^{page 121 note 1} ii, pp. 436, foll.

^{page 122 note 1} Owing to its bad preservation, this interesting piece could not be included in the collection of casts, as had originally been hoped.

^{page 123 note 1} J.R.S. i, 161, and plate xx.

^{page 123 note 2} Haverfield, Romanization of Roman Britain, p. 43.

^{page 124 note 1} See recently Hoffmann, Militärgrabsteine der Donauländer, 1905.

^{page 124 note 2} Arch. Zeitung, 1868, 112; Essex Arch. Trans. 1873, pp. 88, 161; Journ. Arch. Assoc., 1870, p. 240; Proc. Soc. Antiq. IV (1860), 271Google Scholar; C.I.L. vii, 90; figured also in a good many works.

^{page 124 note 3} A similar cuirass, with three rows of flaps, is worn by Caelius. The lorica squamata, which appears on the reliefs of Q. Sertorius Festus (Lindenschmidt, Tracht und Bewaffnung, pl. 1, 6), and of T. Calidius Severus (A.E.M. v, 206) must have been a short-lived, or never very usual, fashion.

^{page 124 note 4} The name is attested only by Martial, xiv, 32, whose words “Arma tribunicum cingere dignalatus”, need not be taken to confine its use to the officer; compare Mommsen, Staatsrecht, i, 433, 434.

^{page 124 note 5} The so-called “Antinous” of the Belvedere and its replicas prove that Praxiteles originated the motive. It is reproduced in various heroic statues (Museo Chiaramonti, 175, etc.) as well as in those draped in military fashion.

^{page 125 note 1} Arch. Ael. ix, 167; Eph. vii, 995: figured by Domaszewski, Fahnen im römischen Heere, fig. 85, and many others.

^{page 125 note 2} See, amongst others, Schröder, B.J. cviii–cix, p. 51.

^{page 125 note 3} C.I.L. vii, 68; Archaeologia, xxvii, p. 215, pl. xiv.

^{page 125 note 4} Fahnen, p. 70.

^{page 125 note 5} Cagnat Année épigr. 1906, no. 119. The reading in C.I.L. viii, 9291 (see addenda, p. 1975) is doubtful. See Domaszewski, Rangordnung, p. 55.

^{page 125 note 6} B.J. cxvii, p. 279, pl. 1.

^{page 125 note 7} Fahnen, fig. 89; C.I.L. xiii, 6233.

^{page 128 note 1} Proc. Soc. Antiq. Scot. 1871, p. 109; C.I.L. vii, 1088; Macdonald, Roman Wall in Scotland, p. 304, plates xiv, xlii.

^{page 128 note 2} Roman Wall, pp. 295, 303, 312, 314, 343.

^{page 128 note 3} It may be suggested that as the number 3666½ passus practically equals two thirds of 11 Roman miles, it was intended to divide a section of that length in equal portions between the three units engaged in constructing the Limes.

^{page 130 note 1} Bulletin du Comité, 1896, pl. 13; Münchener Abhandlungen, 1903, pl. x, 3.

^{page 130 note 2} Arch. Journ. xxxviii, 107 (King), 207 (Watkin); Raine, Handbook to York Museum (1891), p. 36; Eph. vii, 925, 927, 929. Now in York Museum.

^{page 132 note 1} For bibliography, see the full account in V.C.H. Somerset, i, 230, ff. Most of what follows is based on that account.

^{page 133 note 1} This has the appearance of having wings attached to the extremity, and might conceivably be a thunderbolt. The hands which crown the signa of the legions, however, are always empty and uplifted.

^{page 135 note 1} The same explanation will serve for the Medusa from Celeia (Cilli), which seems also to be bearded (cf. Catalogo della Mostra Archeologica, Rome, 1911, p. 52)Google Scholar.

^{page 136 note 1} Arch. Aeliana, 1909, p. 322.

^{page 136 note 2} Arch. Aeliana, 1909, p. 311, Revue Arch. 1909, ii, p. 468; Cumont, Revue de l'bistoire des religions, 1910 p. 152, fig. 21, and in Daremberg-Saglio, Dictionnaire des Antiquités, s.v. Sol, p. 1383, fig. 6500.

^{page 136 note 3} Eph. ix, 12.

^{page 137 note 1} Cumont, Textes et Monuments, ii, p. 202, f, no. 18b, fig. 29.

^{page 137 note 2} Revue de l'histoire des religions, loc. cit.

^{page 137 note 3} cf. the coin types of Pontus (Reinach Mithradate Eupator, pl. xi), where the winged horse appears between sun and moon, and the coins of Gallienus, Cohen, nos. 978–981.

^{page 137 note 4} Cumont, , Textes et Monuments, ii, p. 195Google Scholar, no. 8, fig. 20. This relief, which is in the Museo Torlonia, was found in the Colonna gardens near the so-called Torre Mesa, and was therefore, supposed to have belonged to Aurelian's temple of the sun. Apart from the fact that the site of this temple is to be bought elsewhere (cf. Jordan-Huelsen, Topographie der Stadt Rom, iii, 1, p. 457, ff.), the inscriptions on the relief are earlier than the time of Aurelian.

^{page 138 note 1} It may be mentioned that Ritterling has suggested a third explanation of the winged horse, viz. that it is the well-known emblem of Legio II Augusta found on other monuments discovered in Britain, e.g. the relief from Condercum (Bruce, Roman Wall, p. 115; C.I.L. vii, 517). In that case we could only suppose that the relief was a private monument, set up in honour of a commander of this legion, but the radiate crown seems fatal to this hypothesis. The nearest parallel would be the curious sarcophagus found in 1908 on the Via Collatina (Notizie degli Scavi, 1908, p. 234, ff, figs. 5–11) on the face of which are seen combats between Romans and barbarians, while on one of the short sides a pegasus is galloping towards a trophy (fig. 9).

It is natural to suppose that the person buried in this sarcophagus had commanded one of the legions (Legio I Adiutrix, Legio II Adiutrix, Legio II Augusta) which bore the pegasus as their emblem, and it is remarkable that, although the sarcophagus belongs to the Antonine period (like the similar one in the Cortile del Belvedere, no. 39), he is represented as a beardless youth in “heroic” nudity except for a small chlamys.

^{page 138 note 2} Textes et Monuments, i, pp. 85, iii, 301; cf Phil. i, 189, Jul. Or. 4, 147a, Lyd. mens. 4, 17.

^{page 138 note 3} Furtwängler, Antike Gemmen, pl. x.

^{page 138 note 4} Cohen ii, p. 427, n. 182, foll.

^{page 138 note 5} Proc. Soc. Antiq. xviii (1900), 177, ffGoogle Scholar.

^{page 138 note 6} Eph. ix, 998.

^{page 140 note 1} On the Deae Matres see Ihm, Der Mutter- oder Matronen-Kultus und seine Denkmäler (Bonner Jahrbücher 1887), and for geographical distribution of the monuments, Haverfield, in Arch. Aeliana, 1892, 314, ff. Those found in Gaul may be conveniently studied in Espérandieu's Bas Reliefs de la Gaule Romaine (4 vols. at present published). It seems very doubtful, however, whether the numerous single figures described by Espérandieu as Déesses-mères are entitled to that name. It doubtless true that the triad is not universal. At Saintes, for example, either two (Espérandieu, ii, 1317, 1322, 1327, 1329, 1330) or one (ibid. 1326, 1328, 1333, 1334) are represented, and the presence of the child in several examples leaves no doubt regarding the nature of the divinity, but these finds are exceptional (two Matres are represented in a monument found at Poitiers, cf. Daremberg-Saglio, Dictionnaire, fig. 4857) and it seems rash to conclude that any female figure with the attributes of the patera and cornucopiae is to be classed as a Mater, if indeed the word was ever used in the singular. Sometimes, where two are represented, a third has most probably been lost; cf. Espérandieu, iii, 1827 (Autun), 2288 (Entrains).

^{page 140 note 2} Arch. Zeitung, 1876, 61; Baumeister, Denkmäler des klass. Alterthums, ii, p. 892, fig. 966; Daremberg-Saglio, Dictionnarie, fig. 4859, ect.

^{page 139 note 3} Espérandieu, iv, 3377.

^{page 142 note 1} Strong, Roman Sculpture, plate vii, etc.

^{page 142 note 2} Schreiber, Hellenistische Reliefbilder, pl. xxxi.

^{page 142 note 3} Petersen, Ara pacis Augustae, p. 49, ff.

^{page 143 note 1} Haverfield, Archaeologia, lx, pl. 10; Cumont, Textes et Monuments, ii, p. 389, no. 267a, fig. 304; Eph. vii, 816.

^{page 143 note 2} Such dedications were no doubt often made by initiates to commemorate their admittance to a Mithraic grade. Compare the inscription of L. Apronius Chrysomallus, ob gradum persicum dedicavit (Cumont, T. et M. ii, p. 163, no. 496; C.I.L. xii, 1324). For such a use of factus compare Mater nata et facta, C.I.L. xiii, 8244; Cumont, p. 476, itself, however, a little abnormal.

^{page 144 note 1} In J.R.S. i, 163, called crab by a slip of the pen.

^{page 144 note 2} Cumont, T. et M. ii, 273d. Compare the mosaic from Sentinum, ibid. no. 298.

^{page 144 note 3} A small part was found in 1882, the remainder in 1885 in or near Cleveland Street, S. Shields, where was a Roman cemetery. Eph. vii, 1002; Arch. Aeliana, x, p. 311 (cf. p. 249). Both this and the Regina stone are now in the South Shields Museum.

^{page 144 note 4} Eph. iv, 718a; Arch. Aeliana, x, p. 239.

^{page 145 note 1} Eph. ix, 1153a; Proc. Soc. Antiq. London, May, 1912; Arch. Ael. 1912, p. 189.

^{page 145 note 2} Cast at Saint-Germain, no. 23936; no. 23941, also from Sens, is a similar figure without casket. See the catalogue of Sens Museum, pl. 6, 2 and 3.

^{page 147 note 1} Lapid. Sept. 600, 602.

^{page 147 note 2} Lanckoronski, Pisidien, ii, pl. 4; Butler, American Archaeological Expedition to Syria (Architecture, etc.), p. 367; Durm, Baukunst der Römer, ed. 2, pp. 403, 417, figs. 446, 447 and 465. Durm's treatment, however, is inadequate.

^{page 147 note 3} Strzygowski, J.H.S. xxvii, p. 99.

^{page 147 note 4} See Studniczka, Tropaeum Trajani (Abhandlungen der k. Sächs. Ges. 1904), p. 61, figs. 30–32; Baumeister, Denkmäler, fig. 1927.

^{page 147 note 5} Arch. Ael. 1st ser. iv, 68; C. R. Smith, Gent. Mag. Nov. 1851, and Coll. Ant. ii, 193; Lapid. Sept. 418.

^{page 148 note 1} Arch. Aeliana, 1908, p. 236, pl. ii. Now at Corbridge.

^{page 149 note 1} Hay, Drummond, Letter to the Committee …. (Colchester, 1821)Google Scholar; C. Roach Smith, Collectanea Antigua, ii, 37; Essex Archaeological Transactions, iv (1859), 237Google Scholar. Now in the Colchester Museum.

^{page 149 note 2} Usener, De Carminis Iliadis quodam Phocaico, 1875; cf. J.R.S. i, p. 19 (Strong).

^{page 149 note 3} J.H.S. xiv, 109, ff, 113.

^{page 149 note 4} Textes et Monuments, i, 101, f; Arch. Epigr. Mitth. xvii, 24. Cumont, however (Textes et Monuments, ii, p. 440, on no. 330), rightly rejects the Mithraic explanation in the case of the group of a lion and boar from Godesberg and similar monuments.

^{page 149 note 5} cf. J.R.S. i, p. 19, note 1.

^{page 149 note 6} Bruce, Lapid. Sept. p. 40.

^{page 149 note 7} In an example from Châlons (Espérandieu, iii, 2160; Reinach, Répertoire, iii, p. 268, 7) the victim wears the costume of a Samnes. This must be due to the artist's reminiscences of the arena.

^{page 149 note 8} Arch. Epigr. Mitth. xvii, 24, f.

^{page 151 note 1} Haverfield, Catalogue, no. 91.

^{page 151 note 2} Bonner Jahrbücher, lxxvii, pl. 1.

^{page 151 note 3} Local custom counted for much in the choice of types; at Cologne the “harpy” or siren between lions is regularly found. The sphinx between lions (without heads between the claws) occurs on the tombstone of the eques Rufus Sita at Gloucester (C.I.L. vii, 67; Eph. ix, p. 519), while the central position is taken by a pine cone on the aedicula from Maros-Németi (J.R.S. i, p. 19).

^{page 151 note 4} V.C.H. Somerset, i, 259, fig. 34a.

^{page 152 note 1} Archaeologia, lx, pl. 9 (Haverfield); cf. Cumont, Textes et Monuments, ii, p. 389, f, no. 267c, fig. 306.

^{page 152 note 2} Haverfield, Archaeologia, lx, pl. 8; Cumont, loc. cit. no. 267b, fig. 305.