which is not always to be learnt from the ordinary maps, neither description is of very mucli use: and there is considerable danger that it might be supposed that two different places were referred to. But, further, the discovery of the pavement of the road is mentioned only in Bull. Com., while for the measurements of the cippus and to learn the depth at which it was found, one must go to Not. Scav. It would seem to the unprejudiced observer that it would be a better method of proceeding to give a complete account of the discovery in one periodical: the course at present adopted is somewhat annoying (cf. Class. Rev. 1903, 329; 1904, 137).

^{page 41 note 3} The sepulchral inscription, Kaibel, I.G.I. 1857, was found here in the 16th century. We may notice the discovery in the restoration of 1620 of the series of eight basreliefs, which are now in the Palazzo Spada (Helbig, , Führer, ii. nos. 989–996Google Scholar: two others in the Capitol, ibid. i. 469, 470, belong to the same set). The place whence they came is uncertain: at S. Agnese they were used as building material, while the other two were found, one in the Piazza SS. Apostoli, the other on the Aventine. A statue of Hercules (wrongly restored as killing the Hydra) also in the Capitol, was found here (ibid. i. no. 412) and so was the statue of the drunken old woman (ibid. i. no. 439, cf. Bartoli, mem. 100, in Fea, , Misc. i. 250Google Scholar). The cippus bearing the funeral inscription of Q. Vedennius Moderatus, who rose to be architectus armamentarii imperatoris (C.I.L. vi. 2725) was also found close to S. Agnese in 1816. It has an interesting relief of a Roman catapult (Röm. Mitt. 1904, 255). For later discoveries see Not. Scav. 1885, 251; 1901, 423; and for Christian antiquities ibid. 1901, 14, 489; 1902, 366. See also Addenda, infra, 208.

^{page 42 note 1} Excavations in the interior (Not. Scav. 1888, 507, 570, 732) led to the discovery of a baptismal font and several fragments of inscriptions: the building would seem, therefore, to have served as a baptistery at one time, though it is probable that this was not its original purpose.

^{page 42 note 2} Vacca, mem. 124, records on his father's authority the discovery in the Vigna of Angeluccio da Viterbo near S. Agnese of the statue of a sow—in the mouth of which was a metal plate with the inscription amplius si laboraveris. Whether there is any truth in any part of the story I hardly know.

^{page 42 note 3} Fea's note (la forma è rotonda) shows that he misunderstood the reference, which is not to the mausoleum itself, but to the space in front of it.

^{page 49 note 1} Whether a copy of C.I.L. xv. 7626 was also found here is doubtful: Amati may have been mistaken in attributing it to Fidenae as well as to Ficulea (infra, 59).

^{page 49 note 2} In Eph. Epigr. vii. 1271, the locality is vaguely given as Fidenae.

^{page 49 note 3} Gell (op. cit. 45 and map) indicates two other tumuli, one some way to the E. of the Casale S. Colomba, near the source of the Fosso Bettina, the other to the S. of S. Colomba, and apparently a little way S. of the Fosso Bettina (though his map is not very clear). I have not examined either.

^{page 50 note 1} There are various forms of the name.

^{page 54 note 1} The statue is preserved in the courtyard of the bank itself.

^{page 55 note 1} The remains further to the S. E. will be described infra, 104 sqq., in connexion with the Strada Vecchia di Montecelio—and so also the inscriptions C.I.L. xiv. 3993–5. Other sepulchral inscriptions found in the tenuta are given ibid. 3996–9, and two lead pipes, 4000, 4000a ( = xv, 7621, 7709).

^{page 57 note 1} I refer to the path coming N.W. from Tivoli.

^{page 58 note 1} At Castel Chiodato the inscription C.I.L. xiv. 3930/1 was seen upon the holy-water basin <no doubt, as so often, a Roman cinerary urn converted to this use): but it is no longer in existence.

^{page 58 note 2} Westphal (op. cit. 125) notes that after the so-called Molino del Moro traces of antiquity are wanting in any of these roads.

^{page 59 note 1} Here was probably found the inscription C.I.L. xiv. 4005, the provenance of which is quite uncertain, as it was used in a ‘selciato’ or road pavement.

^{page 59 note 2} So Nibby, in Analisi, ii. 50Google Scholar, and Schede; iii. 36; in Schede, i. 177, he states that it was used as a step at the Casale.

^{page 60 note 1} It may be worth while to give the original text of the two accounts, which differ slightly in details. That in Schede, iii. 36 runs thus: ‘Nel Novembre dell“ anno 1824 furono intrapresi scavi nel tenimento della Cesarina poco oltre il casale in una fimbria di colle parallela a quella nella quale sorge il casale. Essi furono continuati nei mesi seguenti anche di la di quella punta, e dappertutto vi furono trovati indizi ed avanzi di villette che insieme doveano formare uno o più paghi siccome dalla iscrizione che più sotto riporto (C.I.L. xiv. 4012) può dedursi. Il fiumicello che scorre sotto la Cesarina forse fu detto Ulmano. Deviando dalla via Nomentana poco oltre il Torraccio della Cecchina al V. m(iglio) a sinistra si scende all' Ulmano e passatolo sopra ponte si sale al casale della Cesarina che è circa 1 m. dopo il diverticolo. Ivi trovasi impiegato come gradino un masso quadrilungo di travertino colla iscrizione seguente (C.I.L. xiv. 4011). Oggi questo masso è in Roma. (Il casale fu degli Sforza e n'è prova l'arme rapp(resentante) un' orso legato ad una colonna.)

‘Nello scavo sulla fimbria opposta a questo casale furono trovati pavimenti di camere da bagno, e queste iscrizioni sepolcrali’ (C.I.L. xiv. 4033 and the fragment given above) ‘… leggesi in una figlina ivi rinvenuta’ C.I.L. xv. 509. 2. (A.D. 133)—508a, which I found there belongs to the same year, and has practically the same legend, ‘Hibero et Sisenna cos. ex. pr. Ulp(i) Ulpian(i) Sal(arese).’ ‘Ivi pure fu trovato un condotto colla epigrafe C.I.L. xiv. 4018 sovente ripetuta. A poca distanza da questo scavo, forse un mezzo miglio più oltre, fra i ruderi di camere ben decorate di marmi fini, porfido, serpentino etc. frammenti di colonne si trovò la lapide seguente (C.I.L. xiv. 4012) … che sembra essere stata inserita in un muro.’ The account in Analisi, ii. 50 is similar. On the copy is the note ‘trovata alia Cesarina più di ½ miglio a nord del casale.’

The account in Schede, i. 117 says ‘Il casale è di costruzione per quanto apparisce intieramente moderna: vi si trovano però d' intorno sparsi massi di travertino: un tempo fu de’ Colonna come si vede da un arma rovesciata non antichissima che ha una colonna sormontata da un aquila a cui è incatenato un orso: per uno scalino è impiegato un gran travertino (C.I.L. xiv. 4011).

‘Un buon quarto di miglio a sett, del casale sono stati fatti nell' inverno dell' anno 1824 scavi, e si sono trovate camere appartenente ad una villa sontuosa del primo periodo del secondo secolo, con pavim. di marmo ma che sembrava aver sofferto l'ultima distruzione: ivi si sono scoperte (un leone add.) parecchie teste una delle quali di Lucilla moglie di Lucio vero, mold frammenti di giallo, rosso, alabastro, africano, verde e serpentino, e due iscrizioni’ (given above).

^{page 62 note 1} Liv. ii. 16 tells us that it was across the Anio, Plut. Popl. 21, that it was near the Anio, so that it was probably rather to the S. of a straight line between the two places. Cf. Wissowa, in Pauly-Wissowa, R.E. iii. 2650Google Scholar.

^{page 62 note 2} Hülsen (Pauly-Wissowa, , R.E. iv. 1604Google Scholar) proposes to read Φειδηναίονς: for Tellenae was in a different direction (Strabo, v. 3. 4, p. 231; Nibby, , Analisi, iii. 146Google Scholar). See Addenda, infra, 208.

^{page 63 note 1} So Cingolani's map, Nibby, , Analisi, ii. 424Google Scholar—Tomassetti (op. cit. 50) puts it on the right of the Via Nomentana, at about the 12th kilomètre from Rome, but wrongly.

^{page 63 note 2} The list of the inscriptions found with 4001 given by Dessau in loc. (‘Borghesiane, Ficulesi, Aprile 1826,’ praescripsit Amati huic et eis quae hanc sequuntur) is full of errors, and the inscriptions themselves must be consulted.

^{page 64 note 1} With this inscription were found others, including one of travertine with the epitaph of a scurra, the text of which is not given.

^{page 65 note 1} This is the site selected by Bormann (Altlalin. Chorographie, 255) for Corniculum.

^{page 66 note 1} Hirschfeld (Die Kaiserlichen Verwaltungsbeamten bis auf Diocletian (1905), 369 sqq.) decides in favour of Dessau's view.

^{page 67 note 1} The theory that Nomentum was situated on the Monte d'Oro has little to recommend it.

^{page 68 note 1} So in the text, though illustrations of three only (two inscribed on both sides) are given, and the fourth is not further mentioned.

^{page 69 note 1} Fabretti, (inscr. 241, 655)Google Scholar saw 3961 ‘at the 13th mile of the Via Nomentana’ (i.e. near the tombs alluded to supra, 67, 68) and (ibid. 371, 148, 149) 3973, 3974 ‘at the 11th mile near the ruins of an ancient theatre’ (i.e. the reservoir near Monte Gentile, supra, 65), but the whole group are so much alike that Dessau thinks that they must have been found together.

^{page 71 note 1} Strabo calls the Albulae, Aquae ‘cold’ (infra, 117, n. 1)Google Scholar.

^{page 73 note 1} These objects are in the possession of Signor Bonfigli of Palombara, who kindly allowed me to examine them.

^{page 74 note 1} Various conjectures as to its ancient name are given by Nibby, , Analisi, ii. 347Google Scholar.

^{page 78 note 1} Typical blocks measured 58 × 45, 60 × 35, 54 × 40, 66 × 70 × 65 cm. The stick which appears in Figs. 6, 7 is 93 cm. high.

^{page 79 note 1} In my opinion an exception must be made in favour of the lowest of the group towards the N.W. extremity, which ascends somewhat sharply in a curve.

^{page 81 note 1} Here a dedicatory inscription to Diana is said to have been found (C.I.L. xiv. 3928).

^{page 81 note 2} As Nibby's description of Le Rotavelle—the district with which we are dealing—in Analisi, ii. 534 is somewhat brief, it may be of interest to give the full text of his notes (Schede, iv. 34): ‘Giovedi 29 Maggio (1823) ci dirigemmo a Monticelli e Palombara (da Tivoli). Passato il ponte dell' Acquoria si trova poco dopo un rudere di emplecton di scaglie di selce (infra, 151) quindi una cappelletta ed un bivio; noi prendemmo a sinistra benche la strada più breve a Monticelli sia la destra: seguendo la via a sin. vi rimarcai molti poligoni che per antica fanno riconoscere questa strada, la quale sembra la stessa che come via antica Tiburtina trovasi descritta in Cabral e nella sua Topografia (infra, 110, 152). Due miglia dopo Tivoli a qualche distanza a destra vidi ruderi di pietre quadrate di sostruzione sotto un colle (infra, 167) e ½ miglio dopo altre rovine pur di pietre quadrate sopra di un colle forse avanzi di qualche antica città in questi dintorni“— Caenina (?) (infra, 173). After describing his visit to Monticelli, where the only ruins of which he could hear were those near Colle Ferro at Le Caprine (infra, 119), he continues: ‘scendendo verso Palombara, si traversa una macchia dentro la quale appena disceso il colle di Monticelli circa un miglio distante dal villaggio dopo il lavatore cominciasi a trovare ruderi antichi: quindi veggonsi avanzi di un’ acquedotto di opus mixtum ed altri ruderi s'incontrano ad ogni tratto ora reticolati ora laterizi di buona costruzione (infra, 179)). Altri ruderi si trovano poco prima di Palombara … Nella Casa Ferretti che è nella parte nord-est dell' acropoli e sotto di essa vidi un frammento di orologio solare concavo, una testa di cervo, ed una iscrizione di

cui mi presento il propvietario solo la copia manoscritta come trovata a Rotavelle' (C.I.L. xiv. 3929: according to other authors here cited, the inscription itself was at Slazzano).

‘Il Sig. Ferretti ci accompagnò la mattina del 30 a Rotavelle luogo circa 2 m. distante da Palombara nella direzione di Moricone: ivi osservammo ruderi incogniti di reticolato non regolare, molti altri rud(eri) d' incertum: un capitello ionico di lavoro grossolano di travertino. Nella vigna Ferretti che la prima percorremmo fu trovato l'orologio solare: quindi passammo nella vigna Belli dove riconoscemmo a fior di terra mura di poligoni della epoca terza, cioe politi da tre parti, che certamente furono recinto di citta rintracciandosene gli avanzi per un miglio almeno, forse Cameria. Ivi dappresso trovammo una conserva a tre navi di opera reticolata con otto archi la cui volta è moderna, e un acquedotto d' incertum con canaloni. Nella vigna Imperiali si vede un pezzo di opera laterizia che è ad angolo retto colle rovine preced(enti). Vi si veggono altri ruderi informi ed astraco. Dopo questa uscendo e diriggendosi verso occid(ente) veggonsi avanzi di una crepidine di via che va da sett(entrione) a mezzo giorno e dopo questa seg(uendo) la direz(ione) merid(ionale) trovasi un' altra conserva con volta crollata di un' emplecton finissimo: un' altra se ne trova dopo sepolta, e quindi nel luogo denominato Martini si vede il giro di una piscina circolare di circa 90 palmi di diametro: evidente ivi si vede l'andamento di una via antica della quale se ne può precisare la larghezza a 14 palmi ( = 3·10 mètres) la quale secondo la relazione de' naturali viene dalla Fiora e può considerarsi come parte dell' altra e communicazione fra le vie Salaria, Nomentana e Valeria’ (cf. Analisi, ii. 293).

^{page 83 note 1} Above it, further E. are the remains of what appears to be a church, to the N. of which are the terraces of Monte Madano (supra, 78).

^{page 84 note 1} For the division of the roads radiating from Rome into these two classes, see Papers of the British School at Rome, i. 127.

^{page 86 note 1} A block of peperino, built into the later wall which follows the line of the S. wall of the Praetorian camp, and measuring 0·70 by 0·50 mètre, bears the following inscription:

Nibby and Gell (Mura di Roma, 336) give it thus:—

^{page 88 note 1} Platner (Topography of Ancient Rome, 120) attributes the foundations of the towers to Aurelian.

^{page 89 note 2} The roads apparently ran at the old level, while these huge heaps of rubbish accumulated on either side of them: when the rubbish was levelled down, the roads were correspondingly raised. Cf. Mon. Linc. i, 476, ‘la via Tiburtina dei tempi di Augusto corre a tre metri sotlo la soglia della porta di Onorio, ed è fiancheggiata da sepolcri costruiti a bugna di sperone, e da colombai di maniera reticolata.’

^{page 90 note 1} The description is vague— ‘larger side’ would have been more correct, unless indeed the meaning is that the slope is across the axis of the road. But this is improbable, and it is most likely that this is a portion of the road from the Porta Maggiore to S. Lorenzo, the pavement of which was discovered in 1881 in a vineyard at a depth of 10 mètres below the present ground level, together with a tomb (Bull. Com. 1881, 203; C.I.L, vi. 22076). If this is so, the line of the road is not correctly given in Papers, i. map i.

^{page 97 note 1} On its S.E. edge, not far before the bridge is reached, is the rectangular concrete foundation of a tomb with one block of travertine still in situ.

^{page 98 note 1} Uggeri, Giornata a Tivoli, 15, says that he found the width of the road at various points, further on to be only 12 feet, and that it was the narrowest Roman highroad that he knew. Cf. also infra, 101, 114, and, for a much greater width, 120, 124, 126.

^{page 98 note 2} Uggeri, , Vues des Environs de Rome, vol. xviii. (Tivoli), Pl. IVGoogle Scholar, gives a view of the bridge, showing the position of these blocks.

^{page 99 note 1} Excavations made at S. Basilio by Castellani (apparently on behalf of Antonio d'Este, director of the Vatican Museum—cf. Mon. Sabini, i. 225) in 1811 are described by Guattani (Mem. encicl. vii. 83). A building of opus reticulatum, decorated with paintings, and later on converted into a tomb, was found: in it was a large sarcophagus, 13 palms long, 6 high, 6 wide (the measurements are given as 15 × 11 × 7½ in Mon. Sabini, loc. cit.) (a palm is 0·2223 mètre), very roughly worked, within which were two bodies dressed in cloth of gold, of which Nibby, (Analisi, i. 288Google Scholar) says that the costume was thought to belong to the 6th century A.D. A marble cornice was found with the names P. CORNELIVS and IVLIA CORNELIA and fragments of other inscriptions bore the name of the same gens—among them possibly, as Tomasselti conjectures (op. cit. 38, n. 1), C.I.L. vi. 16111.

Some way to the E. of the Casale S. Basilio is the Casale Monastero, at which is a rectangular building in inferior brickwork: the interior, with a large niche on each side and an apse at the end, measures about 8 by 7 mètres: the ceiling has quadripartite vaulting. A marble door-jamb and a fragment of the cornice over the door are still in situ. Foundations of other buildings may be seen close by. To the S.W. are the remains of a villa.

^{page 100 note 1} A sepulchral inscription discovered ‘in a vineyard near the 7th kilomètre’ is published in Bull. Com. 1899, 262.

^{page 101 note 1} Hirschfeld (op. cit. supra, 66 n., p. 163, note 4) rejects Bruzza's interpretation, R(ationis), preferring R(ecognitum) as suggested by Dressel, who compares a similar mark on amphorae (C.I.L. xv. p. 562, i.).

^{page 102 note 1} This is the casale on the W. edge of the ancient road, between it and the river.

^{page 102 note 2} Elsewhere in his papers Revillas notes at the 6th mile of the modern road, ‘Ponte antico sotto l'osteria detta la Casetta de' Cavallari posta a mano diritta. Vedonsi dalla stessa parte in lontananza i rottami d'antico luogo’ (the reference is perhaps to some mediaeval ruins N. of the 8th kilomètre of the modern road).

^{page 102 note 3} Diss. iii. Tab. i.

^{page 102 note 4} Cf. the map of the Dorsum Praenestinum et Tusculanum added to the 2nd edition (opp. p. 90).

^{page 102 note 5} Cf. Eschinardi (ed Venuti, 1750), Descrizione di Roma e dell' Agro Romano, 235: ‘La strada fin qua (Settecamini) è tortuosa, e arenosa, il che non credo fosse dell' antica Via Tiburtina, vedendosi a luogo a luogo vestigi dell' antica via selciata.’ He does not however tell us on which side of the modern road, nor at what points, these fragments of pavement were to be seen.

^{page 103 note 1} There are, however, indications of paving in the bank of the stream, belonging to a road ascending to the N.E. of the tower, which must have joined the Via Tiburtina at the loth kilomètre.

^{page 103 note 2} On the higher ground to the W. and N. are the remains of other buildings.

^{page 103 note 3} The inscriptions C.I.L. vi. 1933, 13143Google Scholar were copied here in the 18th century.

^{page 104 note 1} To the N. in the Riserva dello Spavento are the remains o a water reservoir and of other buildings, and to the N. of it again a building with two square niches on each side and one at the E. end, the arches of which have large impost blocks of travertine. There are some blocks which look like paving-stones in a bridge on the path which runs N. to Marco Simone, not far from these ruins: but I have no certain proof that it follows the line of an ancient road.

^{page 104 note 2} They are republished in C.I.L. xiv. 3993–5.

^{page 105 note 1} The name Inviolata is a corruption of In Via Lata, the tenuta having been the property of the church of S. Maria in Via Lata (Nibby, , Analisi, ii. 157Google Scholar).

^{page 106 note 1} The travertine quarries at this point are in the main modern: the remains of a villa in opus reticulatum with a portion of a water reservoir may be seen in and above them. The reservoir has one gallery perfectly preserved, with the wall dividing it from the next, but whether it had more than two chambers is uncertain. There are four arches in the dividing wall (which is 77 cm. thick), the two central ones measuring 1·23 mètre in height with a span of 1·52 mètre, while the two side ones are 99 cm. high with a span of 1·15 mètre: the chamber which is preserved measures 8·40 by 2·89 mètres, and 2·55 in height to the top of the vaulted roof.

A large mass of fallen concrete in the quarry may or may not belong to this reservoir. A travertine column drum 66 cm. in diameter may be noted in the field above.

^{page 109 note 1} Revillas in his notes makes the following calculation in canne of 10 palms each = m. 2·223.

Divide by the ancient mile at 660 canne 4 palms, and the result is that the estimated distance to Tibur is just over 194 miles. Most of the items to be added are, however, put at rather too high a figure. From the Porta Viminalis to the Porta Chiusa is only 570 mètres (nearly 256 canne), while the additions to the length between Ponte Mammolo and Settecamini (or il Forno) are probably excessive; and finally, it does not seem clear to me that anything ought to be added in respect of the difference in distance from the Porta Chiusa and from the Porta S. Lorenzo to the point of junction of the earlier and more recent roads (supra, 87, 93).

Revillas' calculations do not, therefore, remove our difficulties.

Duchesne, (Lib. Pont. i. 326, n. 14Google Scholar) quotes Bruzza's account of the church of S. Severino, built by Honorius I., ‘iuxta civitate Tiburtina, miliario ab urbe Roma XX,’ which he identified in 1883 as being 1½ miles beyond Tivoli on the road to Vicovaro, precisely 20 ancient miles from the wall of Rome. Cf. Bruzza, Regesto della Chiesa di Tivoli, 95.

^{page 110 note 1} Among them is Ansaloni (the author of an unpublished work on Tivoli, the MS. of which is preserved in the Jesuit college there), who wrote in 1791 (i. 241–243).

^{page 110 note 2} It is just possible that these may have belonged to a road from the Lago della Regina to Ponte Lucano (see infra, 126, n. 1).

^{page 110 note 3} Canina's plan (Edifizi, vol. vi. tav. 120) shows it as running 500 yards to the E. of the modern road.

^{page 111 note 1} Promis (op. cit. 24) describes it as having six—perhaps counting the entrance—and. gives its measurements as 5·97 by 3·05 mètres.

^{page 111 note 2} A cutting existing to the S. of the modern road, and parallel to it, close to the 11th. kilomètre, does not appear to be of ancient origin: a similar one may be seen at the 13th kilomètre, where it cannot be other than modern.

^{page 113 note 1} He has just been speaking of the relief of a lion from a tomb near Ponte Lucano (infra, 141, n.).

^{page 113 note 2} See Lanciani, , Storia degli Scavi, i. 115Google Scholar.

^{page 114 note 1} Revillas notes the existence of another, further still on the same side.

^{page 114 note 2} Revillas saw the crepido as far as Le Tavernucole. The bridge he notes as ancient, but it has been entirely rebuilt since his day. Many of the paving-stones now serve as the floor of the courtyard of the Osteria.

^{page 116 note 1} The only exception is when Revillas quotes Volpi's reading of C.I.L. xiv. 3554.

^{page 116 note 2} The modern road is probably slightly to the S. of the ancient line here (Bulgarini, Notizie di Tivoli, 132 init.).

^{page 116 note 3} In the tenuta of Tor dei Sordi, or else in that of Lunghezza, was found the Greek metrical sepulchral inscription published by Grossi-Gondi, Il tempio di Castore e Polluce sul Tuscolo (1901), 17.

^{page 117 note 1} Strabo (v. 3. 11, p. 238) calls it cold like that of the Aquae Labanae (supra, 71).

^{page 117 note 2} His account is reprinted in Giorn. Accad. lxxi. (1837), 61Google Scholarsqq.

^{page 117 note 3} The following account of the state of the building towards the end of the 16th century may not be without interest: (f. 103^v) ‘la magnanima et Regal fabrica fatta da Cesare Imperatore Augusto la quale si ritroua discosto dalle d^e Accque da cinquanta passi … le Accque passauano dal Luogo oue risorgono per Aquidotto sotto la Terra, e si conduceuano al d° Luogo … Ordinato con gran disegno et Artifitio con le sue stufe, et scali di diuersa fatturâ con Pauimenti di musaico adorni con con un Teatro ricinto intorno di un Ordine di Bellissime Colonne di Breccia uerde le quali sono (f. 104) tanto in prezzo con statue di mafmo diuerse, Le Colonne sono da 30 Palmi grosse proportionatamente di Ordine Toschano con suoi Bellissimi Capitelli, e sue Basi, et acció si habbi da Creder che le d^e Colonne siano di Valore, e di Bellissima Natura, La felice memoria di Papa Giulio 3° ne hebbe Notitia di queste Colonne Vaghe, però subbito Ordinò che douessino andare a Roma delle quali se ne serui per Ornare la sua Uigna Uicino al Palazzo del Papa, che hoggi di si fa chianare la Uigna di Papa Giulio, ma doppo che le forno polite, et imbronite le furno apprezzate mille e cinquecento scudi l'una. delle Medeme Colonne se ne ritrouano anche quattro altre in la Chiesa di S. Pietro della Citta Nostra di Tiuoli … et altre si ritrouano in lo Medema Luogo di Bagni per il Medemo Adornamento ui si ritrouano anche in d° Luogo di Bagni di Molti Altri pezzi rotti per terra.…

‘Si Ritrouano Anche in d° Luogho di Bagni Muraglia grosse più di Venticinque palmi con li suoi Seggi Ordinatamente per la Comodità di coloro che doueuano prendere li Bagni in de Muraglie (f. 104^v), si uedono Aquidotti e credo che seruiuano per Sfumatorij delle Stufe, perche altra effetto non poteuano fare per ritrouarsi cosi alti in de Muraglie più di Ottanta palmi.’

The passage is taken from a copy of Zappi's MS. history of Tivoli (the original of which is preserved in the Municipal Library there) which I acquired at the sale of Prof. Costantino Corvisieri's library in 1902, and which came from the library of the Briganti Colonna family. The original dates from 1583 (C.I.L. xiv. p. 371, xiii.).

Kircher (Vetus Latium, 203) seems to have seen columns of serpentino verde as well: ‘erat fabrica haec Thermarum … columnis ex ophite, quem serpentinum vocant, lapide suffulta, quaedeinde avulsae Romam allatae feruntur, et eae putantur esse, quas Constantinus imperator ia Ecclesiae Lateranensis a se fundatae ornamentum applicuit.’ The value of his statement is, however, uncertain—for on p. 119 he repeats a conjecture that the columns in the Lateran were found at Torre Nuova (Papers of the British School at Rome, i. 233).

^{page 118 note 1} To the inscriptions found here must be added C.I.L. xiv. 3541. Revillas (Herculis Fanum) gives the following account of it: ‘praestat … alium lapidem proferre recentissime, anno scilicet MDCCXXXVI ad Aquas Albulas effossum, nonnihil tamen mutilum

The indication of its find-spot in the Corpus is incorrect, for the passage quoted from Volpi runs infull (the italics are mine and denote the portion omitted):

‘trasportata, Dio sa quando, nel sito del Collegio vecchio de' Padri della Compagnia di Gesù, è stata ritrovata e scoperta dal P. Girolamo Tebaldi della medesima Compagnia, sagace investigatore ed amatore sollecito della venerabile Antuhità in quest' anno 1736.’ How, in the very year of its discovery, Volpi could profess such ignorance of the place where it was found is another matter.

^{page 123 note 1} Besides a short description of the tomb given on the authority oi Sig. Antonio Petrucci, its owner, the following passage occurs: ‘in distanza di palmi 20 dal monumento e stato disotterrato un masso rettangolare, in una facciata del quale si scorge un perfetto triangolo, nel cui mezzo è il corpo lunare, che incomincia a presentare le sue fasi, e sonovi inoltre tre stelle.’

^{page 125 note 1} The name ‘Tiburtinus (with ‘Lunensis’ and ‘Lesbius’) lapillus’ occurs in a sepulchral inscription—C.I.L. vi. 13830; cf. also infra, 201. We may cite the description of Zappi (f. 138): ‘… et in quel luogo ui sono Restate quattro Memorie, in questo Modo hanno lasciato del Medmo Sasso a radicato della Terra Proprio, si come la Natura l'ha creato dico un quatrangolo grosso più di dodici palmi per faccia et alto 30 palmi, et questo li Scarpellini di quel tempo si risolsero a lasciare queste quattro Memorie acció si conoscesse la Moltitudine delli quadri, che da quel luogo si Cauauano. … et tutte quelle scaglie, ouero scarpellature che il scarpello Buttaua uia per Abozzare il quadro, et altri Lauori, ne ridussino tanta gran Moltitudine che ne fecero un Monte il quale oggi di si chiama il Montarozzo della Uiuara, Contrafaceuano quel Monte di Testaccio in la Città di Roma.’

^{page 126 note 1} The existence of a good many paving-stones in the fieldwalls along the first part of a lane which runs W.N.W. from Ponte Lucano to the Lago della Regina may indicate its antiquity: but the evidence is not sufficient to assert it positively. A similar doubt must be expressed with regard to the line—tentatively marked as ancient in the map—which follows the modern road until a point to the E. of the 21st kilomètre, and falls into the road from Ponte Lucano to Palombara.

^{page 126 note 2} For this work on Tibur, of which only one printed copy exists—from which various MS. copies are derived—see C.I.L. xiv. p. 371.

^{page 127 note 1} This is also the view of Zappi, f. 136^v: ‘Ricinta d^a Mole con un Teatro di Colonne di Pietra Tiburtina alte dodici Palmi di Ordine Ionico con le sue nicchie di Mezzo Rilievo, ma il Teatro resta Riquatrato dico con quattro facce di tal sorte che la d^a Mole o' Sepoltura la si ritroua restare in mezzo circondata intorno come lo dico con due Bellissimi epitaffi.’

^{page 129 note 1} The name fundus Ceseranus occurs in a document of 924 A. D. (Reg. Subl. f. 185) published by Bruzza, Regesto della chiesa di Tivoli, 112. In a bull of 978 the confines of it and of the fundus paterno (infra, 148, n.) are given as (1) via publica, (2) paterno, (3) flumen (the Anio), (4) carrarea (a cart track—qy. the ‘Carrara di Paterno’ (?). But just after this we have ‘fundum silicatum. fundum caccabelli. ab uno latere silice qui descendit a ponte lucano. a secundo latere alia silice qui pergit ad palatium antiquum. A tertio latere via publica. et a quarto lapide fundum gostanti.’).

^{page 131 note 1} From this point a road (possibly of Roman origin) diverges to S. Vittorino, passing through an archway cut in the rock, which is known as the Porta Nevola, and which is of quite uncertain age. Somewhere between Porta Nevola and S. Vittorino below the hills of S. Germano (wherever they may be) is a nymphaeum cut in the rock and decorated with mosaics and shells—according to Raffaele del Re, in his edition (1883) of the first five chapters of Antonio del Re's Storia di Tivoli, p. 225. It is locally known as La Grotta di Paris.

Upon the Colle Fiorilo, Kircher (Latium, 188; cf. map opp. p. 142) places the site of Aefula—wrongly: but Fabretti (De Aquis, map opp. p. 90) marks here ‘rudera alterius oppidi ad viam Collatinam.’ Revillas, on the other hand, places the ruins of Aefula on the Colle Tasso, where there are the remains of another large villa.

Another building of which I do not know the site is mentioned by Maria Graham (Lady Callcott), Three Months in the Mountains East of Rome (1820), 14. ‘We [going from Le Capannelle towards Poli] entered a thicket that clothes the steep banks of the stream. As we ascended, we passed the foundation of some large antique building, formed of great square blocks of Peperino, and observed a fluted marble column lying across the path. These remains are near the little unhealthy town of San Vetturino’ (sic).

^{page 132 note 1} See however Fonteanive, Avanzi Ciclopici, 85; Giovenale, in Diss. Accad. Pont. vii. 333Google Scholar; cf. Mélanges de l'École française, 1905, 185. But the instances cited are rather cases of the use of tufa in ‘polygonal’ masonry in the narrower sense, i.e. where there is an intentional avoidance of horizontal bedding.

^{page 132 note 2} The form Aefula is given by the best MSS. of Horace and is also found (in the cognomen Aefulanus) in an inscription from Carthage of the Republican period (C.I.L. ii. 3408); cf. also C.I.G. 3187; C.I.L. vi. 34220, 34221. Hübner, , Hermes, i. 426Google Scholar.

^{page 132 note 3} The bricks forming the floor of a room discovered here in January, 1745, bore the stamps C.I.L. xv. 1061, 1075a, 2385, and a fragment ‘ex pr. Domitiae Lucillae.’.

According to letters written by Silvestro Petronselli of S. Gregorio to Revillas (on Sept. 7th and 13th, 1739), which I acquired in the Corvisieri sale, mosaic pavements were found in the vineyard of Lorenzo Lupidii, situated upon the Colle Faustiniano, also vaulted chambers with walls 8 palms in thickness, with bricks bearing a stamp, of which only the letters QSE appear to have been legible (possibly C.I.L. xv. 2385, P. CQSEPTICIORI). One hundred and thirty paces from the so-called Casale Grande an aqueduct 2½ palms (about 56 cm.) in width was found, which probably supplied this villa.

From a sketch-map given by Petronselli the villa might fairly be conjectured to be near the house marked 312 on the Staff Map (Colonna sheet): Petronselli's knowledge of the aqueducts seems to have been extensive, as these maps and his letters show.

Further towards the Mola a large ruined round tomb of opus reticulatum, originally of the size of that near the Ponte dell' Acquoria (infra, 151), but only preserved to a height of about 8 palms (1·80 mètre), was seen.

^{page 133 note 1} This form is given by Weissenborn, who notes no other reading.

^{page 135 note 1} See Lib. Pont. ii. 11 (ed. Duchesne): ‘et in sancto Angelo in Fagano fecit (Leo III) vestem de fundato.’ ii. 92, ‘hic vero praesul (Sergius II) cum de omnibus ecclesiis sollicite curam gereret, etiam basilicam Sancti Archangeli, quae in cacumine Fagani montis est constituta, largiorem quam pridem fuerat a fundamentis perfecit, ac radientibus picturis luculente pingere iussit, ac sarta tecta eius noviter restauravit.’ The origin of the name Faganum or Faianum (the latter is an alternative reading in the second passage) is not clear (Bruzza, Regesto di Tivoli, 138); Duchesne is inclined to derive it from fagus.

^{page 135 note 2} Graham (op. cit. 19) says, ‘the antique paved way from Tivoli to Palestrina, which runs in a line with the Catena, shows itself in more than one spot in the corn land we passed through.’

^{page 136 note 1} Here, according to Graham (op. cit. 17), ‘there are many ancient substructions; and funeral vases and other antique fragments have been found.’

^{page 136 note 2} The Οὐέρεστις (or Οὐέρεστις) ποταμὸς of which Strabo (v. 3. 11, p. 239) speaks as flowing through the territory of Praeneste, has been variously identified. Many writers before Nibby believed it to be the Fosso dell' Osa, which does not however, as he justly remarks, touch the territory of Praeneste. His own identification of it with the Acqua Rossa is, however, not certain (Analisi, iii. 465), for Strabo's indication of its position, ῥεῖ διὰ τῆς χώρας (Πραινεστοῦ), is not sufficiently definite, and we have no other mention of it.

^{page 137 note 1} The Fosso Saviano is the name given by the Staff Map to the upper portion of the Valle della Mola.

^{page 138 note 1} ‘Passato il ponte [over the fosso di Acqua Rossa] diriggendosi a sin. verso un fenile si riconosce essere questo fondato sopra i ruderi di un' antica villa costrutta di grandi massi di pietra quadrata e di opera reticolata. Credo che le pietre quadrate di tufa fossero un' opera precedente alla quale poi venisse addossata la costruzione reticolata, ovvero che fossero queste tolte dall' antico recinto di Corcolle … Del resto dalle rovine di questa villa si traccia tutta la sua spianata inferiore … il vicino tumulo macchioso [point 72 on the map] copre forse le rovine del piano nobile.’

^{page 139 note 1} The name appears in the 11th century (‘Annales Romaines 1044–1073’ in Lib. Pont. ed. Duchesne, , ii. 335Google Scholar).

^{page 140 note 1} See Piranesi, , Antichità Romane, ii. tav. 38, 39Google Scholar, for plan and view. Cf. Helbig, , Führer, ii. no. 823Google Scholar (who speaks as if this relief were no longer in existence).

^{page 141 note 1} Ligorio, (Bodl. Canon. 138, f. 117Google Scholar) gives the following particulars: ‘Di un altro sepolcro guasto. Questo altro è uicino al sopra detto del quale hoggi non ui è rimasto nulla, perche noi di ueduta le havemo uisto guastare, et il pilo è ben uero che era gittato a terra rottissimo, ma l'arco che il tencua pensile è stato uenduto et toltone uia i sassi dalli tiuolesi: et per esser bella compositione, e' di inuentione no' ho' percio Iasciato che io non lo habbia qui disegnato, che se la fortuna lo ha' fatto spianare, non ha potuto pero far tanto, che non habbiamo potuto col mezzo della carta è del inghiostro fare che non ne sia affatto spenta la memoria: il pilo poteua esser longhoXVpiedi, et largo VI alto otto piedi.’ The sketch which should have accompanied this description is wanting in the MS., and as the passage quoted conies immediately after that cited supra, 113, it is possible that the reference is to a tomb near the 9th mile: but as the first and third of the three tombs of which we have been speaking form the main subject with which Ligorio is dealing, it is possible that he is here describing a fourth member of the same group, especially as he states that it was the people of Tivoli who destroyed it.

On the other hand, it is to be noted that Zappi says nothing of the destruction of any tomb, and speaks only of the three mentioned in the text (f. 135^v):

‘in luogo doue si dice Serena, ui si ritrouario tre bellissime Memorie antique, e rare come ui dirò, si ritroua principalmente un Leone di Marmo del Naturale in un Posamento Rileuato in alto da 35 Palmi in circa, questo Posamento si ritroua esser fatto di quadri di Pietra Tiburtina. Ma il Leone resta scolpito in un quadro di Marmo per ogni faccia da 12 palmi in circa, doppo si ritroua un altra Memoria di un Cauallo, et di un homo grande del Naturale. … La Terza Memoria sono doi Statue rileuate in alto da 30 palmi con il Mede(si)mo posamento, Ma discosto l'una Memoria all' altra da 60 Palmi, e tra esse due Figure ui si ritrouano una Roba Tonda, et si uede essere posata sopra una Tauola …

‘Son forzato dire che quella Memoria del Cauallo tenuto da quel homo per le Redini della Briglia si ritrouauano di tanta Vaga Bellezza che gli forno leuate le Teste ad Ambedue … da un Gran. Sig. Caualiere nel tempo della Guerra di Papa Pauolo quarto dell' Anno 1557, che di rincontro alle d(ett)e tre Memorie notabili ui si ritrouauano li soldati, et esercito del … Re filippo Re di Spagna, gouernato sotto … il Duca d'Alba.’

Bartoli (Gli antichi Sepolcri, tav. 47, 48, 49) gives views of all these tombs—the last after a drawing by Pietro da Cortona.

It would also be interesting to know whether Zappi saw a more direct road from Ponte Lucano to the Villa Adriana than any of which we know at present—say along the path from the bridge to the easternmost of the two houses marked C. Galli on the map, where at present there are no traces of antiquity. He speaks (f. 137^v) of two roads starting from the tomb of the Plautii, the one going to Tivoli, ‘e l'altra uerso la Gran uilla di Adriano Imperatore discosta Mezzo Miglio secondo che si uedono li Vestigij delle Siricate [sic, for Selciate] Antique accosto alia d^a Mole.’

Nicolas Audebert is probably referring in the following passage (Brit. Mus. MS. Lansdowne, 720, f. 306) to a discovery of sarcophagi, and not to the tombs of which we have spoken:

‘Peu plus loing [than Ponte Lucano] on veoit a coste du chemin dedans une vigne un reste de peinture antique qui est contre un vieil mur tout ruiné et y a quelques sepultures et urnettes coe cercueils, en facon d'un long coffre, capable d'un hòmme y estendu, Le tout de fort beau marbre blanc, avec facons de sculpture et aultres ouvrages par dehors. Aussy se veoit un peu plus loing et plus proche du chemin un vieil marchepied ou montoir de pierre contre lequel est grave l'Epitaphe d'une mule de Crassus …

Dis Pedib. Saxum

Cinciae Dorsiferae et Cluniferae'

(naturally a forgery—C.I.L. vi. 3443a^*). He also notices (f. 307) that the ancient paving was well preserved along the road to the Villa of Hadrian, but he is not precise as to its exact course.

^{page 145 note 1} At the end of the second line Petronselli notes ‘le lettere mancanti sono corrose,’ but he only leaves a space for two letters more at most, and shows no traces of a third line.

^{page 145 note 2} He refers to the village to which he belonged—S. Gregorio (supra, 134).

^{page 146 note 1} The plan (a simple hexagon) is given by Dehio and v. Bezold, Kirchliche Baukunst des Abendlandes, t. i. fig. 10.

^{page 146 note 2} Two fragments of a Greek inscription, cut on a slab of white marble, are built into the stairs of the cottage in the vineyard. I have not been able to obtain a satisfactory reading of them, as they are much worn, and therefore give the text with all reserve. The letters are 2 cm. high.

^{page 147 note 1} Petronselli thought that this aqueduct came from the bridge at Pomata (infra, 195).

^{page 147 note 2} A much better sketch is given by Lanciani, Bull. Com. 1899, 35.

^{page 148 note 1} The lane which runs on the left of what we have assumed to be the line of the ancient road is called Carrara di Paterno. The name Paterno is an old one: the act of donation to the church of S. M. de Cornuta, of the year 471, published by Bruzza, Regesto della Chiesa di Tivoli, 15, and Duchesne, , Lib. Pont. i. cxlviGoogle Scholar., mentions ‘fundum Paternum maranus [sic], fundum Mons Paternus’: and though this locality may not be referred to in that document, it seems to be certainly spoken of in the second document published by Bruzza, a Bull of Marinus II (945), in which the property of the cathedral of Tivoli is mentioned (p. 20, l. 25, fundum paterno), as also in a document of 942 A.D. (Reg. Subl. f. 171, Bruzza, p. 117), and in subsequent bulls (supra, 129, n.).

^{page 149 note 1} A precisely similar inscription (C.I.L. xiv. 3583) refers to the restoration of an unknown bridge, for the inscription was not found in situ, though the I onte dell' Acquoria is not improbably referred to.

^{page 149 note 2} Gruter, , Pag. xxxivGoogle Scholar.

^{page 149note 3} Montfauc, . Antiq. Explan. tom. i. 2. 373sqq. (Paris, 1722)Google Scholar.

^{page 152 note 1} The following extract from the bull of Benedict vii. (978), published by Bruzza, Regesto della Chiesa di Tivoli, may be of interest (p. 33, l. 32): ‘miliario a suprascripta civitate plus minus quinto. in campo maiore. Ab uno latere staphiliano. Et a secundo latere silice in qua est ponticello. Et a tertio latere crepidinis. Et a quarto latere piranis. Et a quinto latere columella.’ Cf. a document of 990 A.D. (p. 40, l. 10): ‘partem terra sementaricia in fundum qui sancto Valentino nuncupatur. cum predicta secclesia infra se. cum gripte et parietinis suis antiquis in mini positi … posita territorio tiburtino miliario a civitati tyburtina plus minus tertius. Inter affines ab uno latere incipientem a staphile lapideum deinde venientem per via publica, etc.’ The road referred to is in both cases that leading N. from the Ponte dell' Acquoria (Bruzza, p. 174 fin.), and the ‘campus maior’ is the flat ground to the W. of it and of the Colle Nocelle.

^{page 153 note 1} Some further remarks on the dating of Cyclopean masonry in Italy will be found in the description of La Civita, near Artena, by Dr. G. J. Pfeiffer and myself, which will shortly appear in the Suppl. Papers of the American School of Classical Studies in Rome, i. 87–107. Compare also the reports on the recent excavations at Norba, Not. Scav. 1901, 514–559; 1903, 229–262, and my paper on Monte Circeo in Mélanges de l'École française, 1905, 157 sqq., and especially 181–186. In the last a distinction is drawn between Cyclopean and polygonal structures, the latter term being reserved for those in which an intentional avoidance of horizontal bedding is evident (supra, 132, n. 1).

^{page 155 note 1} Below this villa to the W.N.W. is a large open water cistern, 17·8 mètres in diameter, with five buttresses on the W. side, where the ground slopes away. To the S. of the villa, close to the Anio, is another small one, rectangular in shape.

^{page 156 note 1} In the Archivio Storica dell' Arte, 1890, 196, Prof. Venturi gives some account of excavations made by Ippolito II d'Este in 1550–1560. About the former year a Hercules, a headless Venus, and another statue were found in Tivoli, the former being given to Ippolito's brother, Ercole I, Duke of Ferrara. About 1560, Pirro Ligorio was in charge of excavations at the Villa Adriana and the Villa of Quintilius Varus.

^{page 156 note 2} The passage has been already published by Gori, , Archivio Storico di Roma, iv. 256Google Scholar.

^{page 157 note 1} He seems however, to mean the Acqua di Piavola.

^{page 161 note 1} I cannot locate the excavations described in Bull. Inst. 1832, 6, ‘Nelle vicinanze di Tivoli il Sig. Arduini ha cavato alcuni pezzi di cornice di rosso antico; due piccole colonne, ed un pavimento di marmo palombino a scacchi.’

^{page 161 note 2} Nibby, , op. cit. iii. 221Google Scholar, speaks of remains of marble facing slabs and stucco in some of the rooms, and Antonini, (Manuale di vari ornamenti, ii. 24, 25Google Scholar) gives two rosettes from fragments of ancient sculptures preserved here.

^{page 161 note 3} This chamber measures 10 metres in width by 8 in length, not including the apse, which is 4·5 mètres in diameter.

^{page 162 note 1} In the garden at: S. Antonio I copied two brick stamps,

neither of which appears to be otherwise known. The former, (tegula) Toneiana de (figlinis) Qtiint(ianis), is a case of the mention of two brickworks in one stamp, which seems to occur only where the figlinae Tonneianae are concerned, and for which no explanation has yet been found (C.I.L. xv. p. 188). The combination with the figlinae Quintianae is a new one.

^{page 162 note 2} M. Jullian maintains that this farm was actually within the territory of Tibur (Mélanges de l'École française, 1883, 82); but see Dessau (C.I.L. xiv. p. 368).

^{page 163 note 1} Cf. also Antonini, , op. cit. ii. 23Google Scholar.

^{page 163 note 2} No doubt the 'fundum trulias’ of the Bull of 978 (supra, 129, n.).

^{page 163 note 3} In the church of S. Angelo existed the inscriptions C.I.L. xiv. 3544, 3662.

^{page 164 note 1} Since the above description was written the road has been almost entirely destroyed.

^{page 166 note 1} The name occurs in a Bull of Anastasius IV (1153–1154), Bruzza, op. cit. p. 76, l. 9.

^{page 167 note 1} Some of these were, according to Bulgarini, op. cit. 101, used for the edge of the new fall of the Anio. Its foundations have recently been cleared, and its plan brought to light. It had two chambers both 4·1 mètres wide inside, the one 4·43 mètres long, the other 2·35, with a door connecting them.

^{page 169 note 1} This is not shown on the map.

^{page 169 note 2} A letter from Lesley to Lupi, describing a visit to this villa, and dated May 8th, 1732, is preserved in Cod. Vat. Lat. 9143, f. 89. In it the following passage occurs: ‘ieri fui a Vitriano, viddi quel marmo che Marco havea fatto scavare. L'avevano sepelito di nuovo in mezzo di un campo, è di più rotto in vari luoghi.’ No further description is given, and it is clear from the context that the reference is not to the villa itself.

^{page 172 note 1} Dodwell calls it Medullia, but without adequate reason (infra, 175, 183).

^{page 172 note 2} The style of the walls at Turrita, if closely observed, shows some differences from that of the villas of the district. There has been, it is true, a certain amount of weathering: but though the joints were originally good and the faces of the blocks fairly smooth, there is not the extreme fineness observable in the later work; on the other hand, there is no intentional archaism, such as that which leads to the bossing of the faces of the stones, and no positive avoidance of the horizontal line. The wall is 1·28 mètre in thickness: the blocks in the foundation are larger than those above. The material is pudding-stone quarried on the spot. Mortar was not originally used, but has been introduced with the addition of the mediaeval superstructure.

The ancient walls are well preserved on the S.W. and W.: on the W., where alone this projecting knoll joins the rest of the hill, there is a slight projection in the line of the wall, as though there was an entrance at this point—which would indeed be the natural one. On the N. and E. but little of them is to be seen.

The site measures (maximum measurements) 185 paces from E. to W. and 70 paces from N. to S.; but the E. end is almost the apex of a triangle.

The name Turrita occurs in a document of 1030 (Bruzza, op. cit. p. 67, l. 15), ‘de plebe de sancto Johannes qui est posito intro castello qui dicitur turrita’; and also in a Bull of Anastasius IV (1153–1154), ‘ecclesiam sancte romule de turrita.’

On the opposite side of the railway, on the slope to the N.E. of the station are foundations of roads and walls in Cyclopean work, of which, in their present state, but little can be made.

^{page 173 note 1} That the site of this place is quite uncertain we have already seen (supra, 134).

^{page 173 note 2} The reference is of course to the straight line taken by the track which has but recently been superseded by the modern road with its curves and zig-zags. See the passage of Nibby's Schede quoted supra, 82, n.

^{page 174 note 1} In a garden along the highroad, S.W. of the cemetery, are fragments of columns and an Ionic capital.

^{page 174 note 2} Infra, 176, n. 3.

^{page 176 note 1} Colonnella means ‘a small column.’

^{page 176 note 2} For a sarcophagus found on this hill, see Not. Scav. 1894, 146.

^{page 176 note 3} Probably referred to under the name of Le Grotte by Marocco, Stato Pontijicio, x. 64. He also mentions excavations made hereabouts by a certain Signor Mendola, not long before 1836, in which were found the remains of a temple (supra, 174).

^{page 179 note 1} My reading is in some points uncertain, and I have not had another opportunity of verifying my copy: Signor Bonhgli has, however, kindly done so.

^{page 179 note 2} Whether the modern road up to Montecelio follows an ancient line, or no, I cannot say.

^{page 180 note 1} It is mentioned in a document of 1030 (Bruzza, , op. fit. no. xii. cf. p. 87Google Scholar).

^{page 181 note 1} The church belonged to the abbey of Subiaco.

^{page 181 note 2} Annales Romaines (1044–1073) in Lib. Pont. (ed. Duchesne, ) ii. 334Google Scholar.

^{page 183 note 1} For Cameria, which is generally supposed to have been at or near Palombara, cf. supra, 76.

^{page 183 note 2} Remains of a pagan cemetery, as well as those of the older church, are said to have been found at S. Michele, on the hill to the S. of the village, in 1724, when the foundations of a new monastery were laid (Casimiro, Memorie delle chiese e dei conventi dei Frati Minori della Provincia Romana, 187).

^{page 184 note 1} Whether the latter inscription, which speaks of ordo dec. … Moeniensium, may not have been brought from Castrimoenium is uncertain: it is a fragment found in a wall of a church in 1853.

^{page 185 note 1} Probably 1688, as, although the title-page bears the date 1690. the ‘Approbatio operis’ was given on Jan. 12, 1689.

^{page 186 note 1} As the exact position is not given, I have not marked it upon the map.

^{page 186 note 2} Poggio Cesi, the summit between Montecelio and S. Angelo, which is higher than either, is occupied by a mediaeval castle; and there are no traces of any constructions of an earlier period— a remarkable fact.

^{page 187 note 1} They have inadvertently been omitted from both maps: the circular reservoir is to the E. of them.

^{page 188 note 1} In one place courses of baked bricks ·04 mètre thick, and tufa blocks 0·07 thick by about 0·20 long, are arranged alternately in threes.

^{page 188 note 2} Two very poor views of it are given by Volpi, , Vetus Latium, x. 1Google Scholar, opp. p. 360 (reproduced in Veteris Latii Antiquitatum Amplissima Collectio (Rome, 1776), pt. i. pls. 13, 14)Google Scholar.

^{page 189 note 1} Prof. Hülsen locates the villa, in my opinion, rather too far down the hill; for where Cabral and del Re's plan indicates it there are no remains of a villa, and one would think that they have inaccurately represented the locality of the large villa at the 26th kilomètre (which they place to the N. of the path), for they speak of it (op. cit. 137) as if it were a building of considerable size. It is, however, somewhat difficult to find the road which they call the Strada delle Piaggie (op. cit. 132) on the map: in fact, it is apparently the first of the three roads described by me which corresponds with their Strada di S. Marco, which descends directly to Casale Leonina.

^{page 190 note 1} Stevenson (Vat. Lat. 10552, f. 1) mentions the existence of paving-stones at the 26th kilomètre of the modern road, which probably belong to our road.

^{page 190 note 2} I cite the paging of the reprint from Atti dei Lincei, ser. iii. vol. iv. (1880)Google Scholar.

^{page 191 note 1} Both Nibby, (Analisi, i. 389Google Scholar) and Sebastiani (op. cit. 230) are at fault. Cabral and del Re (op. cit. 165) call this villa the Villa of Brutus; but the contract for the excavations of De Angelis calls it the Villa of Cassius, and so do Nibby and Sebastiani.

^{page 191 note 2} This group, according to Sebastiani (loc. cit.), was bought by Lord Jennings: Bulgarini calls him Penchins: but we really have to do with Thomas Jenkins, English consul and antiquity dealer, who bought it for 600 scudi, and resold it to a ‘Milord Inglese’ for 4,000.

^{page 193 note 1} In Stevenson's MS. notes in the Vatican library (Vat. Lat. 10552, . 23) there is a sketch plan of the tomb, a single rectangular chamber, about 3·70 by 2·70 mètres, cut in the rock: immediately behind it passes the specus of an aqueduct (probably the Anio Vetus) which curves around it in such a way as to indicate that the tomb is earlier in date than it. He also gives a sketch of the stone door slab of the tomb. There follows a copy by him of a document entitled ‘Notitia delle reliquie ritrovate in una grotta del Territorio di Tivoli luogo d° Carciano,’ from which it appears that the tomb was opened on April 28, 1693, in the presence of Canonico Antonio Filippi, delegated by the Reverenda Camera Apostolica. A large slab of travertine was found, which served as the door of the tomb, with what was thought to be a cross upon it; but Stevenson's sketch of the slab (which he saw on the spot in 1879) shows that it was an imitation of four panels with a round hole to represent the keyhole. Five large nails, each one palm (m. 0·223) long, were found in front of it. Behind this slab was the tomb itself, a chamber containing four large sarcophagi, three of them each of a single block of peperino (two of them with heavy lids of the same material), and one of travertine with a cover of the same stone. On the stucco coating of the front of each of them, except of that which had lost its cover, were letters which could not be deciphered, except in the case of the travertine sarcophagus, where MAIOREO M could be clearly made out.

^{page 195 note 1} A ‘fundus fusci’ is mentioned in the Bull of A.D. 945 (supra, 148, n.), and this villa is called by some writers the ‘Villa of Fuscus,’ but without sufficient reason.

^{page 196 note 1} Some paving-stones may be seen in a newly-made wall along this path.

^{page 196 note 2} It is on the E. edge of the Colonna sheet of the Staff Map.

^{page 196 note 3} Presumably this is the villa on the Colle Marcoraino, S.E. of pt. 492 near the W. edge of the Palestrina sheet of the Staff Map.

^{page 198 note 1} C.I.L. vi. p. lxiii. no. cii.

^{page 198 note 2} Stevenson (Cod. Vat. 10552) givesacopy of part of a letter relating to this very map from Canina to Coppi (dated Nov. 7th, 1855), lent him by Card. Nardoni in 1890, which runs as follows: ‘Stando a Tivoli mi venne dato di rinvenire nella libreria dell' Episcopato il rame della importante carta topografica della Diocesi Tiburtina rilevata dal Revillas che non si sapeva dove esistesse e resa molto rara, ma solo cognita per una nuova incisione fatta dal Petrowski che cancellando il nome del Revillas se la fece propria. Mi feci rimettere il detto rame da M. Vescovo di Tivoli e ne feci tirare 50 copie.’ Either Canina considerably overrates the rarity of the map, or else the copies which not infrequently come on the market in Rome are some of the fifty which were printed by Canina.

^{page 198 note 3} I obtained a copy of this map in Rome in April, 1905. It is as described by Mommsen, (C.I.L. ix. p. 347)Google Scholar, who only saw a drawing of it, and bears the date 1735 and the dedication ‘ill. ac rev. domino Josepho Baronio vigilantissimo Marsorum episcopo’: it was engraved at Rome by Sintes.

^{page 198 note 4} See Michaelis, , Ancient Marbles in Great Britain, § 36, p. 61Google Scholar.

^{page 199 note 1} In this letter Sir Charles Frederick remarks: ‘io vi avrei ciò mandate per mezzo di My Lord Carlisle, ma la difficoltà ch' io trovai di farne una versione in Italiano mi ha preoccupato.’ Lord Carlisle (the fourth earl) was much interested in the antiquities of Rome, and may be the person referred to above, whose patronage Revillas sought or thought of seeking.

^{page 199 note 2} He speaks of him a little further back (f. 9) as his informant with regard to the Servian wall: ‘quum etenim ab vinitore, qui a multis iam annis vineam colebat, percontarer, num in terrae effbssionibus veterum murorum quandoque emersissent, absque haesitatione respondebat, muri non lateritii, sed quadratis lapidibus exstructi ingentia frusta, multa humo cooperta non uno in loco secus aggere inventa fuisse, in parte eius extima.’ He refers either to the Vigna dei Certosini or to one next to it within the Villa Peretti (Montalto).

^{page 199 note 3} De aquis (ed. 1788), Diss. iii. tav. ii.

^{page 199 note 4} From the draft of a paragraph, ‘De diversione et mensura viae Tiburtinae ab urbe ad Albulas,’ we may add the following: ‘ubi viatrium laevorsum occurrit vinetis inserviens’ (the Vicolo delle Mattonelle, as Gori calls it—Porta e Basilica di S. Lorenzo, 73).