¹ Frere, S. S., Verulamium Excavations i, Soc. Antiq. Res. Rpts. XXVIII (1972).Google Scholar

² Wacher, J. S., Antiq. Journ. xlii (1962), 9 ff.Google Scholar

³ JRS xlix (1959), 114 ff.Google Scholar

⁴ Biddle, M., Antiq. Journ. xlviii (1968), 281.Google Scholar

⁵ H. Hurst, ibid, lii (1972), 39 ff. and figs. 7-9; 50 and figs. 5-6.

⁶ Jahrbuch der Schweizerischen Gesellschaft für Urgeschichte xlix (1962), 63.Google Scholar

⁷ W. Schleiermacher, Cambodunum-Kempten, Eine Römerstadt Im Allgäu (1972), 60.

⁸ Das Rheinische Landesmuseum, Bonn, Berichte 3/69, 37 f.; H. Hinz in H. Temporini (ed.), Aufstieg und Niedergang der römischen Welt 11.4 (1975), 851 ff. and Abb. 8.

⁹ H. Hellenkemper, ibid. 809 f. and Abb. 19.

¹⁰ The rent at Orange was a denarius per foot of frontage occupied: A. Piganiol, Les Documents Cadastraux De La Colonie Rotnaine D'Orange, XVI^e supplèment à Gallia (1962), 330.

¹¹ The pes Monetalis takes its name from the standard kept in the Temple of Juno Moneta in Rome. As Grenier points out (Manuel d'Archéologie Gallo-Romaine III, i (1958), 35)Google Scholar, examples of the p.M. in use varied by up to 6 or 7 mm, but the average length was that given above. The approximate metric equivalents of the fractions of the actus (120 p.M.) used in the text are:

¹² Hyginus, , De Condicionibus Agrorum (Lachmann, 123, 9).Google Scholar

¹³ G. C. Boon, Silchester: the Roman Town of Calleva (1974). 92 points out the unreliability of some of the published plans of Silchester.

¹⁴ Fr. Hultsch, Griechische und Römische Metrologie, 2nd ed. (1882), 692 ff.

¹⁵ The pes Drusianus may fall under this heading: see Pauly-Wissowa, , Realencydopädie der Klassischen Altertumswissenschaft v (1905), 1740 (Hultsch).Google Scholar

¹⁹ Columella, de re rustica v.i.6, mentions that in Gaul the actus was known as the arepennis. He also refers to the Gallic candetum, an area 100 ft. square in towns and 150 ft. square in the countryside.

¹⁷ E.g. the Gallic libra and parallela employed in Narbonensis: Hyginus, op. cit. (note 12), 122, 7. For other possible local units of measurement in Gaul see Grenier, op. cit. (note n), 35 ff.

¹⁸ A point emphasised by Hyginus, op. cit. 123, 11 ff. and Varro, de re rustica 1.10,1.

¹⁹ Only the more convincing examples from each town will be discussed in the text. Dimensions conforming less exactly to Roman units, although probably still close enough to be significant, are listed in an appendix.

²⁰ House iii, S.: Archaeologia Ivii (1901), 301 ff. and pl. XL; ii,S.: lviii (1902), 121 ff. and pL. VIII.Google Scholar

²¹ House vii,S.: ibid. 138 ff.; viii.S.: ibid. 147 ff.; xiv and xv,S.: lxii (1911), 421 ff. and pl. LX.

²² Houses iii and iv,N.: ibid, lix (1904), 101 ff. and pl. IX.

²³ ibid. Ix (1906), 117 ff. and pl. XVII. The wall lies c. 41·2 m from the western side of the insula and c. 40-6 m from the eastern.

²⁴ The notion of 1 actus p.M. plots gains some support from the insula containing House xxv,S. (ibid, lxiv (1913), 441 ff. and figs 1-4). The building has a north frontage of 33·6 m, although this is increased to 34·8 m by including the interval separating it from House x,S. Since the west wall of the latter appears to correspond to the line of a north-south wall dividing the insula (ibid, lix (1905), 290 ff. and pl. LXVI), it could be that we are dealing with a plot I actus wide in the northwestern corner of the insula. Indeed, since the north frontage of xxv,S. lies c. 37·3 m north of the north wall of House xi,S., it may well be that the plot was I actus square.

²⁵ House xv,S.: see note 21; xvi,S.: Archaeologia lxii (1911), 427 ff.Google Scholar and figs. 11-14; xviii,N.: ibid. 7 ff. and pl. 1; xxvi,N.: information from Dr G. C. Dunning.

²⁶ In xviii,N. the wing linking the strip-buildings together was built along the street-frontage.

²⁷ The possibility of an original ¼ actus (8·8 m) allotment cannot be entirely excluded here, although, as suggested in the text, the variation in width of the buildings seems to be against it.

²⁸ This notion may derive some support from the fact that in Period 1 the central and eastern blocks of xvi,S. had a combined north frontage of 17·6 m, whilst in the corresponding period of xxvi,N. the western and central blocks presented a south frontage of 17·1 m, i.e. in both cases approximating to ½ actus p.M.

²⁹ The natural tendency along a main street-frontage would be for the owner or lessee to pack as many shops as possible into his allotted space and to sub-let them to tenants. A fragment of the Orange Cadastral Inscription records the allocation of six merides to a single manceps, C. Naevius Rusticus, merides ii to vi measuring 18, 34½, 35, 55½ and 75 ft. respectively. The fact that merides ii and iii are grouped together at 69½ ft. and v and vi at 130½ ft., adding up to the possibly significant total of 200 ft. (2 Gallic candeta? see note 16), may provide a parallel for the arrangement suggested at Caerwent. See Piganiol, op. cit (note 10), 331 ff. and fig. 34; pl. XXIX (frag. M.I).

³⁰ Cf. Sites ii and iii at Wroxeter (note 39).

³¹ Archaeologia lix (1905), 297 ff. and pl. LXVI.Google Scholar

³² Ibid, lxii (1911), 413 ff. and pl. LVII.

³³ Ibid, lx (1907), 431 ff. and pl. XL.

³⁴ Ibid, lvi (1899), 230 ff. and pl. XI.

³⁵ Ibid, liv (1894), 200 ff., and pl. XVI.

³⁶ Atkinson, D., Norfolk Arch, xxiv (1932), 105 ff.Google Scholar and pls. VIII and XII. For the most recent plan of the town see J. S. Wacher, The Towns of Roman Britain (1975), 231, fig. 54.

³⁷ For possible Continental examples of a ¾ actus p.M unit see the Appendix on Cologne, Insula HI and Xanten, Insula vii (19).

³⁸ Bushe-Fox, J. P., Excavations on the site of the Roman Town at Wroxeter, Soc. of Antiq. Res. Rpts. iv (1916), 4 ff. and pls. xxx-xxxi.Google Scholar

³⁹ Id., ibid, i (1913). fig. 8.

⁴⁰ R. E. M., and Wheeler, T. V., Verulamium, A Belgic and two Roman Cities, Soc. of Antiq. Res. Rpts. xi (1936), 96 ff. and pl. xxxi.Google Scholar

⁴¹ Ibid., 100

⁴² Ibid., 100 f. and pl. CXX.

⁴³ Frere, S. S., Antiq. Journ. xxxvii (1957), 13 and fig. 6Google Scholar.

⁴⁴ A. W. G. Lowther, ibid. xvii (1937), 42 f. and pl. XXVI.

⁴⁵ Frere, op. cit. (note 1), 98 ff. and fig. 25.

⁴⁶ Boon op. cit. (note 13), 96.

⁴⁷ Archaeologia liv (1895), 440Google Scholar ff. and pl. XLV.

⁴⁸ Indeed, Block iv and House 3 have frontages which measure almost exactly ¼ and ½ actus respectively, calculated in pedes Drusiani. The other buildings and intervals, however, seem to correspond better to the pes Monetalis.

⁴⁹ Note also that Block iii lies 55·2 m (I½ actus?) from the west side of the insula.

⁵⁰ Archaeologia liv (1895), 450 ff. and pl. XLVI.Google Scholar

⁵¹ Ibid. 454 ff. and pl. XLVI.

⁵² Ibid, lviii (1902), 18 ff. and pl. II.

⁵³ Ibid, lx (1907), 431 ff. and pl. XL.

⁵⁴ To avoid too much tedious cataloguing, other Silchester insulae with significant dimensions are listed in the appendix. Among the more noteworthy are Insulae vi, xix and xxviii.

⁵⁵ Richmond, A., Hod Hill ii (1968), 74 ff. and fig. 42bGoogle Scholar; 78 f. and fig. 43; 79 ff. and fig. 45a.

⁵⁶ Fox, A. and Ravenhill, W., Britannia iii (1972), 56 ff.CrossRefGoogle Scholar and figs. 5, 10, 11 and 13.

⁵⁷ Schleiermacher, op. cit. (note 7), 53 and Abb. 2, 25.

⁵⁸ Ibid. 52; Reinecke, P., Germania 13 (1929), 151, 153.Google Scholar

⁵⁹ Schleiermacher, op. cit. (note 7), 60.

⁶⁰ On the pes Drusianus see Hultsch, op. cit. (note 14), 693 f.; Grenier, op. cit. (note 11), 37). The approximate metric equivalents of the fractions of the actus p.D. discussed in the text are:

⁶¹ Cassius Dio liv, 32 f.

⁶² Frere, op. cit. (note 1), 14 ff. I am grateful to Professor Frere for providing detailed measurements from the original plans.

⁶³ Ibid., fig. 8.

⁶⁴ Ibid. 23 ff. and fig. 10; 41 ff. and fig. 11. In Period IIA Room 10 was 4·7 m wide and in IIB Rooms 4 and 7 were c. 4·4 m wide, these dimensions excluding the thickness of the side walls.

⁶⁵ It is noticeable that the following combinations also have 15·1 m frontages: south-east portico-4-6; 6-8-9.

⁶⁶ Ibid. 54 ff. and fig. 15.

⁶⁷ Ibid. 54.

⁶⁸ Ibid. 73 ff. and fig. 19.

⁶⁹ For the archaeological evidence suggesting the single ownership of this group of rooms see ibid. 13; 78.

⁷⁰ Ibid. 10 f.

⁷¹ Frere, S. S. and Joseph, J. K. St., Britannia v (1974), 33 and fig. 17.Google Scholar

⁷² Britannia iv (1973), 288 and fig. 4.Google Scholar

⁷³ Ibid, v (1974), 431 and fig. 10. The plan suggests that the length of the building was greater than that given in the text.

⁷⁴ Glasbergen, W., De Romeinse Castella te Valkenburg Z.H.: De Opgravingen in de dorpsheuvel in 1962, 1967 (1972).Google Scholar

⁷⁵ Ibid. fig. 46.

⁷⁶ The lengths of the Valkenburg barracks given in Fox and Ravenhill, op. cit. (note 56), 86 are incorrect.

⁷⁷ Note also that the Principia had an east frontage of c. 29·1 m (¾ actus p.D.).

⁷⁸ Cf. notes 60, 61.

⁷⁹ Castellum I is now thought to be of Caligulan date: Glasbergen, op. cit. (note 74), 151 f.

⁸⁰ L. Berger in Studien Zu Den Militärgrenzen Roms: Vorträge Des 6. Internationalen Limes-kongresses in Süddeutschland (Beihefte der Banner Jahrbücher 19, 1967), 98 ff. and Beilage.

⁸¹ J. Ewald in Provincialia (eds. E. Schmid, L. Berger and P. Bürgin 1968), 80 ff. has proposed an original subdivision of Insula xxx into twenty blocks 38 ft. square. This theory rests on a very small number of fragmentary timber walls or foundation-trenches (compare Abb. I and 3), on the somewhat arbitrary assignment of these to individual houses (note especially Abb. 3, Houses 5, 6 and 10) and the assumption that the insula was sub-divided in the ratio 5 to 4. The explanation offered here at least has the merit of limiting itself to those walls which could be traced satisfactorily, and it brings the early subdivision of Insula xxx into line with that of the adjacent Insula xxiv.

⁸³ It is worth noting that the existence of a south portico in the early phases of the insula is uncertain: Berger, op. cit. (note 80), 103. The designations of individual houses, used in the text for convenience, are Berger's.

⁸³ Even if the division into individual houses proposed here is incorrect, the timber structures to the north of the most south-easterly building still occupy an area measuring ⅜ actus p.D. from north to south.

⁸⁴ ‘House B’ itself is divided by an east-west wall into two ⅛ actus p.D. units.

⁸⁵ Jahrbuch der Schweizerischen Gesellschaft für Urgeschichte 49 (1962), 56 ff.Google Scholar and Abb. 22-28.

⁸⁶ Abb. 23.

⁸⁷ Ibid. 63.

⁸⁸ Ibid. 61, 63.

⁸⁹ Abb. 24-25.

⁹⁰ Ibid. 62.

⁹¹ Wacher, op. cit. (note 36), 63.

⁹² The north wall of 8a also lies ¾ actus p.M. from the southern edge of the insula.

⁹³ Abb. 26.

⁹⁴ It is also worth pointing out that the plan of block 17/20 somewhat resembles that of block 18/24 in Period IID of the Insula xiv timber buildings at St Albans: see note 68.

⁹⁵ Op. cit. (note 85), 63 f.

⁹⁶ Abb. 22.

⁹⁷ Hellenkemper, op. cit. (note 9), Abb. 1, 19 and 20.

⁹⁸ Schleiermacher, op. cit. (note 7), Abb. 2.

⁹⁹ Hinz, op. cit. (note 8), Abb. 8b.