¹ For accounts of the fort, see Smith, Tom C. and Shortt, J., The History of the Parish of Ribchester (London, 1890), 1–41Google Scholar and 270: also Watkin, W. T., Roman Lancashire (Liverpool, 1883), 125–163Google Scholar; and Hopkinson, J. H., The Roman Fort at Ribchester (3rd edn., by Atkinson, D., Manchester, 1928)Google Scholar.

² For accounts of the roads: Manchester-Ribchester, Trans. Lit. & Phil. Soc. Lanes. & Cheshire vi–vii; Wigan–Walton-le-Dale, Baines, , Hist. Lancs. iii, (1836) 573Google Scholar; Ribchester–Overborough, Lancs. & Cheshire Antiq. Soc. xxxi, 69–87; Ribchester–Lancaster, op. cit., 221.

³ Ribchester–Ilkley, Bradford Antiquary vi, 44.

⁴ Ribchester–Fylde, Whitaker, T. D., History of Richmondshire (1823) ii, 458Google Scholar; Watkin, op. cit., 70 ff.; JRS xviii, 198.

⁵ For Ribble navigation, see Whitaker, loc. cit., VCH Lancs. vii, 80, and Baines, , Hist. of Lancashire (1836) iii, 377Google Scholar. Stukeley, Iter Boreale 37, mentions ‘Anchor Hill’, just west of the fort, where there had been discoveries of ‘anchors and great quantities of iron pins of all sizes for ships or barges’. Since it seems certain that the river was not navigable to sea-going vessels thus far, the ships must have been rivercraft. For the site at Walton-le-Dale, see Watkin, Roman Lancs. 202–205.

⁶ For cavalry at these sites, see CIL vii, 929 (Stanwix); Eph. Ep. vii, 995 (Corbridge); CIL vii, 423, 427, 429 (Binchester) and ibid. 287, 288 (Lancaster).

⁷ ibid. 218: the credit for first noting that the regiment were Sarmatae goes to J. Horsley, Brit. Rom. 302. The name ‘Bremetennacum’ appears to be formed from an adjectival bremeton (cf. nemeton), connected with βρέμω, with the common Celtic suffix in -nn (cf. Cebenna, Arduenna, Tarvanna, etc.) and the familiar addition in -acum denoting personal property. Bremetenna might well be a stream name.

⁸ The Antonine Itinerary gives Bremetonnacum, see Cuntz, , Itin. Rom. i, 74, 481Google Scholar, 5; the Ravenna Cosmography (431, 3) Bresnetenaci veteranorum; the Notitia Dignitatum, Oc. xl, 54, Bremetenraco.

⁹ ibid.

¹⁰ CIL vii, 229, 230.

¹¹ Yale Classical Studies vi, 99, note 68; cf. Stein, E., Die kaiserlichen Beamten und Truppenkörper im römischen Deutschland unter dem Prinzipàt (Wien, 1932), 238Google Scholar.

¹² cf. Mommsen, , Gesammelte Schriften vi, 111Google Scholar, n. 5, who considers the use as an abuse and compares the ala exploratorum Pomarensium, see op. cit. 109, n. 1. Rowell, P-W, s.v. ‘numerus’, remarks ‘vielmehr können wir erst im 3 Jhdt. die Verwendung eines numerus in einer Truppe einer anderen Gattung feststellen’. He then cites two instances, the numerus Palmyrenorum, which became on the one hand a cohort (CIL iii, 908), and on the other an ala (ILS 9472), and the numerus Palmyrenorum at Coptos under Caracalla, which seems to be the ala VIII Palmyrenorum cited in Not. Dign. Or. xxi, as among the troops of the Duke of the Thebaid. But the first change is so odd that Dessau (ILS, loc. cit.) describes it as mirum, while the second is not of the third century. No support for the Ribchester phenomenon as normal practice can be derived from either example.

¹³ CIL vii, 221.

¹⁴ Illustrated by Watkin, Roman Lancs. 158, and found 200 yds. upstream from the fort in the south bank of the river: see also Proc. Soc. Ant. Lond. ser. 2, vii, 30, with Watkin's illustration.

¹⁵ Rattenbury, , CR lvi, 113Google Scholar; lvii, 67.

¹⁶ Ptolemy, , Geogr. iii, 7Google Scholar, 1, puts the Iazyges north-west of Lake Maeotis (the sea of Azov) as does Strabo, , Geogr. vii, 3, 17Google Scholar = 306 C., based, however, upon sources of the second century B.C., though the name still lingered there, in the Portae Sarmaticae (Ptol. Geogr. v, 9, 11Google Scholar) together with some of the people, allied to the Hiberi (Tac. Ann. vi, 33Google Scholar). Ovid, however, knew of them in west Sarmatia (Ex Ponto iv, 7, 9, and Tristia ii, 191) just as Roman generals met them north of the Danube (ILS 852, 853, 986), as Tacitus also records (Hist. i, 79). But Tacitus also associates them with the Suebi (Ann. xii, 29–30, Hist. iii, 5), as do Pliny, (NH iv, 80Google Scholar) and Ptolemy, (Geogr. iii, 7Google Scholar), the last named calling them the Iazyges Metanastae. For a review of the whole evidence, see E. H. Minns, Scythians and Greeks 120–121, and Syme, R., ‘The Argonautica of Valerius Flaccus,’ CQ xxiii (1929), 130–133Google Scholar.

¹⁷ Hermes xix, 227 (Gesammelte Schriften vi, III).

¹⁸ Hist. iii, 5; cf. Minns, l.c.

¹⁹ Dio Cass., lxxi, 16, Ἰάзυγες ἐς ὁμολογίαν ἧλθον….καὶ ἱππέας εὐθὺς ὀκτακισχιλίους ἐς συμμαχίαν οἱ παρέσχον, ἀφ᾿ ὧν πεντακισχιλίους καὶ πεντακοσίους ἐς Βρεττανίαν ἔπεμτε.

²⁰ Tac. Hist. iii, 5Google Scholar, ‘vim equitum qua sola valent.’

²¹ On the title ‘Sarmaticus’, see Mattingly-Sydenham, , RIC iii, 304, 1154–1157Google Scholar.

²² SHA, Vita Marci, 27.

²³ Tac. Hist. i, 79Google Scholar; cf. Vegetius, , de re mil. ii, 23Google Scholar, quoted in n. 26 below.

²⁴ Cichorius, Die Reliefs der Traianssäule, Taf. 28, scene 37. To these illustrations should be added the Chester stone (Cat. Grosvenor Museum, no. 137, p. 69, illustrated p. 121), which shows a horseman in tightly fitting scale armour (the scales being clear upon leg and arm) with a dragon-pennon. It may be remarked that, if the Chester stone had been a little more weathered, the trooper would have looked naked; and this explains the Ribchester stone, now lost, of a ‘naked horseman’, recorded by Camden (Britannia, edn. 1607, 302), which must in fact have been the weathered figure of a mailed Sarmatian, since the inscription (CIL vii, 230) mentioned the unit.

²⁵ Tac. Hist. i, 79Google Scholar.

²⁶ Vegetius, , de re mil. iii, 23Google Scholar, ‘catafracti equites, propter munimina quae gerunt, a vulneribus tuti, sed propter impedimentum et pondus armorum capi faciles et laqueis frequenter obnoxii, contra dispersos pedites quam contra equites in certamine meliores: tamen aut ante legiones positi aut cum legionariis mixti quando comminus … pugnatur, acies hostium saepe rumpunt.’

²⁷ See n. 19 above.

²⁸ The size of Ribchester is 597 by 443 feet over the ramparts, that is, slightly larger than the forts for alae of Hadrian's Wall (Arch. Ael. ⁴ xix, 6), where quarters were packed tight. As for the size of numeri, ILS 2531 mentions one 600 strong, while De mun. castr. 30 notes analogous detachments of 500, 700, and 900 men.

²⁹ The significance of this piece was first observed by the late R. C. Bosanquet, to whose manuscript notes I owe my own knowledge of it. He never published it.

³⁰ Eph. Ep. vii, 1041.

³¹ Not. Dign. Oc. xl, 21, praef. eqq. catafractariorum, Morbio, under the dux Britanniarum. The position of Morbium is not known.

³² CIL vii, 218.

³³ ibid. 222.

³⁴ Whitaker, History of Whalley 17: Hist. Richmondshire ii, 459.

³⁵ ibid., and Haverfield, , Trans. Hist. Soc. Lancs. & Cheshire liv, 198Google Scholar.

³⁶ It should be noted that, since there are reasons for thinking that the Imperial titles on the stone were erased (p. 24), the workmen may have been confronted by an inscription of which the initial lines were so largely erased as to seem unworthy of preservation. In any case, it is evident from Whitaker's account that they had already broken the stone before they realized that it was inscribed (Hist. Richmondshire ii, 459).

³⁷ Haverfield, l.c.

³⁸ ibid. 199.

³⁹ ibid. 197, where he describes the stone as ‘fragmentary and in part obscure’ and the dedicator as ‘a high Roman military official’, evidently taking leg. on the stone as legatus.

⁴⁰ The flourish is like the lower two-thirds of a pot-hook, with the hook upon the sinister side.

⁴¹ CIL vii, 222; Whitaker, o.c. 460, read praepotenti numini et reginae; Hübner praepositus numeris et rector; the proposal rector was apparently made by McCaul (see Watkin, Roman Lancashire 147), and both he and Watkin read noster for n.

⁴² The correct title, as shown by CIL vii, 287, would have been praeses noster.

⁴³ See n. 35 above.

⁴⁴ This flourish is not a pothook, but a wide V-shaped attachment, with the top of the initial stroke slightly above the top of the L, to which it is attached.

⁴⁵ CIL vii, 218, for collected references.

⁴⁶ e.g. MS. B.M. Cotton Julius, F x, fol. 137 v, and Camden (ed. 1586), 431.

⁴⁷ CIL vii, 218.

⁴⁸ A. von Domaszewski, Rangordnung des römischen Heeres, 108.

⁴⁹ This point was kindly confirmed by Mr. R. P. Wright, after making a squeeze of the lettering and scrutiny of the stone. The traces are clearly visible in a good raking light.

⁵⁰ Whitaker referred ‘reginae’ to Minerva or to the goddess to whom he supposed the stone to have been dedicated. But this is not the natural position in the dedication for such a title, which should come at its head.

⁵¹ Mr. Collingwood's restoration was very kindly communicated to me by Mr. R. P. Wright, with whom I have also discussed the stone, and I desire to make the fullest acknowledgment to my old friend and mentor's inspiration and to Mr. Wright's kindness. I would also mention with gratitude the helpful comments of the editorial committee, for which I am much indebted throughout.

⁵² ILS 2768.

⁵³ ILS 2769.

⁵⁴ Cosm. Rav. 431, 3.

⁵⁵ Kornemann, , Klio xi, 390 f.Google Scholar: ‘Bis jetzt ist wohl angenommen worden, dass in Ägypten ebenso wie die κλῆροι der Ptolemäerzeit, so auch die Veteranengüter der Kaiserzeit mitten zwischen den anderen Aeckern zerstreut gelegen haben…. Wenn wir in Römerreich nach Ähnlichem Umschau halten, gibt es nur eine Parallele zu diesem Sachverhalt: die coloniae, die innerhalb der gallo-römischen Civitäten nachweisbar sind.’ For the difficulties of veterans under this system in Egypt, see Segré, A., JRS XXX, 151–2Google Scholar.

⁵⁶ ILS 6803; see Schulten, , Philologus liii, 650–651Google Scholar.

⁵⁷ ILS 9400 from Sidi Soltan.

⁵⁸ ILS 6885.

⁵⁹ Cuntz, O., Itin. Anton., i, 22Google Scholar (163, 3), 23 (169, 4). For the origin of the name, see Lesquier, L'armée romaine d'Égypte, 392, and, for the normal Egyptian arrangement of veteran land-grants, see Kornemann, l.c.

⁶⁰ Cuntz, o.c., 5 (35, 4).

⁶¹ CIL viii, 4578, 4590, 4605, 4606, 4607; cf. Cagnat, L'armée romaine d'Afrique (1913), 414.

⁶² Gsell, and Graillot, , Mélanges d'archéologie et d'Histoire, xiv, 526Google Scholar, no. 18.

⁶³ Pliny, , NH iv, 45Google Scholar.

⁶⁴ Rowell, , Yale Classical Studies vi, 100Google Scholar.

⁶⁵ ibid. 106 f.

⁶⁶ See Rostovtzeff, Storia economica e sociale dell'impero romano, 250 ff., for a statement of the case. For Britain the two Yorkshire stones, CIL vii, 200, and Eph. Ep. vii, 920, may be taken as typical of the class.

⁶⁷ The bath-house, for example, continued throughout of the small military type, see JRS xvii, 193, fig. 16.

⁶⁸ Schulten, Die römischen Grundherrschaften, 66 for general discussion, 68–9 for quotation.

⁶⁹ This follows from their exceptionally heavy equipment and armour, which the ordinary small native horse of Western Europe would have been unequal to carrying. The Fylde is very like the ‘campos et plana’ of their native land: see Pliny, , NH iv, 80Google Scholar.

⁷⁰ Camden, Britannia (1590), ‘tota est campestris, unde THE FILD, pro FEILD, appellatur’; cf. Baines, , Hist. Lancs. iv, 427Google Scholar, ‘This tract is a champaign country, inclined to peat moss, but yielding excellent crops of wheat and other kinds of grain.’ Also VCH Lancs. ii, 419–436, passim, for the fertility of the Fylde.

⁷¹ See p. 15, n. 4 above.

⁷² JRS xxiv, 204–5, xxvi, 248–250; Arch. Journ. xci, 133, pl. xxvi; Fox, Archaeology of the Cambridge Region, 222–3.

⁷³ Cod. Iustin. xi, 60, 3.

⁷⁴ Tac. Ann. i, 17Google Scholar, ‘ac si quis tot casus vitae superaverit, trahi adhuc diversas in terras ubi per nomen agrorum uligines paludum vel inculta montium accipiant’: cf. Pliny, loc. cit. on Deultum, n. 63, above.

⁷⁵ That is, after twenty-five years' service, if this was the term applied to men in numeri: see Rowell, , Yale Classical Studies vi, 102Google Scholar.

⁷⁶ Bormann, , Römische Limes in Oesterreich ii, 140–1Google Scholar, on CIL iii, 14356, 3a.

⁷⁷ SHA Vit. Alex. Sev., 52, a passage whose significance is fully discussed by Mommsen, in Gesammelte Schriften vi, 210–211Google Scholar.

⁷⁸ CIL iii, suppl. I, p. 2002, no. xc, is restored by Mommsen as praeterea [liberis eorundem] decurionum et centurion[um qui probatis filiis in] provinc(ia) ex se procreatis [milites ibi castel]lani essent, and this is accepted by Nesselhauf, , CIL xvi, p. 118Google Scholar, no. cxxxii. Both authorities date the diploma to the middle of the third century.

⁷⁹ Cod. Iustin. xi, 60, 2 (A.D. 423): ‘quicumque castellorum loca quocumque titulo possident cedant atque deserant … quod si ulterius vel privatae condicionis quispiam in his locis vel non castellanus miles fuerit detentor inventus, capitali sententia cum bonorum publicatione plectatur’.

⁸⁰ ibid., ‘quia ab his tantum fas est possideri castellorum territoria quibus adscripta sunt et de quibus indicavit antiquitas.’ A block grant by Vespasian to 800 soldiers at Emmaus is, however, recorded by Josephus, , Bell. Iud. vii, 217Google Scholar.

⁸¹ JRS xxvi, 227–235, where the question is discussed by Mr. A. H. M. Jones.

⁸² Rostovtzeff, Storia economica e sociale dell'impero romano, 494, on the African castella, quoting Carcopino, , Syria vi, 30 ff.Google Scholar, and Rév. arch. xx (1924), 316 ff.Google Scholar; and ibid. 495–6, on Thracian ἐμπόρια.

⁸³ Hoey, , Yale Class. Studies vii, 189Google Scholar, n. 906, points out that mater castrorum was part of the official titulature of these Empresses, though it had been borne by Faustina Junior, ibid., n. 811; cf. R. Noll, Carnuntum 1885–1935, 13.

⁸⁴ There is no doubt on this point, small though the fragment is. The erasure has been made by rude pecking following the actual lines of the letters, and has broken off the original surface of the stone between the lower bars of the X. Only the lower halves of the letters survive.

⁸⁵ It will be realized that there is no part of their titles into which the letters EX could enter, except Caracalla's pontificate, which is in turn excluded by the fact that the words are here in the genitive case.

⁸⁶ JRS xii, 70–71, no. 44: Fulvianus.

⁸⁷ See Haverfield, Roman Britain in 1914 (Brit. Acad. Suppl. Papers iii) 31, and Cumb. and West. Arch. Soc. Trans. (n.s.) xviii, 223 ff., for the vicani of Chesterholm, and Birley, , Arch. Ael.⁴ xii, 208–211Google Scholar, for other examples. For territoria of auxiliary castella, see ILS 5969 and Eph. Ep. vii, 986, recently discussed in Arch. Ael. ⁴ xxii, 83; cf. JRS xxxiv, 88. Legionary lands (territoria or prata) are more commonly mentioned: see ILS 2454–6, 5968

⁸⁸ ILS 2768, 2769, discussed in Mommsen, Staatsrecht ii³, 1075, n. 2.

⁸⁹ Dessau ILS 4920 = CIL vii, 45. Mommsen, l.c., compares with the cases cited in note 88, but in Strafrecht 312, n. 1, he further cites Pliny, Ep. ad Traj. 97, 98, and Hirschfeld, Sitz.-Ber. Berl. Akad. 1891, 864, on a ἑκατόνταρχος ῥεγιωνάριος from Antioch in Pisidia and BGU 522, for an ἐπὶ τῶν τόπων ἑκατόνταρχος.

⁹⁰ von Domaszewski, op. cit. 107, on praefecti civitatium (CIL ix, 2564, xiv, 2954), and 129 on a praepositus et princeps Iapodum (CIL iii, 14324).

⁹¹ See p. 23, n. 76, above.

⁹² ILS 5949, 5950, 5952, 5953. The work is quite distinct from that of mensores.

⁹³ See Dessau, ILS 4306, 4314, 4315a, 4316, 4317, 4318, 4320, 4322.

⁹⁴ Hoey, , Trans. Amer. Philol. Assoc. lxx, 478Google Scholar.

⁹⁵ ibid. 459: cf. for the older view, von Domaszewski, Die Religion des römischen Heeres 60 and Loeschcke, , Bonner Jahrbücher cvii, 66–72Google Scholar, followed by Macdonald, , PSAS lxvi, 276Google Scholar, and Roman Wall in Scotland (edn. 2), 416–417.

⁸⁶ cf. Excav. Doura, Fifth Season, 221, pls. xxix, xxvii.

⁸⁷ Ehrle und Liebaert, Specimina codicum lat. Vat., p. vii: ‘bifariam dividitur haec scriptura in capitalem et uncialem; capitalis autem in quadratam sive monumentalem et in rusticam sive actuariam.’

⁹⁸ The most readily accessible illustrations of normal documents from military tabularia are Eph. Ep. vii, 456–467, and pl. iii; Amer. Acad. Rome ix, pl. 3; Yale Class. Stud. vii, 23, nn. 16–17; Excav. Doura, Fifth Season, pl. xxxi, 2. Inscriptions based upon this type of writing are, in Britain, CIL vii 1080 (Newstead), ibid. 739 (Carvoran), Eph. Ep. vii 1001, Eph. Ep. ix, 1055: for Italian examples, cf. Inscr. Italiae, xiii, fasc. iii, nos. 85, 86, based upon written elogia.

⁹⁹ It must have been the script tradition of such legionary tabularium that evolved the attached initial flourish to indicate a centurion of the legion, just a very different forms of flourish are used for the lea seals from Brough-under-Stainmore, see CW ² xxxvi, 119–121.

¹⁰⁰ Schultze, Zur Geschichte lateinischer Eigennamen, 436–8, does not cite the name Floridius, but deals with the termination -idius. Floridus occurs as a cognomen in the Veneto, Conway, R. S. and Johnson, S. E., The Prae-Italic Dialects of Italy i (London, 1933), 293Google Scholar.

¹⁰¹ Watkin, Roman Lancs. 131 sqq.

¹⁰² The hole on the top of the stone is a lewis-hole, not a dowel-hole, and was doubtless employed in setting the stone in position when erecting the monument of which it formed part. The upper surface of the stone is then chisel-dressed at the margins with a slight trace of anathyrosis to ensure a tightly fitting joint.

¹⁰³ cf. Thiersch, , Jahrbuch d. k. d. Inst. xxviii, 266–272Google Scholar, for a study of votive columns, illustrating the antiquity of the form and the fact that many other gods besides Jupiter were honoured in this way.

¹⁰⁴ A statue would be more likely to have a plain base, which would not detract attention from the statue itself: this, at least, was the normal practice.

¹⁰⁵ The only other probable example is that cited by MrWright, R. P., JRS xxxiii, 37Google Scholar and pl. ii, right hand bottom illustration. The deity is discussed in Arch. Ael. ⁴ xxi, 207–210. Hübner's suggestion, that there was here a third female figure, is pure guesswork, CIL vii, 218.

¹⁰⁶ The turreted crown to indicate cities seems due to the Hellenistic sculptor Eutychides in his statue of Antioch (Gardner, P., JHS ix, 75–6Google Scholar, pl. v = Toynbee, Hadrianic School, pl. xxx, 1–2, also figuring four derivative statues, pl. xxix). Comparison may be made with the B.M. statue of Alexandria, first published by Toynbee, ibid. 42 and pi. xxiii, 3; also the cities on the south wall of the passage-way of Trajan's Arch at Beneventum (Strong, Roman Sculpture, pl. lxvi, upper figure, or La scultura romana, ii, 198, fig. 114) and the Louvre relief assigned by Toynbee to the early second century (op. cit., pl. xxi, 1; p. 21, n. 4). Turreted crowns for provinces soon followed, as on the Ince Blundel l statue of Cappadocia (Toynbee, op. cit., pl. xxiv) and, on coins of Hadrian, for Asia (ibid., pl. iii, 2–4), Bithynia (iii, 5–9), Cappadocia (iii, 17–20) and Pannonia (vi, 6); on coins of Antoninus Pius for Asia (vii, 9), Cappadocia (vii, 12–15), Spain (viii, 4) and Syria (viii, 14); on coins of Severus for Britain (xii, 3–4) and on coins of Postumus for Gaul (xiv, 8). An earlier instance on coins is provided by the two companion Spanish provinces on Civil War issues of M. Minatius Sabinus of 46–44 B.C. (xv, 7–8). The standardization of the convention is well shown upon the African mosaic from Biregik, now in Berlin (Jatta, Le rappresentanze figurate delle provincie romane, figs. 1 (Britannia), 2 (Gallia), 3 (Hispania), and 5 (Raetia), while its persistence into later antiquity is shown by the figures of Insulae, Hellespontus, Palaestina and Septem Provinciae in the Notitia Dignitatum, comparable with the cities on the south face of the base of the Column of Arcadius in Constantinople (Archaeologia, Ixxii, pl. 17. The elderly veiled figure which we are identifying as Britain on the Ribchester stone might well have been copied from the type on the coinage of Severus (Toynbee, op. cit., pl. xii, 5) of the province veiled but without a crown, combined with the normal turreted convention. It may also be noted that the subdivision of a province occurs thus attired on coins of Antoninus Pius, where Phoenice in Syria is thus shown (op. cit., pl. vii, 8); for Phoenicia, though it had been distinguished from Syria as early as Augustan shown times (cf. Lapis Tiburtinus, ILS 918), was not created an autonomous province until the time of Severus.

¹⁰⁷ Watkin, Roman Lancashire, 133, says that the object is ‘too worn to be recognizable’.

¹⁰⁸ See E. Strong, Roman Sculpture, 217, pl. lxiv, or La scultura romana ii, 195, fig. III, where Trajan, presenting land to veterans, carries a roll representing the grant.

¹⁰⁹ See p. 23, n. 80, above.

¹¹⁰ Toynbee, op. cit., pls. vii, 11, and viii, 3 and 5.

¹¹¹ The action is wholly intelligible if the younger figure had handed her gift and is now having it pressed back into her possession.

¹¹² CW ² xxxvi, 113, 116 : cf. also Newstead, A Roman Frontier Post, 309, pl. lxxxii, 19–20, and p. 2, quoting Milne (1743), A Description of the Parish of Melrose, 6, for the statement ‘ a great deal of lead got and some curious seals’.

¹¹³ cf. ILS 6890, 7100.

¹¹⁴ CIL vii, 423, 1345; Arch. Ael. ⁴ xxi, 207–210.

¹¹⁵ CIL vii, 1345.

¹¹⁶ Arch. Ael. ⁴ xxi, 194–5.

¹¹⁷ Stein, E., Geschichte des spätrömischen Reiches, i, 90Google Scholar.

¹¹⁸ Yale Classical Studies vii, 202–210, especially 209.

¹¹⁹ Arch. Ael. ⁴ xxi, 162–176; in particular, 174–5.