^{page 14 note 2} English Historical Review, xxxii (1917), pp. 531–533, 543Google Scholar. The text is reproduced at the end of the present article.

^{page 14 note 3} Dr James prints the extracts from the Rolls Series edition. E.H.R. pp. 533–536.

^{page 14 note 4} The Mirabilia is sometimes described as a guide book for pilgrims, but Duchesne has pointed out its true character as ‘le plus ancien essai de topographie erudite,’ though ‘cette exegèse savante est tout ce qu'on peut voir de plus artificiel et de moins réussi.’ This interest in antiquity was due to the rise of Roman political independence. Everything points to the Mirabilia being the work of Benedictus Canonicus, and written before 1143 when the Commune of Rome was established. Introduction to Le Liber Censuum de l'Église Romaine, vol. i (Paris, 1910), pp. 98–104Google Scholar. See also Gregorovius, , History of the City of Rome in the Middle Ages, English translation (2nd ed.), iv, p. 306Google Scholar.

^{page 14 note 5} The earliest manuscript of the oldest form of the Mirabilia (Cod. Vat. 3973) belongs to the end of the twelth century (Urlichs, Codex Topographicus Urbis Romae, p. 91). The Liber Censuum, to which it forms an appendix, was compiled by Cencius Camerarius (Honorius III) in 1192. Duchesne, Introduction, p. 1.

^{page 14 note 6} Sections 4, 27, 29.

^{page 15 note 1} H. Omont, Les Sept Merveilles du Monde au Moyen Age. Bibl. de l'École des Chartes, xliii (1882), pp. 43 ff. Migne, , Patr. Lat. xc, 961 ffGoogle Scholar.

^{page 15 note 2} I have called attention to most of these as they occur.

^{page 16 note 1} Cf. Huelsen, , La Pianta di Roma dell' Anonimo Einsidlense (Atti d. Pontif. Accad. Rom. di Arch. Ser. ii. vol. ix, 1907), pp. 44Google Scholar: ‘L'itinerario che per la cronologia sta più vicino all’ Einsidlense, l'Ordo Benedicti Canonici, fa vedere gli enormí cambiamenti che la città aveva sofferti fra il secolo VIII ed il XII.’ This inability to read Roman inscriptions, which seem so plain to us, may be illustrated by the remark of Dr.James, in his Wanderings and Homes of Manuscripts (London, 1919; p. 31Google Scholar) that ‘MSS. written in capitals or uncials … were uniformly despised and neglected by the readers of later centuries’ as being use ess and illegible.

^{page 16 note 2} E.g. the sections about Fabricius (25), and Scipio Africanus and Hannibal (26). ‘Marius Suetonius’ (29) may be due to confusion with his continuator, Marius Maximus, whose name would be known from the references in the Historia Augusta.

^{page 16 note 3} The allusions are: urbis opus in 15 from Aen. v, 119; and moritura pallescit (of Cleopatra) in 22, which seems to be derived from pallentem morte futura of Aen. viii, 709. See below, p. 40, n. 1.

^{page 16 note 4} Iudicio nostro vincit utramque Venus in 12 is evidently derived from Ars Amat. i, 248 (of Paris): cum dixit Veneri ‘vincis utramque Venus.’ The comparison of the bronze bull of the Castle of Crescentius to the group of Europa and the bull (3) is perhaps an allusion to the description in Met. ii, 851 sqq.: mugit et in teneris formosus inambulat herbis. Compare the words of Gregory: mugituro et moturo similis.

^{page 16 note 5} In the description of the reliefs on the Arch of Augustus (22) bella detestanda seems to be derived from bellaque matribus detestata of I Carm. i, 24; while superba mulier (Cleopatra) may well be a reminiscence of Horace's superbo non humilis mulier triumpho, the concluding words of the ode (I, xxxvii) on the victory of Actium, which, as we have pointed out elsewhere (p. 40, n. 1), lays stress on just the same episodes as those mentioned by Gregory.

^{page 16 note 6} Gesta Regum, iv, 351 (Rolls Series, ii, p. 403).

^{page 17 note 1} Lanciani, , Storia degli Scavi, i, p. 5Google Scholar; The Destruction of Ancient Rome, pp. 162, 165.

^{page 17 note 2} Cp. Gregorovius, v, p. 43.

^{page 18 note 1} There are three references to Gregory the Great as a destroyer of statues (4, 6, 12). Notice also the reference to the temple of Pallas (16): multo sudore Christicolarum deiectum.

^{page 18 note 2} It is just possible that there is a local allusion in his account of the temple of Pallas (16), where, he says, the Christian martyrs were condemned in the days of the persecutions (Duchesne, , Le Liber Pontificalis, i, p. 152Google Scholar). He goes on to mention one by name—Hippolytus; and it is not obvious why he should have done so, unless he had some special interest in him. In the early Middle Ages it was believed that the Abbey of Saint Denis possessed the body of Hippolytus, and his day (Aug. 13th) was kept there with peculiar solemnity. Acta SS. vol. xxxvii, pp. 8 ff. This allusion would be still more appropriate if we could believe that Gregory came from the ancient monastery of St. Hippolytus (Sankt-Pölten) in Lower Austria (Augustinian from the eleventh century), which, it may be noted, was taken under the protection of Alexander III, c. 1180, and of Innocent III in 1206. Felgel, and Lampel, , Urkundenbuch etc. (Vienna, 1891), i, pp. 18, 31Google Scholar.

^{page 19 note 1} Graf, A., Roma nella memoria e nelle immaginazioni del medio evo (Turin, 1915), p. 86Google Scholar, quoting Angilbert and Dante. This regular approach to Rome over Monte Mario (Mons Malus) is illustrated by Paschal II in IIII sending his envoys to meet the emperor Henry, Vin montem Gaudii qui et mons Malus dicitur (Annales Romani, M.G.H. Script, v, p. 474Google Scholar; Duchesne, , Lib. Pont, ii, p. 340Google Scholar. Cp. Gregorovius, iv, p. 338), and by Frederick Barbarossa's entry in 1155 (Gregorovius, iv, p. 539). The names ‘Mons Malus’ and ‘Mons Gaudii’ first appear in the early middle ages, but Duchesne says that ‘leur origine précise est inconnue’ (op. cit. p. 262, n. 4). The latter is generally explained as the pilgrims' name for the spot whence they caught their first view of the city. It may, however, be due to the principle of euphemism, like Benemons ventum for Maleventum.

^{page 19 note 2} Gregorovius, iv, p. 691.

^{page 19 note 3} Gesta Regum, iv, 352 (R.S. ii, pp. 404–408).

^{page 19 note 4} Duchesne, , Lib. Pont, i, p. 152Google Scholar.

^{page 20 note 1} Urlichs, Codex Topographicus, pp. 92, 115, 127; Nichols, F. M., The Marvels of Rome (London, 1889), p. 8Google Scholar. I have usually given the more important references to the Mirabilia from these texts, the second being a translation. There are also editions by Parthey and Jordan (Topographie, vol. ii).

^{page 20 note 2} Urlichs, p. 79; Nichols, p. 168.

^{page 20 note 3} ye Solace of Pilgrimes (ed Mills, C. A.: Oxford, 1911), p. 12Google Scholar: porta colina v(e)l colatina. One wonders whether the author, John Capgrave, got it, directly or indirectly, from Gregory.

^{page 20 note 4} Cp. in a late edition of the Mirabilia (Urlichs, p. 142): Porta Collina … per quam itur ad cum qui bodie dicitur mons marus [sic].

^{page 20 note 5} Nichols, p. 166, note 383. Urlichs, p. 186. Lanciani, , L'Itinerario di Einsiedeln Einsiedeln (Monumenti Antichi, i) p. 105Google Scholar. Huelsen, Anon. Eins. p. 30.

^{page 20 note 6} The real Porta Latina was open throughout the middle ages, and was not closed till 1808. Nibby, , Roma Antica, i, p. 148Google Scholar.

^{page 21 note 1} Nibby, , Roma Antica, ii, p. 501 ffGoogle Scholar.

^{page 21 note 2} Nibby, l.c., p. 497.

^{page 21 note 3} Urlichs, pp. 106 and 119; Nichols, p. 79. Jordan (Top. ii, p. 430Google Scholar) says that Petrus Mallius borrowed his account from this source.

^{page 21 note 4} Helbig, , Führer (3rd ed.), i, p. 543, nos. 954, 955Google Scholar. But it seems that these were found on the site of a large ancient building, to which they probably belonged.

^{page 21 note 5} ii, 851–875: mugit et in teneris formosus inambulat herbis, etc. For medieval references to the story see e.g. Flores Historiarum (R.S.), i, 16; Higden, , Polychronicon, i, 168Google Scholar (citing Isidorus XIV, iv), 196; ii, 340. It was, apparently, a favourite subject with the Byzantine ivory-workers of the ninth-tenth centuries, for it appears twice on the Veroli Casket, and also on a fragment in the Salting Bequest, both, in the Victoria and Albert Museum.

^{page 22 note 1} Duchesne, , Lib. Pont. ii, p. 236Google Scholar (Sergius III), and n. 2.

^{page 22 note 2} Marangoni, Cose Gentilesche, p. 337; Nibby, , Rom. Mod. i, p. 257Google Scholar; and cp. Urlichs, p. 117.

^{page 22 note 3} The Castle of Sant'Angelo was sometimes called domus Thiederici. Jordan, ii, p. 430.

^{page 22 note 4} Saresberiensis, Jo., Polycraticus, vi. 15, ed. Webb, ii, p. 40. Florus, ii, 26 (iv, 12).Google Scholar

^{page 22 note 5} Urlichs, pp. 109, 121; Nichols, p. 97. Cp. Lanciani, Ruins and Excavations of Ancient Rome, p. 126.

^{page 23 note 1} Urlichs, pp. 93, 128; cp. 157; Nichols, p. 21.

^{page 23 note 2} Michaelis, in Röm Mitt, vi (1891), p. 14; Helbig, i, p. 547, no. 959, and p. 548, no. 960.Google Scholar

^{page 23 note 3} Suetonius, Nero, 31; Plinius, H.N. xxxiv, 45Google Scholar; Urlichs, pp. 36, 98.

^{page 23 note 4} Urlichs, p 194.

^{page 23 note 5} See the 4th cent. Regionary catalogues—the so-called Curiosum and Notitia—passim: sometimes with a name attached, e.g. Insulam Felicles (Reg. ix; Urlichs, p. 14.

^{page 23 note 6} E.g. Urlichs, p. 55 (8th cent.): Insulae qui inter vicos sunt horti. Id. p. 80,1.4, and 95,1.18, etc.

^{page 23 note 7} By way of illustrating the confusion of ‘Rhodes’ and ‘Herod,’ Mr. G. F. Hill reminds me that the legend ΡΟΔΙΟΝ (taken for [Η]ΡΟΔΙΟΝ) on coins of Rhodes was perhaps one of the reasons why specimens were identified in the fifteenth century with the Thirty Pieces of Silver. See his article in Archaeologia, lix (1904), pp. 246–251Google Scholar, reprinted in The Medallic Portraits of Christ (Oxford, 1920)Google Scholar.

^{page 23 note 8} Suetonius, Nero, 31: praecipua caenationum rotunda quae perpetuo diebus ac noctibus vice mundi circumageretur. Cf. Lanciani, Ruins, p. 361. The fanciful description of the Coliseum in the later editions of the Mirabilia (Urlichs, p. 136; Nichols, p. 62) seems to be derived from the same source; as also, probably, the story of the tower of Chosroes mentioned by Graf, p. 100.

^{page 23 note 9} Urlichs, p. 136; Nichols, p. 64.

^{page 24 note 1} Rev. Arch. xliii (1882), p. 26Google Scholar.

^{page 24 note 2} Röm. Mitt, vi (1891), p. 14Google Scholar.

^{page 24 note 3} Graf, p. 148 ff. For the Mirabilia see Urlichs, p. 99, etc.; Nichols, p. 46; Jordan, ii, p. 622. The idea of a collection of statues of the provinces may be derived from a tradition of the existence of such series adorning the monuments of Imperial Rome. The most notable are the reliefs from the so-called Basilica of Neptune (Temple of Hadrian) in the Campus Martius, sixteen of which survive. Mrs. Strong, Roman Sculpture, pp. 243–246; Helbig, i, p. 503, no. 888.

^{page 25 note 1} 12: cum ab hospicio meo duobus stadiis distaret. The ancient Greek measure cannot be meant, as this would make the distance the trifling one of a quarter of a mile. Gibbon suggested that in a Greek letter of Gregory III of the year 727 στάδιαmeans ‘miles’ (ed. Bury, v, p. 259, n. 34).

^{page 25 note 2} Cerasoli, F., Ricerche storiche intorno agli alberghi di Roma dal secolo xiv al xix. Studi e documenti di Storia e Diritto, iv (1893), p. 383Google Scholar.

^{page 25 note 3} Helbig, i, p. 447, no. 803; H. Stuart Jones, Catalogue of the Mus. Capitolino, p. 183. Lanciani, , Storia d. Scavi, i, p. 231Google Scholar, gives the passage from P. Sante Bartoli, Memorie, 27: ‘Si cavo in tempo di Clemente X [1670–76] nell'orto de Signori Stazi ove si scoperse gran parte delli bagni di Agrippina, nelli quali fu trovata una statua di Venere di altezza da 9 palmi [m. 2] quasi che intatta, e anche bella quanto la Venere de' Medici.’ For the ‘Orto degli Stazi’ see Nolli, Pianta di Roma (1748), p. 20.

^{page 25 note 4} Huelsen, i, 11: Sci. Vitalis in vico longo ubi caval(li) opt(imi).

^{page 25 note 3} Though the authority for the statement repeated by the guide-books that the Venus was found walled-up may be wanting, as Helbig remarks (i, p. 447), that statement probably represents a fact, even if it were originally only an inference from the condition of the statue. It may be noted that an untouched Mithraic shrine, containing a perfect image of a lion-headed god, was found walled-up close by. Vacca, Memorie, p. 116, quoted by Lanciani, Scavi, iii, p. 200Google Scholar. Cumont, , Textes et Monuments relatifs aux myslères de Mithra, ii, p. 196Google Scholar, no. 10. Graf (p. 670) refers to Gregory's description (as reproduced by Higden) as illustrating William of Malmesbury's Roman story of the bronze Venus and the ring (Gesta Regum, ii, 205, Rolls Series, vol. i, p. 256). The collection of pagan statues brought from Rome in 1151 by Bishop Henry of Blois to adorn his palace at Winchester may be recalled as another rare instance of the appreciation of ancient sculpture in the Middle Ages. Historia Pontificalis (by John of Salisbury), xl (Mon. Germ. Hist. Script, vol. xx, p. 517)Google Scholar. I owe this reference to Dr. R. L. Poole.

^{page 26 note 1} Urlichs, p. 96; Nichols, p. 39.

^{page 26 note 2} Helbig, i, p. 412.

^{page 26 note 3} Jordan, ii, p. 527: ‘die Trümmer, welche damals “domus Corneliorum” hiessen, die beiden auf dem Bufalinischen Plan an der Nordwestecke (gegen SS. Apostoli) der Thermen Constantins eingetragenen Nischen sind, die zu diesen Thermen gehört haben müssen. Dort standen die jezt vor der Treppe des Senatorenpalastes zu beiden Seiten, der Roma angebrachten Flussgötter.’

^{page 26 note 4} Urlichs, p. 122, 1. 28; Nichols, p. 109.

^{page 27 note 1} Rossi, De in Bull. Arch. Crist., S. 5, i (1890), p. 87Google Scholar and n. 2; Jordan, p. 527; Armellini, Chiese. p. 182. Later the name de Cornutis appears as de Corneliis, either because inscriptions were found belonging to that family (so De Rossi), or, perhaps, because the name was better known than that of the Cornuti.

^{page 27 note 2} Kaibel, Inscr. Gr. Sic. et It. 1786 = C.I.G. 6506: Κορνούτου ἰατροῦ καὶ Ῥουϕίνς θυγ;ατρὸς apparently the epitaph of a freedman or dependent of the family.

^{page 27 note 3} E.g. those in the British Museum, reproduced in Reinach, , Rép. de Reliefs. ii, p. 511Google Scholar. They also appear at the angles of candelabra bases (ibid. pp. 468–472).

^{page 27 note 4} The Greek epitaph just referred to (note 2) seems to show that there was something of the kind.

^{page 27 note 5} But Rossi, De in Bull. Comm. xviii (1890), p. 280Google Scholar, connects a forum Tauri with the name caput Tauri (Jordan, ii. 319Google Scholar), suggesting that it was built by one of the Statilii Tauri, and adorned ‘as was usual at the end of the Republic’ with objects derived from the name, e.g. perhaps a frieze of bulls’ heads.

^{page 27 note 6} Servius on Verg. Aen. iv, 196Google Scholar (Hiarbas … Hammone saius). When Bacchus was suffering from thirst in the Libyan desert, Jupiter sent a ram to point out a spring. Unde factum est Iovis Hammonis ab arenis dicti simulacrum cum capite arietino. There is a herm of Ammon in the Capitoline Museum (Stuart Tones, Catalogue, p. 114), but its history is not traced beyond the Cesi collection. An analogous herm (rather smaller) is in the Lateran Museum (Benndorf-Schoene, p. 266). See also Helbig, i, p. 42. no. 69. A herm of a god would, perhaps, be more appropriate for a private house than a statue, which would rather belong to a temple.

^{page 27 note 7} Aen. iii, 390; viii, 43.

^{page 28 note 1} Papers of the British School at Rome, iv (1907), pp. 247 ffGoogle Scholar; Reinach, , Rép. de Rel. i, p. 380Google Scholar; Helbig, ii, p. 177, no. 1412. Cp. Lanciani, Ruins, p. 200 and fig. 75; Platner, Ancient Rome, p. 318.

^{page 28 note 2} Lanciani, Ruins, p. 198; Middleton, , Ancient Rome, ii, p. 223Google Scholar. The Parian marble sow in the Vatican Museum (Gall. d. Animali: Amelung, , Vat. Mus. ii, p. 373Google Scholar; Helbig, i, p. 113, no. 176) does not agree with Gregory's description, for it has only twelve young ones, and it was found on the Quirinal, at a considerable distance from the Coliseum.

^{page 28 note 3} Urlichs, p. 148. Cp. Graf, p. 75.

^{page 28 note 4} This was the principal entrance to the palace, on the north side. Cf. Lib. Pont. (Duchesne, , i,. pp. 502, 518Google Scholar): under Hadrian I one hundred poor were to be fed daily in porticu quae est iuxta scalam quae ascendit in Patriarchio.

^{page 28 note 5} Petersen's articles on the ‘Lupa Capitolina’ in Klio, viii (1908), p. 440Google Scholar ff. and ix, p. 29 ff. make it highly probable that the wolf belonged to the group mentioned by Cicero in Capitolio as struck by lightning in 65 B.C., and that it had lost the twins long before it was set up at the Lateran. Cp. Helbig, i, p. 562, no. 983.

^{page 29 note 1} Röm. Mitt. vi (1891), p. 13Google Scholar.

^{page 29 note 2} Unless it be the ‘altra figuretta di bronzo’ seen by G. Ruccellai in 1450 beside the ‘lupa pregna’ of the Lateran. Lanciani, , Scavi, i, p. 61Google Scholar. Cf. Annali, xlix (1877), p. 379ffGoogle Scholar.

^{page 29 note 3} C.I.L. vi, 930; Lanciani, , Scavi, i, p. 37Google Scholar.

^{page 29 note 4} Bruns, Fontes Juris Romani (7th ed.), pp. 8, 35.

^{page 29 note 5} Middleton, ii, 179.

^{page 30 note 1} In the Mirabilia (Urlichs, 109; Nichols, 93) the Forum of Nerva appears, but the temple is that of the deified Nerva, no doubt owing to the dedicatory inscription on its front. The survival of the name ‘temple of Pallas’ was, of course, due to the Acta Martyrum.

^{page 30 note 2} Duchesne, , Lib. Pont. i, p. 152Google Scholar, n. 11; Nichols, p. 31, n. 69.

^{page 30 note 3} The description makes it clear that this is not the figure of Minerva still to be seen on the attic of the surviving fragment of the wall of the Forum (‘le Colonnacce’), which is perfect and in relief.

^{page 30 note 4} Lanciani, Ruins, p. 311, referring to Lampridius, Alexander Severus, 28, 6.

^{page 30 note 5} The most important of these views are: Cock, H., Praecipua aliquot Romanae antiquitatis ruinarum monumenta (Antwerp, 1551), pl. v and xGoogle Scholar, reproduced by G. B. Pittoni (Venice, 1561, etc.), pl. 6 and 7 (reversed); du Pérac, E., Vestige dell' antichità di Roma (Rome, 1575), pl. 6Google Scholar; Gamucci, B., Le Antichità delta città di Roma (Venice, 1569), plate facing p. 52Google Scholar. Cp. Ashby, , Topographical Study in Rome in 1581 (Roxburghe Club, 1916) fig. 54, etcGoogle Scholar.

^{page 30 note 6} Urlichs, p. 140 from the Polistoria of Giov. Cavallini (14th cent.) : dicitur bodie a Romanis archa Noe ubi habetur hodie macellum prope turrim Comitum.

^{page 31 note 1} Gregorovius, v, pp. 41 ff, pp. 667 ff.

^{page 31 note 2} De Mas Latrie, Trésor de Chronologic, col. 1189. The suggestion for the explanation of Horreum Cardinalium given above was provided by Mr. W. St. Clair Baddeley.

^{page 31 note 3} E.g Lanciani, Ruins, 310. Jordan (ii, 469) saw that the name belonged to the temple, quoting Signonrili's (15th cent.) localization of Nerva's inscription as in oratorio Nervae in loco qui dicitur. corrupto vocabulo arca Noe, ad honorem Nervae. The similarity of Minerva and Nerva will account for the disappearance of the former.

^{page 31 note 4} Genesis, vi, 21.

^{page 31 note 5} Gregorovius, v, pp. 1–49.

^{page 31 note 6} Lanciani, Ruins, pp. 141 ff.

^{page 31 note 7} C.I.L. vi, 1106: inscription on the arch of Gallienus. Gregory goes on to say that Augustus refused the title of dominus (Qui cum esset dominus urbis et totius orbis appellationem tamen domini omnino vitavit), which may be an allusion to the legend of the vision of Augustus (Graphia, 20: Urlichs, p. 121: Cum Sim mortalis dominum me dicere nolo), unless it is derived from the passage in Orosius (vi, 22): eodemque tempore hic ad quem rerum omnium summa concesserat, dominum se hominum adpellari non passus est; immo non ausus, quo verus dominus totius generis humani inter homines natus est, which Graf suggests (p. 245) may have been the germ from which the story sprang.

^{page 32 note 1} Ruins, pp. 141 ff.

^{page 32 note 2} Ruins, p. 186, with reference to his Comentarii di Frontino, pp. 211, 234. Cp. C.I.L. xv, 7246.

^{page 32 note 3} The words a montanis fontibus per spatium unius diete mean that the sources of the aqueducts were a day's journey from Rome. The dieta seems to have been reckoned as twenty miles (see Ducange), which is approximately the distance from Rome to the hills.

^{page 32 note 4} Lanciani, I. Comentarii di Frontino, pp. 3 f. Frontinus, Aq. i, 4Google Scholar: Ab urbe condita per annos CCCCXLI contenti fuerunt Romani usu aquarum quas aut ex Tiberi aut ex puteis aut ex fontibus hauriebant.

^{page 32 note 5} See the life of Hadrian I (772–95) in the Liber Pontificalis (Duchesne, i) for his restorations of the Aqua Sabbatina (p. 503), Iovia (a branch of the Marcia, p. 504), Claudia (or Lateranensis, p. 504). and Virgo (p. 505), the four ancient aqueducts which were more or less maintained during the earlier Middle Ages (Adinolfi, P., Roma nell' età di Mezzo, i, p. 158Google Scholar). There seems to be an allusion to this in Gregory's words: quare a quatuor partibus urbis per artificiosos meatus Romani ueteres aquas recentes uenire fecerunt. But it is not clear whether the antithesis between ueteres and recentes means that the Romans of ancient times brought fresh (i.e. spring) water to the city, or (as I am inclined to think) that the Romans (of medieval times) have caused the ancient aqueducts to run afresh.

^{page 33 note 1} Gregory's story that Apollo (or Apollonius) set the bath alight cum una candela consecrationis partly recalls the story in Gregory of Tours about the spring at Grenoble, (M.G.H. Script. Rer. Meroving. i, p. 862Google Scholar); Quod si cereum vel tedas admoveas protinus ut flammas adtigerint, comprehenduntur.

^{page 33 note 2} I leave open the question whether the change of Apollonius Tyaneus into Apollo Bianeus be due to Gregory's copy of the ‘Seven Wonders,’or to corruption of his own text. As far as I know, Apollonius left no traces in Rome. It was in the eastern half of the empire that his legend flourished.

^{page 33 note 3} According to Murray's Handbook for Rome (17th ed.), p. 543, the water is sulphurous, with a temperature of 113° Fahrenheit.

^{page 33 note 4} Stuart Jones, Catalogue, Glad, no. 7; Helbig, i, p. 494, no. 878; Ashby, in Papers of the British School at Rome, iii (1906), pp. 117–119Google Scholar.

^{page 33 note 5} Smith, , Dict. of Gk. and Rom. Geography, i,p. 94Google Scholar. Pliny (H.N. xxxi, 10Google Scholar) calls the Aquae Albulae egelidae, but perhaps this only refers to the phenomenon mentioned by Pausanias (iv, 35, 10).

^{page 33 note 6} In the Middle Ages the Campagna was far too unsafe for the existence of bathing establishments. The Aquae Albulae seem to have been abandoned all through the Middle Ages, and were first restored (temporarily) to use in the sixteenth century. Bacci, A., Discorso delle acque albule (RomaGoogle Scholar, 1563), reprinted in A. Capello's article on the same subject in the Giornale Arcadico (Roma), lxxi (1837), p. 67 ff. Cp. Lanciani, Ruins, pp. 36, 37.

^{page 34 note 1} This medieval use of per meaning ‘by,’ ‘alongside of,‘ ‘in front of,’ occurs regularly in the Ordo of Benedict. Cp. Nichols, p. 167, n. 384. See also Ducange.

^{page 34 note 2} Suetonius, Nero, 31: balineae marinis et albulis fluentes aquis.

^{page 34 note 3} The credit for this suggestion belongs to Mr. St. Clair Baddeley. The bath may have been a small affair, and therefore not noticed by ordinary writers. If it existed, it will have been one of the nameless balnea of the second or third Regions recorded by the fourth-century Regionary Catalogues. The charge made for using it seems to show that it was private property.

^{page 34 note 4} Hence it appears as the forma Lateranensis in the Einsiedeln Itinerary.

^{page 34 note 5} As the domain of the Golden House extended to and probably included the Gardens of Maecenas (Sueton., Nero, 31 : a Palatio Esquilias usque. Cp. Lanciani, Ruins, p. 361), the site of the bath mentioned by Gregory may have been within it. See the beginning of the next section.

^{page 34 note 6} Dr, Ashby has pointed out to me that a serious objection to this explanation of Gregory's statement is the probability that the duct would have been choked by the sulphur deposit long before the twelfth century. Against this I can only appeal to the realism and obvious sincerity of Gregory's description. Mr. Baddeley suggests as an alternative source of the sulphur water the spring of Acqua Bollicante (sulphuretted hydrogen), less than two miles from the Porta Maggiore.

^{page 35 note 1} Urlichs, p. 36.

^{page 35 note 2} On the analogy of the (unidentified) Domus Pinciana of Cassiodorus, Variae, iii, 10Google Scholar. Jordan, ii, p. 402.

^{page 35 note 3} One might think also of S.Prisca on the Aventine, with its tenth-century inscription beginning: Haec domus est Aquilae sen Priscae virginis almae (Rossi, De, I.C.R. ii, p. 443)Google Scholar; or of the Horti Aciliorum on the Pincian (Lanciani, Ruins, p. 421). But the locality of either is irreconcileable with any interpretation of Gregory's statement.

^{page 35 note 4} Mayor, , Juvenal, i, p. 92Google Scholar.

^{page 35 note 6} Fronto, Ep. i, 8Google Scholar—ii, 1, ed. C. R. Haines (Loeb Classics, i, p. 122): Horatius Flaccus memorabilis poeta, mihique propter Maecenatem et Maecenatianos hortos meos non alienus. In buildings connected with the so-called ‘Auditorium Maecenatis,’ leaden pipes have been found inscribed with the names Cornelius Fronto (probabty the orator, and a brother), Cornelius Quadratus C.I.L. xv, 7438; Lanciani, in Not. Scav. iii (1877), p. 85Google Scholar; Forma Urbis, t. 23.

^{page 35 note 6} Friedlaender, Juvenal, p. 91. The passage from Sat. i, 12, is quoted by John of Salisbury, Pol. VIII, xiii (Webb ii, p. 320).

^{page 35 note 7} Compare the maps in Lanciani, Ruins, figs. 19 and 150.

^{page 35 note 8} Urlichs, p. 111; Nichols, p. 109.

^{page 35 note 9} Probably the Temple of Serapis. Jordan-Huelsen, i, 3, p. 423; Platner, p. 492. Cp. Lanciani, Ruins, p. 430. The ruin of its front pediment was known as ‘mensa imperatoris.’ Jordan, ii, pp. 425, 527.

^{page 36 note 1} L'Itinerario di Einsiedeln, p. 62 (494).

^{page 36 note 2} UrlichS, p. 105; Nichols, p. 21; Jordan, ii, p. 429.

^{page 36 note 3} Duchesne, , Mélanges, ix (1889), p. 354Google Scholar; Lanciani, L'Itin. Eins. p. 96.

^{page 36 note 4} The almost identical passage in a late manuscript (16th cent.) of the Mirabilia given by Urlichs (p. 136), to which Dr. James has called attention (E.H.R. p. 541), may have been copied by the writer from Gregory, whose text would therefore be known to him. But as Gregory introduces the quotation from Ovid by ut aiunt, both may have come from a common source.

^{page 36 note 5} Polychron. (Rolls Series), i, p. 212. Quoted in E.H.R. p. 533.

^{page 36 note 6} Urlichs, pp. 93, 115, 128, 157; Nichols, p. 19. Cp. Graf, p. 89.

^{page 36 note 7} De Principis Instructione, Dist. I (Works, Rolls Series, viii, p. 98). The same system seems to be followed in the Waverley Annals (Annales Monastici, Rolls Series, ii, p. 129 ff), the succession after Valentinian I, the thirty-eighth (p. 142), being traced through Valens and the Eastern emperors. This reckoning does not seem to agree exactly with that mentioned by Bryce, Holy Roman Empire (ed. 1904), pp. 64, 269.

^{page 36 note 8} As Prof.Bury, remarks (History of the Later Roman Empire, ii, p. 499Google Scholar), ‘it is hard to define at what period the Roman see ceased to be a part of the Roman Empire.’ The process of dissolution was not complete till the coronation of Charles the Great in 800. It is unnecessary here to dwell on the history of the palace in the early Middle Ages, and it is enough to recall the facts that Heraclius was the last emperor to be crowned in it (629), that Constans II was the last to visit Rome (663), and that, as late as 687, Plato, the father of Pope John VII, still held the office of curator palatii (Rossi, De, I.C.R. ii, p. 442Google Scholar). The explanation suggested above would seem to be the origin also of the sixty kings and sixty queens in Rome mentioned in the Venetian version of the romance of Flore and Blancheflore, to which Graf (p. 108) calls attention.

^{page 37 note 1} Middleton, , Ancient Rome, ii, p. 130Google Scholar. Nibby, , Roma Antica, ii, p. 691Google Scholar: ‘Questa mole [viz. rotunda and portico] ha 260 piedi di profondità, compresa la grossezza de' muri, et 190 di diametro maggiore.’

^{page 37 note 2} Ashby, Topographical Study, p. 131, with refs.

^{page 37 note 3} Lanciani, (Bull. Comm. xi (1883), p. 56Google Scholar) and Huelsen, (Röm. Mitth. vi (1891), p. 125Google Scholar) think that the lions probably came from the temple of Isis; but Marucchi, (Bull. Comm. xviii (1890), pp. 322Google Scholar points out that they were rediscovered in front of the Pantheon under Eugenius IV (c. 1444, and Clement VII, 1531–34: Lanciani, , Scav, i, p. 51Google Scholar, ii, p. 30), has argued that they were set up at the entrance of the temple by Augustus or Agrippa to emphasize the fact that, among other things, it was a monument of the conquest of Egypt and the foundation of the Empire. In any case, this mention by Gregory, the earliest we have, shows that they were there in the twelfth century,

^{page 37 note 4} Urlichs, p. 129; Nichols, p. 14; Lanciani, in Not. Scav. ix (1881), p. 275Google Scholar; Jordan-Huelsen, i, 3, p. 590.

^{page 37 note 5} Nichols, p. 15, n. 29. I note that Boni (N. Antologia, 1st Nov. 1906, ‘Leggende,’ p. 36) has tried to connect Dante's famous description (Purgatorio, x, 73 ff.) with a relief on the Arch of Constantine.

^{page 38 note 1} See the remarks of Marucchi, in Bull. Comm. xviii, p. 324Google Scholar, referred to above (p. 37, n. 3).

^{page 38 note 2} Jordan, , Topographie, i, 2, p. 211Google Scholar and n. 45, quoting the Veronese scholiast on Virgil, Aen. vii, 605Google Scholar (Parthosque reposcere signa): huius facti nicae repraesentantur in arcu qui est iuxta aedem divi Iulii. Cp. Mommsen, Res Gestae d. Aug. (2nd ed.), p. 125.

^{page 38 note 3} The foundations of the triple arch discovered in the Forum confirm the representation on the reverse of a coin of 19 B.C. (Cohen, Médailles, i, p. 75Google Scholar, no. 82; Eckhel, , Doctr. Num. vi, p. 101)Google Scholar. The other arch appearing on coins of Augustus, presumably that of 30 B.C., is a single one (Cohen, i, p. 82, no. 123). Yet Lanciani (Ruins, p. 270; Scavi, ii, p. 200) identifies the foundations in the Forum with the arch of 30 B.C, and has suggested (Not. Scav. x (1882), p. 343Google Scholar) that the inscription C.I.L. vi, 873, in honour of Augustus republica conservata, came from it. But the slab (now lost) was too small, being only about 2·70 m. long, whereas the central passage of the arch discovered in 1883 was over 4 m. wide (Plainer, Ancient Rome, p. 253) Its original position is quite uncertain. Cp. Huelsen, in Jahrb. Inst. iv (1889), p. 231, n. 7Google Scholar; C.I.L. vi, 31188a; Beiträge zur alien Geschichte, ii (1902), p. 239Google Scholar.

^{page 38 note 4} Cp. Jordan, i, 2, p. 211: ‘Ob der andere Bogen [of 30 B.C.] wirklich zur Ausführung gekommen ist, ist unbekannt.’ n. 45: ‘Dass er jemals gebaut worden ist, sagt Niemand.’ It might be suggested tentatively that the inscription C.I.L. vi, 873, belonged to a substitute monument in the Forum, intended to give some effect to that part of the decree of 30 B.C. which referred to the place of the memorial.

^{page 39 note 1} Lanciani discussing the discovery of the foundation of the Arcus Pietatis recorded by P. Sante Bartoli (Mem. 113), expresses the opinion that it commemorated the triumph of some fourth or fifth century emperor. If it had belonged to an emperor ‘dei buoni tempi,’ it would without fail have been mentioned by the historians (Not. Scav. ix (1881), p. 276Google Scholar). But such an important Augustan monument as the Ara Pacis is not mentioned in ancient literature, and it was not till the nineteenth century that its remains were connected with the contemporary notices in the Ancyran Monument (Mommsen, Res Gestae, p. 49) and the Calendars (C.I.L. i, p. 320). Cp. Röm. Mitt. ix (1894), p. 171Google Scholar.

^{page 39 note 2} For the form compare the concluding phrases of the inscription on the arch of Severus in the Forum (C.I.L. vi, 1033)Google Scholar: ob rem publicam restitutam imperiumque populi Romani propagatum, etc.

^{page 39 note 3} C.I.L. vi 873. See p. 38, note 3.

^{page 39 note 4} Mommsen, Res Gestae, p. 145 sqq.

^{page 40 note 1} Dio Cass. li, 21. I have already called attention (p. 16, n. 3) to the reminiscence of Virgil's pallentem morte futura (Aen. viii, 709Google Scholar) in the words of the description superba mulier moritura pallescit, whether due to Gregory or to his authority. Horace's famous ode on the same subject (1 Carm. xxxvii) emphasizes just the same two episodes, the flight and pursuit, and the death of Cleopatra; and his closing words, superbo non humilis mulier triumpho may have suggested the superba mulier of our account.

^{page 40 note 2} Petrarch, , Epistolae de rebus famil. VI, ii (ed. Fracassetti, J., Florence, 1859, i, p. 313Google Scholar: letter to Giovanni Colonna): hic Pompeii arcus, haec porticus, quoted by Nibby, , Roma Antica, ii, p. 616Google Scholar. The name of Pompey and the identity of his theatre were too well preserved during the Middle Ages (see Nibby, l.c.) to make it likely that there has been any confusion with the arch of Tiberius iuxta Pompei theatrum (Sueton. Claud. 11). For the same reason we may reject the arch of Severus with its long train of wagons laden with spoils (Reinach, , Rép. de Reliefs, i, 268Google Scholar), recalling a similar representation to which Gregory calls special attention in describing the arch of Pompey (immensum pondus auri et argenti … quod longo ordine Pompeiano triumpho prelatum est).

^{page 40 note 3} Urlichs, p. 111, 1. 5; Nichols, p. 106. Cp. Jordan, ii, p. 515.

^{page 41 note 1} Reinach, , Rép. de Reliefs, i, pp. 250, 251Google Scholar.

^{page 41 note 2} Reinach, , Rép. de Rel. i, p. 251, no. 5Google Scholar; Jahrb. Inst. xxxiv (1919), p. 144Google Scholar.

^{page 41 note 3} This detail seems to come from the Liber de viris illustribus (printed with Aurelius Victor), 42,6 bausto quod sub gemma anuli habebat veneno absumptus est. Gregory says that Hannibal took refuge with king Lircus (not Prusias, as one would expect). Sir J. E. Sandys suggests that possibly the name conceals that of Seleucus, son of Antiochus III, who received Hannibal on his arrival at Antioch.

^{page 41 note 4} Reinach, , Rép. de Rel. i, p. 252, no. 1Google Scholar.

^{page 41 note 5} Urlichs, p. 98; Nichols, p. 25.

^{page 41 note 6} This is an inference from the fact that the column brought in a rent to its owners. Lanciani, Ruins, p. 509; Destruction of Ancient Rome, p. 166.

^{page 42 note 1} Thus, in the ‘Solace of Pilgrimes’ (p. 140) the subjects of the Antonine column are ‘the stories of the batail of troye.’ We may note that Giovanni Cavallino, about the middle of the fourteenth century, believing that the column was erected by the emperor ‘Antonius’ to commemorate the great Romans of the past, includes Fabricius in the list of those whose images and exploits were sculptured on its surface. Graf, p. 114. For Cavallino see Graf, p. 56, and Urlichs, p. 139, who does not include the passage among his extracts.

^{page 42 note 2} e.g. Reinach, , Rép. de Reliefs, i, p. 302, no. 34Google Scholar.

^{page 42 note 3} For the name of the geometrical figure to which he compares the ‘pyramids’—in summitate acute, figuram h'ecnoidis referentes, see note by Sir T. L. Heath on p. 58.

^{page 42 note 4} Cap. xxi, p. 47 and note 2, where Huelsen derives arvagia by corruption from the Almachia of the 15th cent. Anonymus Magliabecchianu's (Urlichs, p. 161), itself a corruption of Naumachia of the earlier Mirabilia (Urlichs, 106; Nichols, 75). But the legend told by Gregory suggests that the word is not a late corruption.

^{page 42 note 6} There is no record of a pyramid near the Porta Latina. The connexion of the name of Augustus with the pyramid of Cestius does not appear elsewhere. The Mausoleum of Augustus cannot be meant, for it was not near any gate. There was another sepulchral pyramid near the Porta Flaminia (Lanciani, Destruction etc. p. 210), but there is no reason to think that it is the monument referred to by Gregory.

^{page 43 note 1} Lanciani, Ruins, p. 552; Urlichs, p. 221. The ball is now in the Palace of the Conservatori (Bronzi, no. 4). Cp. Lanciani, , Scavi, ii, p. 92Google Scholar.

^{page 43 note 2} Suetonius, Divus Iulius, 85. C.I.L. vi, 882. Cp. Nichols, p. 71, n. 129; Graf, p. 228.

^{page 43 note 3} Urlichs, p. 132; Nichols, p. 73, with contigui instead of congeries, thus missing the rhyme. A version quoted by Graf (p. 231, n. 97) from the Harleian MS. 562 shows that the word came from a third line (coming second in the epigram): Si lapides bini dic ubi contigui.

^{page 43 note 4} William of Malmesbury, Gesta Regum, II, xii. Cp. Graf; p. 232; Jordan, ii, p. 373. Others attributed it to Lothair II (d. 1137); Burmann, , Anthol. ii, p. 153Google Scholar.

^{page 43 note 5} Nichols, p. 72, n. 130; A. Neckham, laud. div. sap. (R.S.), 320 (quoted by Dr. James, E.H.R. p. 542). Dr. Ashby points out that when the obelisk was moved in 1589, its supports were no longer in the form of lions (Fontana, Della transportazione dell' obelisco vaticano, p. 15).

^{page 44 note 1} Omont, H. in Bibl. de l'École des Chartes, xliii (1882), p. 49Google Scholar, from the 12th cent. MS. of the Sorbonne.

^{page 44 note 2} Revue Archéologique, 4 S, xx (1912), p. 330Google Scholar. Deonna, G. (Rev. Arch. xxiv, 102Google Scholar) thinks that it may have been an oscillating statue. Cp. Plin. H.N. xxxiv, 75Google Scholar.

^{page 44 note 3} Choiseul-Gouffier, , Voyage pitt. de la Grèce, i (1782), p. 177Google Scholar: ‘On distingue encore dans ses ruines qui sont considérables, les vestiges … d'un theâtre creusé dans la montagne.’ The theatre is barely mentioned in Th. Wiegand's Latmos (Bd. iii, Heft i of Milet, Berlin 1913), p. 13Google Scholar, but a monograph on Heraclea is promised (cp. p. vii). One might suggest that here the ‘crabs’ are not unconnected with the medieval use of cancer for an arch or vault (Ducange).

^{page 44 note 4} Scriptores Rerum Merovingicarum (Mon. Germ. Hist.), i, p. 859.

^{page 44 note 6} Ibid, p. 855.

^{page 44 note 7} B.C.H. ii (1878), p. 175 ffGoogle Scholar. C.I.L. iii, Suppl. i, 6588.

^{page 45 note 1} retinet] manet. Dr. James also omits the que, but the form hodieque is found in post-Augustan Latin. See De Vit. T.

^{page 45 note 2} cunctabundam] contaliundam.

^{page 45 note 3} calamo in] calamoni.

^{page 46 note 1} porta flammea.

^{page 46 note 2} Hos.

^{page 46 note 3} corr. from auro.

^{page 46 note 4} manus.

^{page 46 note 5} composicionis] composicio. Cf. composicionis causam four lines below. T.

^{page 46 note 6} pre] praua.

^{page 46 note 7} euentui.

^{page 47 note 1} obsidi esse.

^{page 47 note 2} strenuissima.

^{page 47 note 3} fidere.

^{page 47 note 4} antemurale] anum murale. T. Cf. Vulg. Isaia, xxvi, I: murus et antemurale.

^{page 47 note 6} magnum.

^{page 48 note 1} Itaque] In

^{page 48 note 2} ipse add. T.

^{page 48 note 3} inde exiuit] nō exuit. qu. in specie cuculina exiuit. T.

^{page 48 note 4} tanta] canta.

^{page 48 note 5} qu. in. R.

^{page 48 note 6} altera dextera.

^{page 48 note 7} uim add. T. Dr. James supplied uirtutem after significabat.

^{page 48 note 8} dare] dere.

^{page 49 note 1} parte.

^{page 49 note 2} Qua de causa multi] Quare nisti causa.

^{page 49 note 3} tamen] cum.

^{page 49 note 4} appariet.

^{page 49 note 5} mentitur] inrētitur.

^{page 49 note 6} defexis.

^{page 49 note 7} dimisso.

^{page 49 note 8} add. T.

^{page 50 note 1} inesse] inc'e' qu. incendere or incendi. R.

^{page 50 note 2} arcus inuolsura (qu. inuoltura)] artus imolsura, i.e. there are four down strokes before o. Read in arcus uoltura ‘in the groining of the vault.’ R.

^{page 50 note 3} habebantur bis.

^{page 50 note 4} m.]] iii.

^{page 50 note 5} Quod] Quid.

^{page 51 note 1} lapidem.

^{page 51 note 2} marmorie.

^{page 51 note 3} sum reuisere] siue uisere. corr. T.

^{page 51 note 4} hii his.

^{page 51 note 5} clari] elarei.

^{page 51 note 6} quoniam] quoniam. Dr. James corrected to quondam, but better to retain quoniam and omit et before Cornuti. T.

^{page 51 note 7} ciuies.

^{page 52 note 1} Hoc] Hoo.

^{page 52 note 2} simulacri.

^{page 52 note 3} orbis] urbis.

^{page 52 note 4} coctilis.

^{page 52 note 6} annis.

^{page 52 note 6} qu. nociuus. T.

^{page 52 note 7} Romani] Romam. corr. from Polychron. I, xxv (R.S. i, p. 222).

^{page 52 note 6} Polychron (loc. cit.) inserts consummare before licuit.

^{page 53 note 1} frontomiana.

^{page 53 note 2} octauiam.

^{page 53 note 3} This line has strayed from its place and has been inserted after detestanda (22, note 1).

^{page 53 note 4} pirope.

^{page 53 note 5} corr. from Panthaleon.

^{page 53 note 6} antonomastice] antonoma site. Polychron I, xxv (R.S. i, p. 214) has autonomastice.

^{page 53 note 7} Romani] roma in.

^{page 53 note 8} augisti.

^{page 54 note 1} See p. 53, note 3.

^{page 54 note 2} Atticum.

^{page 54 note 3} deopatram.

^{page 54 note 4} A stop after similis is wanted, and perhaps a word or two has dropped out.

^{page 54 note 5} mannis.

^{page 54 note 6} quos.

^{page 54 note 7} sc. Tarpeiam.

^{page 54 note 8} corr. from triumphales.

^{page 54 note 9} sc. Pharnace.

^{page 54 note 10} contra] intra.

^{page 54 note 11} A numeral may have dropped out before or after mensem, e.g. tercium. Cp. Plutarch, Pomp. 18. R.

^{page 55 note 1} effecit] effectus.

^{page 55 note 2} prelatum] prelectum.

^{page 55 note 3} corr. from hoc approbatus. R.

^{page 55 note 4} medic̄o the o being expunctuated.

^{page 55 note 5} qu. secundam pre. T.

^{page 55 note 6} auro] aurum.

^{page 55 note 7} certamino.

^{page 55 note 8} et h'. uinci etc. sic. : h' is for Hannibal. Read et Hannibalis vincendi primus spem. R. and T.

^{page 55 note 9} Hannibal] hec.

^{page 55 note 10} Hannibal] h'.

^{page 55 note 11} canis.

^{page 55 note 12} Hannibalem] hec.

^{page 55 note 13} Cipio] c'.

^{page 56 note 1} add R. Dr. James had [et].

^{page 56 note 2} Hannibal] h'.

^{page 56 note 3} a Cipione] accip'.

^{page 56 note 4} gestabat] gestiebat.

^{page 56 note 5} Cipioni] c'.

^{page 56 note 6} h'ecnoidis or h'ecuoidis cod. hemic[o]noidis Sir T. L. Heath, whose note see on p. 58.

^{page 56 note 7} Crescentii] crescentis.

^{page 56 note 8} ferro] f̣ọ ferro.

^{page 56 note 9} obiturum] obitṛn̄

^{page 57 note 1} fātoribus.

^{page 57 note 2} Thus given: Dafnis e. i s. h. u. a. s. n. for c. p. f. i.

^{page 57 note 3} sic: qu. parvo rogo. R. Dr. James suggested primus regum.

^{page 57 note 4} acum] supplied from Polychronicon.

^{page 57 note 5} mare.

^{page 57 note 6} aque.

^{page 57 note 7} Doloseum.

^{page 57 note 8} suis] sciis.

^{page 57 note 9} Piramidis elene.

^{page 58 note 1} lupam.

^{page 58 note 2} in] et.

^{page 58 note 3} corr. from insideatur.

^{page 58 note 4} emittit] remt.

^{page 58 note 5} afforissm.