^{page 141 note 1} This paper was read to the annual meeting of the Classical Association in January, 1912, at King's College, London, and is here printed with a few alterations and the addition of some footnotes. Some introductory matter, dealing with the methods of teaching history through archaeological illustration, is here omitted: it will be found in the “Proceedings” of the Association. I have to acknowledge help from Mr. Philip Norman, LL.D. F.S.A. in preparing this paper, but he is, of course, in no way responsible for my conclusions. I should add that since this article was in type, the Clarendon Press has issued Sir L. Gomme's volume on The Making of London which contains seven chapters on Celtic and Roman London (pp. 1-95). I am afraid that, although it appears under the authority of a University Press, I am unable to accept many of the statements in it. Several of them, such as those referring to various Celtic dwellings and to the “territorium” and “pomerium” of Roman London, and to the derivation of Londinium, I had already noted in my article as being in my opinion quite untenable (p. 170 below).

^{page 143 note 1} Westminster, Nichols and Sons, 1870, 1873, 1880.

^{page 143 note 2} London, 1909, pp. 1-146.

^{page 144 note 1} There was direct traffic from the lower Rhine to Colchester, and troops were, it seems, sometimes conveyed directly to northern Britain. See also p. 169 below.

^{page 144 note 2} The late Mr. J. R. Green took, as is well known, a very different view of the strategic character of the English plain, and his books have popularised it. Most recent writers have, however, recognised that he was wrong, and have taken the view given in the text. The one strategic feature of the district which does deserve notice is the fact that the two strips of higher ground, formed of Jurassic and Cretaceous limestones, which run across it, slope abruptly to the north and west and gently to the south and east, and thus positively aid the man, whether invader or merchant, who enters Britain from the south-east, may refer further to the Cambridge Medieval History, i, 367.

^{page 145 note 1} This etymology was suggested by M. d'Arbois de Jubainville and has been accepted by Holder, Bradley, H. (Essays and Studies by members of the English Association, i, 17Google Scholar) and other competent scholars. The name “Londinos” does not actually occur, but it is a reasonable Celtic form. The common derivation from Llyn din is phonetically impossible; had Londinium been connected with these two modern words, it would in Roman times have been “Lindu-dunum.”

^{page 145 note 2} Coloured illustration and description in the British Museum Guide to the Antiquities of the Bronze Age, frontispiece and p. 93: see also Rice Holmes, Ancient Britain, p. 245, fig. 42. Compare a bronze helmet found in the Thames at Waterloo Bridge, Guide, as above, p. 88, fig. 67.

^{page 146 note 1} Some pile-foundations near Walbrook have been interpreted as Celtic pile-dwellings and a couple of skulls found among them have been called trophies of the heads of the pile-dwellers' enemies (Gomme, , Geographical Journal, xxxi (1908), 491Google Scholar). But this idea is an anthropological abberration, see Archaeologia, lx, 182, and Arch. Journ. lx, 137.

^{page 146 note 2} For example, a piece of Samian with the stamp of Ateius (A E. in a foot-shaped label) found in 1841 at London Bridge railway station and now at Bethnal Green (de la Beche, Museum of Practical Geology, 1876, p. 65: omitted in Vict. Hist.). Mr. Walters, H. B., Proc. Cambridge Antiq. Soc. xii, 111Google Scholar (note), wrongly ascribes it to Greenwich, and is, I think, also wrong in assigning it to the Flavian period. The fragment is also noted by Hübner, , C.I.L. vii, 1336Google Scholar, 96, but with incorrect reference and provenance. The Victoria History lists once or twice note “early first-century” wares in London (pp. 96, 140, etc.), but these may be doubtful. However, three pieces now in the British Museum (L. 160, 167, 169), with the stamps of Cornelius, Secundus and Xanthus, might belong to the period A.D. 10-40; compare Loeschke, S., Westfal. Mitteil. v, pp. 174, 182, 188Google Scholar. The first-named seems to have been found in Southwark. Southwark potsherds of such early date would help to explain why Ptolemy puts Londinion in the territory of the Cantii, that is, south of the Thames (Geogr. ii, 3, 12Google Scholar).

^{page 148 note 1} Vict. Hist. p. 30 and plan B; Smith, R. A., Journal of the Royal Soc. of Arts, lix (1910), 114–126Google Scholar, from which fig. 17 is by leave reproduced. A similar theory is maintained by Codrington, Roman Roads (S.P.C.K. 1903), p. 56, who rashly alleges that “there is no doubt about it,” and gives no reasons.

^{page 149 note 1} Tac. Ann. xiv, 33, cognomento quidem coloniae non insigne sed copia negotiatorum et commeatuum maxime celebre.

^{page 149 note 2} Knorr, Südgallische Terra Sig. Gefässe von Rottweil (1912), p. iii, calls them “Ueberbleibsel der imposanten Unternehmungen des Claudius.” It will hardly do to limit them so closely.

^{page 149 note 3} Mr. R. A. Smith, Vict. Hist. p. 42 suggests. that “at the time of the Claudian conquest a legion was posted here to guard the river-passage, but soon passed on to the front” and assigns to its camp certain “vast wall-foundations.” Had there been such a fortress, the place could not have been unwalled in A.D. 60. See further note 3, p. 150.

^{page 150 note 1} Arch, xi, 41-49. Another conjectural area for the earlier Roman London (from St. Paul's to Birchin Lane and from Cheapside to Poultry) was suggested by Mr. A. J. Kempe (Gent. Mag. 1842, i, 267) and has been accepted by Mr. Codrington. It, too, seems to me unproven and improbable.

^{page 150 note 2} No traces of a Roman bridge have yet been found: Archaeologia, lx, 228, and references there given. The oldest mediaeval bridge (eleventh-century) is said by Stow to have been near Botolph's Wharf (see plan, fig. 29).

^{page 150 note 3} This area was first detected by Mr. R. A. Smith, Vict. History, 42 But he explained it as the site of a legionary fortress, a view which seems to me interesting but nevertheless untenable (see note 3, p. 149, above).

^{page 150 note 4} Vict. Hist. 79.

^{page 150 note 5} Vict. Hist. 79. At the corner of Bow Lane and Cannon Street a tile tomb was unearthed in 1839 at fifteen feet depth and the tomb contained a skeleton with a coin stated to be of Domitian's reign. But the attribution seems uncertain, and an inhumation burial of such a date does not seem very likely; see Vict. Hist. 92.

^{page 150 note 6} Vict. Hist. 124.

^{page 151 note 1} See C.I.L. vii, 22 (of which Hübner's expansion is a trifle too long), 28, 1235. No. 28 is misinterpreted in Vict. Hist. 114. It may be added, in respect of no. 28, that, despite the legal prohibition of unions between free women and slave men, the slaves of the Roman state and of some public bodies were allowed, among other privileges, to marry free women: see Mommsen, , Staatsrecht, i, 324Google Scholar and Dessau, Inscr. select. 4983. 6673, 7022 (I owe the references to Prof. Dessau), Hence in London the slave of the Province, Anencletus, has a free woman to wife, Claudia Martina.

^{page 152 note 1} Amm. Marc, xxvii, 8, 7, “Lundinium vetus oppidum quod Augustam posteritas appellavit;” xxvii, 3, 1, “Augusta quam vetcres appellavere Lundinium.” So also on certain coins: see Numism. Chronicle, 1867, pp. 61, 329.

^{page 152 note 2} See the lists of the Council of Aries (Haddan and Stubbs, i, 7).

^{page 152 note 3} Maurice, Numism. Chronicle, third series, 1900. p. 108, Webb, ibid. 1907, p. 46. The evidence for other mints, at Richborough or Colchester or elsewhere, is much weaker than is usually admitted, The London mint began seemingly with Carausius (fig. 18). It is possible that nearly a hundred years earlier Albinus coined here, but this cannot be discussed in this essay.

^{page 152 note 4} Amm. Marc, xx, 1, 3; xxvii, 8, 7; xxviii, 3, 1.

^{page 152 note 5} Not 360 acres, as was lately alleged in an official paper of the London Society of Antiquaries.

^{page 153 note 1} For the bath in Strand Lane see Chas. Knight's, London (1841), ii, 165–167Google Scholar, and a drawing by Thos. Wykeham Archer in the British Museum Print Room; it still survives, as Dr. Norman tells me, but much defaced. Holborn has yielded a mosaic, found “deep under ground” near St. Andrew's church, and small objects: see Grew, N., Musaeum regalis societatis (London, 1681), p. 130Google Scholar; Guildhall Catalogue, pp. 50, 60, 81, etc. A point in Islington north of King's Cross (formerly Battle Bridge) has yielded a bit of tombstone (C.I.L. vii, 26), probably a waif, and some tiles from a building (Gent. Mag. 1842, ii, 144Google Scholar).

^{page 153 note 2} Vict. Hist. 136, 137. An alleged temple of Apollo at Westminster is, I think, a pure fiction.

^{page 156 note 1} Mr. J. E. Price held the bastions to be wholly post-Roman (Bastion of London Wall in Camomile Street, Westminster, 1880Google Scholar). But this seems a quite untenable view (Vict. Hist. 49, 56; Norman, , Archaeologia, lx, 214Google Scholar.)

^{page 156 note 2} Fox, Archaeologia, lii, 615.

^{page 156 note 3} Norman, , Archaeologia, lx, 212Google Scholar. Dr. Norman has kindly given me some details of the 1908 discoveries, which are as yet unpublished.

^{page 156 note 4} Archaeologia, lix, 125, and Vict. Hist. 65.

^{page 156 note 5} Archaeologia, xxix, 152, and lx, 177.

^{page 157 note 1} Vict. Hist. 49; Archaeologia, lx, 183. The conclusion is based chiefly on the evidence of coins found in the Walbrook, the latest of which are said to date from about A.D. 180. I cannot say that this seems to me conclusive, though Albinus might just possibly have built the wall.

^{page 158 note 1} Bonding-tiles appear in the Emissarium of Claudius at the Lago Fucino, in a staircase of Domitian's time at the back of S. Maria Antiqua in Rome, in the Flavian temple at Brescia, in the baths of Trajan at Rome, in Hadrian's Villa at Tivoli and on many other first and second-century sites in Italy. North of the Alps they have been noted in an aqueduct of this period at Lyons (St. Irénée). I know of no case in Britain or northern Gaul or Germany which can be certainly ascribed to so early a date. But the assertion of Burckhardt-Biedermann, (Westdeutsche Zettschrift, xxv, 174Google Scholar) that they belong exclusively to a late date is not tenable on our present evidence.

^{page 158 note 2} Much that has been written about them is demonstrably wrong. Gracechurch Street, for example, which Mr. Roach Smith called the original north and south highway of Londimum, is now known to be mediaeval in origin. See Archaeologia, xxix, 154; lx, 226.

^{page 161 note 1} Gent. Mag. 1835, i, 493, 618Google Scholar; C. R. Smith, Illustr. of Roman London, 65; Bernouilli, , Ikonograpbie, ii, 2, 115, no. 92Google Scholar; Vict. Hist. 110, with good plate.

^{page 162 note 1} For details see Vict. Hist. 109, 127. The tessellated pavement said there (p. 111) to come from the Thames, really came from the approach to London Bridge (Guildhall Catalogue, p. 72, 6).

^{page 163 note 1} See my note in Archaeologia, lx, 43.

^{page 164 note 1} C. R. Smith, 73, and plate xxii; Vict. Hist. 109, and fig. 48.

^{page 165 note 1} Chaffers, , Archaeologia, xx, 543Google Scholar. See also C. R. Smith, Illustrations, 71, and Vict. Hist. 119.

^{page 167 note 1} The lion and stag sculpture is figured in the Vict. Hist. fig. 37, where it is oddly styled an “architectural fragment.”

^{page 168 note 1} Smith, C. Roach, Archaeologia, xxx, 548Google Scholar; Cumont, , Mystères de Mithra, ii, 432Google Scholar. Some bronze statuettes found in the Thames about the same place, though not the same time, representing Apollo, Attis, Mercury, Ganymede and so forth (Archaeologia, xxviii, 40; Vict. Hist. fig. 50) may be decorative objects. If so, they no more prove definite worship than the figures of eastern deities which may be seen to-day in many modern European houses.

^{page 168 note 2} For the relief see C. Roach Smith, Illustrations of Roman London, p. 33, or Vict. Hist. fig. 46. The rude sculpture of three standing figures, now in the British Museum (see Smith, ibid. p. 45, plate vi, I) perhaps portrays the same deities, as Ihm thinks, Mütterkultus, no. 343: see also my list in Arch. Ael. xv.

^{page 168 note 3} Compare my map in Arch. Ael. xv, 320, and Cumont's Mithraic statistics.

^{page 168 note 4} First edited by Price, J. E., Proc. Soc. Antiq. xi, 178Google Scholar, and Watkin, , Arch. Journ. xliv, 126Google Scholar (bad copy); see further Eph. Epigr. vii, 1141, and my note in Arch. Journ. xlvii, 236. It is now in the Guildhall Museum. Mr. Lambert has sent me an excellent photograph of the tile, from which plate xxvi is reproduced. Owing to a slight abrasion of the surface, it is not clear whether Austalis was absent from work for VIII or XIII or XIIII days; but for our present purpose, that is immaterial.

^{page 170 note 1} The idea that there were bishops of London in the fifth and sixth centuries (see e.g. Wakeman, Hist. of Church of England, p. 9, note) seems to rest on Geoffrey of Monmouth and his fictitious bishop Theonus (Hist. brit. xi, 3) and the like, and not on real evidence.