¹ Politics, 7. 10. 1–2 (= 7. 1330 a, b).

² Diodorus, 12. 10. 5–6; Herodotus, 4. 158.

³ Frontinus, de aquaeductibus urbis Romae, 4.

⁴ 10. 4. 1.

⁵ Strabo, 4. 1. 5. Cf. Sherwin-White, A. N., Racial prejudice in imperial Rome (Cambridge, 1967), 1–13.Google Scholar

⁶ Perkins, J. B. Ward, ‘Etruscan engineering’ in Hommages à Albert Grenier, vol. 3 (Brussels, 1962), 1636–43Google Scholar; disproving Deman, E. B. van, The building of the Roman aqueducts (Washington, 1934), 4.Google Scholar

⁷ Burns, A., Technology & Culture xv (1974), 405Google Scholar. The Hellenistic aqueduct at Pergamon (Kastenbein, W., Archäologischer Anzeiger 1960, 178–98) was a single small pipeline whose delivery will only have sufficed to supply one fountain, and not even abundantly at that.Google Scholar

⁸ Vitruvius, de architectura, 8. 6. 1–2; Frontinus, de aquaeductibus, 105–6.

⁹ Perkins, J. B. Ward, Cities of ancient Greece and Italy (London, 1974), 25Google Scholar. e.g. the Etruscan aqueduct at Narce (Potter, T. W., A Faliscan town in south Etruria (London, 1976), 11–2)Google Scholar, and the elaborate system of drainage at Veii (Stefani, E., Notizia degli Scavi Antichita, ser. 8, vii (1953), 43).Google Scholar

¹⁰ Hirschfeld, Y., Israel Exploration Journ., xxviii (1978), 91.Google Scholar

¹¹ Grenier, A., Manuel d'archéologie gallo-romaine, vol. 4 (Paris, 1960), 23.Google Scholar

¹² At Cosa, for example, the thermae were supplied from cisterns (Brown, F. E., Memoirs of the American Academy in Rome xx (1951), 82–4)Google Scholar, as must also have been the case at Bilbilis (Bueno, A. Martin, Hispania Antiqua v (1975), 202–22).Google Scholar

¹³ Wright, R. P. & Richmond, I. A., Catalogue of the Roman inscribed and sculptured stones in the Grosvenor Museum, Chester (Chester, 1955), 48. 199, pl. xliv. 199.Google Scholar

¹⁴ Relieving arches were constructed where the pipes passed under the walls (Bidwell, P. T., The legionary bath-house and basilica and forum at Exeter (Exeter, 1979), 35), so that the pipes were either laid at the same time as the erection of the walls, or were inserted as soon as these had been built.Google Scholar

¹⁵ Frontinus, de aquaeductibus, 78. Strabo (5. 3. 8) exaggerated when stating that almost every house at Rome was supplied with piped water: Martial complained about the lack of it in his own dwelling (Epig., 9. 18. 3–8).

¹⁶ Ep., 2. 17. 25.

¹⁷ ILS 5771; Cotton, M. A., The late republican villa at Posto, Francolise (London, 1979), 5, 27Google Scholar. The examples of British villa aqueducts in Liversidge, J., Britain in the Roman empire (London, 1968), 262Google Scholar, do not represent anything like a complete list, omitting, for example, Kingsweston (Boon, G. C., Trans. Bristol & Gloucs. Arch. Soc. lxix (1950), 10)Google Scholar and Cobham (Tester, P. J., Arch. Cant. lxxvi (1961), 91).Google Scholar

¹⁸ Casado, C. F., Aquaeductos romanos en España (Madrid, 1972).Google Scholar

¹⁹ Grenier, A., op. cit. (note 11), 91–5.Google Scholar

²⁰ Alcantará–C. F. Casado, op. cit (note 18). Narni–Choisy, A., L'art de bâtir chez les romains (Paris, 1873), 460Google Scholar. Cherchel–Leveau, P. & Paillet, J., L'Alimentation en eau de Caesarea de Mauretanie et l'aqueduc de Cherchel (Paris, 1976), fig. 82.Google Scholar

²¹ Grenier, A., op. cit. (note 11), 68–75Google Scholar; Haberey, W., Die römischen Wasserleitungen nach Köln (Dusseldorf, 1971)Google Scholar; Casado, C. F., op. cit. (note 18). Professor Wacher (The towns of Roman Britain (London, 1974), 131), has suggested a c. 32-km aqueduct at Lincoln. This may well be correct, cf. note 74.Google Scholar

²² JRS xxx (1940), 168, pl. xii. 2.Google Scholar

²³ Pócsy, K. Sz., Archaeologiai Ertesito xcix (1972), 15–30Google Scholar; Mócsy, A., Pannonia and Upper Moesia (London, 1974), 159–62.Google Scholar

²⁴ Blondel, L., Geneva vi (1928), 33–55.Google Scholar

²⁵ Duncan-Jones, R. P., JRS liii (1963), 86–9.Google Scholar

²⁶ Brogan, O., Roman Gaul (London, 1953), 157.Google Scholar

²⁷ NH, 34.161. The figure is quoted in a passage concerning solders, and Mr G. C. Boon points out that lead in the mass may have retailed at a lower price. On the other hand, the price of Spanish cinnabar was artificially controlled (NH, 33.118). Limits imposed upon British lead production (NH, 34.164) may imply similar controls. In any case, the prices used here–one third Pliny's level at Lyon and one quarter at Bath-make ample allowance for price differentials.

²⁸ Drinkwater, J. F., Britannia vi (1975), 135–6.Google Scholar

²⁹ Cunliffe, B., Roman Bath, Research Report of the Society of Antiquaries 24 (London, 1969), 126.Google Scholar

³⁰ Green, H. J. M., ‘Roman Godmanchester’ in Rodwell, W. & Rowley, T. (eds.), The ‘small towns’ of Roman Britain BAR 15 (Oxford, 1975), 196, 198.Google Scholar

³¹ Britannia ix (1978), 451Google Scholar; Essex Arch. Hist.³ x (1978), 198.Google Scholar

³² Britannia vii (1976), 343.Google Scholar

³³ Birley, R. E., Arch. Ael.³ xlviii (1970), 111–12Google Scholar; Birley, R., Vindolanda (London, 1977), 65.Google Scholar

³⁴ Cunliffe, B., op. cit. (note 29), 99; 104–5.Google Scholar

³⁵ Thorpe, H., Trans. Birmingham Warwicks. Arch. Soc. lxxiv (1956), 28.Google Scholar

³⁶ JRS 1 (1960), pl. xxii. 1Google Scholar; Wacher, J. S., ‘Yorkshire towns in the fourth century’ in Butler, R. M. (ed.), Soldier and civilian in Roman Yorkshire (Leicester, 1971), 170. I am grateful to Professor Wacher for providing me with additional information about Catterick.Google Scholar

³⁷ Britannia viii (1977), 397.Google Scholar

³⁸ Rivet, A. L. F. & Smith, C., The place names of Roman Britain (London, 1979), 157Google Scholar; Ramm, H., ‘Native settlements east of the Pennines’ in Branigan, K. (ed.), Rome and the Brigantes (Sheffield, 1980), fig. 4.6.Google Scholar

³⁹ Lake–Field, N. H., Proc. Dorset Natur. Hist. Arch. Soc. xci (1969), 188–9Google Scholar. Great Chesters–Bruce, J. C., The Roman Wall, ed. 1 (London, 1851), 257–66Google Scholar; Richmond, I. A., JRS xxxv (1945), 80Google Scholar. Lanchester–Hodgson, J., Arch. Ael.¹ i (1816), 118–21Google Scholar; Steer, K. A., The archaeology of Roman Durham, unpublished doctoral thesis (University of Durham, 1938), 211–23Google Scholar. Bowes–Tomlin, R., Yorks. Arch. Journ. xlv (1973), 181–4Google Scholar. Burrow-in-Lonsdale–Hildyard, E. J. W., Trans. Cumb. & West. Antiq. Arch. Soc.² liv (1954), 99.Google Scholar

⁴⁰ Hall, D. V. & Nickerson, N., Arch. Journ. cxxiv (1967), 73.Google Scholar

⁴¹ Birdoswald–Bruce, J. C., The Roman Wall, 3rd ed. (London, 1867), 261Google Scholar. Bainbridge–Wade, W. V., Proc. Leeds Phil. & Lit. Soc., 7. 1 (1952), 13Google Scholar. Birrens–Christison, D. et al. , PSAS xxx (1895–1896), 99, pl. iv. k–1Google Scholar. The aqueduct at Birrens was a stone channel entering the fort through the north gate (Richmond, I. A., PSAS lxxii (1937–1938), 306–9Google Scholar). The triangular channel was, therefore, part of the distribution-system. It probably supplied the building outside the north-east angle, which may have been the mansio (Joseph, J. K. St, JRS xli (1951), 57Google Scholar). Another triangular ‘drain’ at the villa at Westland (Radford, C. A. R., Somerset Arch. Natur. Hist. lxxiv (1928), 128), was probably also a pipe casing.Google Scholar

⁴² Catterick–JRS 1 (1960), 218, pl. xxii. 1Google Scholar. Corbridge–Forster, R. H., Arch. Ael.³ iv (1908), 272–9Google Scholar; Richmond, I. A. & Gillam, J. P., Arch. Ael.⁴ xxviii (1950), 158–68.Google Scholar

⁴³ Hutchinson, W., History of the county of Cumberland (Carlisle, 1794), 578Google Scholar. The unit was cohors IV Delmatarum (JRS lv (1965), 222. 7).Google Scholar

⁴⁴ Macdonald, G., The Roman Wall in Scotland, 2nd ed. (Oxford, 1934), 330, 444–5, pl. lxxvii. 5.Google Scholar

⁴⁵ Green, H. J. M., op. cit. (note 30), 198.Google Scholar

⁴⁶ Boon, G. C., Silchester, 2nd ed. (Newton Abbot, 1974), 144.Google Scholar

⁴⁷ op. cit. (note 21).

⁴⁸ Wacher, J., Roman Britain (London, 1978), 77.Google Scholar

⁴⁹ JRS lii (1962), 166Google Scholar; Wacher, J., op. cit. (note 21), 401, 443 n.73.Google Scholar

⁵⁰ Thompson, F. H., Arch. Journ. cxi (1955), 106–28Google Scholar; Wood, K. F., Ant. Journ. lxi (1981), 107–10Google Scholar. An earthenware pipeline has also been found running downhill to the east of the lower colonia (Richmond, I. A., Arch. Journ. ciii (1946), 37Google Scholar). This suggests that the aqueduct comprised two pipelines in its Roaring Meg section, one of which ran from there to the lower colonia, and the other to supply the distribution-tank near the Newport Gate (Whitwell, J. B., Roman Lincolnshire (Lincoln, 1970), 32)Google Scholar. Alternatively, the second pipeline represents a second aqueduct. Certainly a single pipeline cannot have supplied all the colonia's water: much of its delivery will have been consumed by the thermae–perhaps located around Cottesford Place (JRS xlviii (1958), 136)–and there is still no indication as to how the fountain in the lower colonia was suppliedGoogle Scholar (Thompson, F. H., JRS xlvi (1956), 34–6).Google Scholar

⁵¹ Crummy, P., Britannia viii (1977), 100Google Scholar; Current Arch. iv (1974), 237–43.Google Scholar

⁵² Dunnett, B. R. K., Arch. Journ. cxxiii (1966), 31.Google Scholar

⁵³ Hope, W. H. St J., Archaeologia lv (1896–1897), 422Google Scholar; Cotton, M. A., Archaeologia xcii (1947–1948), 129–30.Google Scholar

⁵⁴ JRS xxix (1939), 214Google Scholar; Frere, S. S., Britannia ii (1971), 22.Google Scholar

⁵⁵ Arch. Journ. vi (1849), 398, 408–9Google Scholar; VCH, Hants 1, 286Google Scholar; Biddle, M., Ant. Journ. lv (1975), 115.Google Scholar

⁵⁶ Ashby, T. et al. , Archaeologia lvii (1900–1901), 309; lx (1906–7), 122, 126; lxi (1908–9), 566; lxii (1910–11), 426.Google Scholar

⁵⁷ idem, Archaeologia lix (1904–5), 93–4.

⁵⁸ Kenyon, K. M., Excavations at the Jewry Wall site, Leicester, Research Report of the Society of Antiquaries 15 (London, 1948), 40–1Google Scholar; Wacher, J., op. cit. (note 21), 345Google Scholar. A fountain is also known in Insula XXII (ibid., 347Google Scholar; Fox, G. E., Arch. Journ. xlvi (1889), pl. opp. p. 58).Google Scholar

⁵⁹ Barnes, W. M., Proc. Dorset Natur. Hist. Arch. Soc. xxii (1901), 84–90Google Scholar; Forster, P., ibid. xlvi (1925), 1–13Google Scholar; Richardson, K. M., Ant. Journ. xx (1940), 435–40Google Scholar. This must date from no later than the late-first or early-second century, when water was being distributed in the city (Britannia xiv (1983), 326).Google Scholar

⁶⁰ Pocock, R. W., Trans. Shrops. Arch. Soc. xlvii (1933), 78–80Google Scholar; Webster, G. & Hollingsworth, D., ibid. lvi (1959), 133–7.Google Scholar

⁶¹ Britannia xiv (1983), 323.Google Scholar

⁶² Kenyon, K. M., op. cit. (note 58), 24Google Scholar; Wacher, J., op. cit. (note 21), 342, 441 n. 132Google Scholar; Kenyon, K. M., Archaeologia lxxxviii (1938), 184.Google Scholar

⁶³ Green, C. J. S., Proc. Dorset Natur. Hist. Arch. Soc. xc (1969), 172Google Scholar; Britannia ix (1978), 462.Google Scholar

⁶⁴ RCHM Eboracvm (London, 1962), 51, pl. 21.Google Scholar

⁶⁵ Wheeler, R. E. M., London in Roman times (London, 1930), 39, pl. xiiGoogle Scholar. Distribution is attested in the pre-Boudiccan period (Philp, B. J., Britannia viii (1977), 15)Google Scholar. Cf. Wacher, J., ‘The water-supply of Londinium’ in Bird, J. et al. (eds.), Londiniensia (London, 1978), 104–8.Google Scholar

⁶⁶ Hassall, M. & Rhodes, J., Trans. Bristol & Glos. Arch. Soc. xciii (1974), 26–7, 32Google Scholar; Hurst, H., Ant. Journ. lii (1972), 52Google Scholar; Hurst, H., Ant. Journ. liv (1974), 31.Google Scholar

⁶⁷ Wheeler, R. E. M. & Wheeler, T. V., Verulamium, Research Report of the Society of Antiquaries 11 (London, 1936), 70Google Scholar; Kenyon, K. M., Archaeologia lxxxiv (1934), 216Google Scholar; Lowther, A. W. G., Ant. Journ. xvii (1937), 41Google Scholar; Frere, S. S., Ant. Journ. xxxix (1959), 16Google Scholar; Frere, S. S., Ant. Journ. xlii (1962), 150Google Scholar. The aqueduct was still functioning c. 450–470, when a wooden pipeline was laid (Frere, S. S., Verulamium Excavations 2, Research Report of the Society of Antiquaries 41 (London, 1983), 226).Google Scholar

⁶⁸ Bede, , Vita S. Cuthberti, 27.Google Scholar

⁶⁹ Ferguson, C. J., Trans. Cumb. West. Arch. Soc.¹ xii (1893), 356Google Scholar; Hogg, R., Trans. Cumb. West. Arch. Soc.² lxiv (1964), 21–3Google Scholar; Charlesworth, D., Arch. Journ. cxxxv (1978), 120Google Scholar. An impressive settling-tank is also known (Lysons, D. & Lysons, S., Magna Britannia iv (London, 1816), ccviiGoogle Scholar), but the presence of pointed arches–which appear to have been developed in the late-Roman period (Hirschfeld, Y., op. cit. (note 10), 92)Google Scholar–shows that it is probably mediaeval. It may, however, be a mediaeval rebuild of a Roman settling-tank, for it is not unlike one of the settling-tanks of the Aqua Virgo (Deman, E. B. van, op. cit. (note 6), 173).Google Scholar

⁷⁰ Wacher, J. S., Ant. Journ. xli (1961), fig. 1Google Scholar; Wacher, J. S., op. cit. (note 21), 300.Google Scholar

⁷¹ Bennett, P., Arch. Cant. xcv (1979), 271Google Scholar. Professor Wacher has suggested that a drain in Rose Lane (Frere, S. S., Arch. Cant. lxviii (1954), 118–20Google Scholar) may have conveyed water from the aqueduct (op. cit. (note 21), 185).Google Scholar

⁷² Down, A., Chichester excavations III (Chichester, 1978), 79, 144.Google Scholar

⁷³ James, T., Carmarthen: an archaeological and topographical survey, Carmarthenshire Antiquarian Society Monographs 2 (Carmarthen, 1980), 17, fig. 2.3.Google Scholar

⁷⁴ F. H. Thompson, op. cit. (note 50). It is virtually certain that no Roman pump was capable of pumping water uphill from the Roaring Meg to the colonia–cf. Wacher, J., op. cit. (note 21), 129–31Google Scholar; Smith, N. A. F., Hist. Technol. i (1976), 70 n.37. It was for this reason that Dr Smith suggested that the colonia was supplied by a chain of pots (halysis) lifting water to the top of a high wooden water-tower, from which the pipelines were supplied. Professor Wacher, on the other hand, believes that the colonia was supplied from a different source to the Roaring Meg (cf. note 21), with the substruction designed to support the pipelines over marshy ground. This seems the most satisfactory of the alternatives, and implies that Lincoln may have been supplied by the longest aqueduct yet known in Roman Britain.Google Scholar

⁷⁵ Grenier, A., op. cit. (note 11), 41–55, 160–3, 171–80Google Scholar; Montauzon, C. Germain de, Les aqueducs antiques de Lyon (Paris, 1873). Corbridge was supplied by a channel-aqueduct (cf. note 42), but this must be classed as a military installation, in which category it has a number of parallels (cf. note 106).Google Scholar

⁷⁶ Implicit in Vitruvius, de architectura, 8. 1.

⁷⁷ de re rust., 1. 17.

⁷⁸ Dixey, F., A practical handbook of water supply (London, 1931), 61Google Scholar. Untreated river water was, of course, provided until modern times–e.g. from the Dee at Chester (Shrubsole, G. W., Journ. Chester Arch. Soc.² v (1895), 32).Google Scholar

⁷⁹ Hope, W. H. St J. & Fox, G. E., Archaeologia lix (1904–1905), 362–3Google Scholar; Boon, G. C., op. cit. (note 46), 88Google Scholar. Similar, but more robust, wooden ‘flashes’ were employed on London's New River aqueduct to cross depressions and streams (Gough, J. W., Sir Hugh Myddleton (Oxford, 1964), 54–6).Google Scholar

⁸⁰ Econ. Hist. Rev.² xviii (1965), 31.Google Scholar

⁸¹ C. F. Casado, op. cit. (note 18).

⁸² Frere, S. S., Britannia (London, 1967), 257.Google Scholar

⁸³ Grenier, A., op. cit. (note 11), 180–91.Google Scholar

⁸⁴ The only enclosed channel known in a municipal context is at Winchester, where it measures a derisory 0.36 m by 0.23 m. I use ‘high-delivery’ in the technical sense–viz., delivering much water.

⁸⁵ Braudel, F., The structures of everyday life (trans. Reynolds, S.; London, 1981), 229–30.Google Scholar

⁸⁶ Wacher, J., op. cit. (note 21), 368, 375, 319–20.Google Scholar

⁸⁷ Enterkin, H. & Reynolds, G., Estimating for builders and surveyors (2nd ed.; London, 1978), 91–3.Google Scholar

⁸⁸ Rees, S. E., Agricultural implements in prehistoric and Roman Britain BAR 69 (Oxford, 1979), 321–6, 417–18.Google Scholar

⁸⁹ Watson, G. R., The Roman soldier (London, 1969), 91.Google Scholar

⁹⁰ This is a purely arbitrary figure. Dr R. P. Duncan-Jones has guessed that an agricultural labourer in first-century Italy might have been paid c. HS 1·5 per day (The economy of the Roman empire (Cambridge, 1974), 54)Google Scholar. Wage-rates will, in any case, have varied from province to province, and from area to area within a province–e.g. Hunt, E. H., Regional wage variations in Britain, 1850–1914 (Oxford, 1973)Google Scholar; Bowley, A. L., Wages in the United Kingdom in the nineteenth century (Cambridge, 1900).Google Scholar

⁹¹ CIL 13. 596, 600.Google Scholar

⁹² Pliny, Ep., 10. 37. 1.

⁹³ IGGR 3.804. This was presumably spent on the known second-century aqueduct (Perkins, J. B. Ward, PBSR xxiii (1955), 115–23).Google Scholar

⁹⁴ Philostratus, VS, 548–9. It was paid for by the father of Herodes Atticus, and was regarded as an extremely expensive, ultra-prestige project.

⁹⁵ Dig., 50. 1. 17, 22, 37, 38; 50. 4. 3. 3.

⁹⁶ CIL 2. 5439, ch.98.

⁹⁷ Historical Manuscripts Commission, The manuscripts of the House of Lords ( = Third report) (ed. Riley, H. T.; London, 1872), 345.Google Scholar

⁹⁸ Brunt, P. A. (ed.), The Roman economy (Oxford, 1974), 5.Google Scholar

⁹⁹ Richmond, I. A., Arch. Journ. ciii (1946), 35–6Google Scholar; Whitwell, J. B., op. cit. (note 50), 30–4.Google Scholar

¹⁰⁰ Fullbrook-Leggatt, L. E. W. O., Trans. Bristol Glos. Arch. Soc. lv (1933), 82Google Scholar; Richmond, I. A., Trans. Bristol Glos. Arch. Soc. (note 99), 70–1.Google Scholar

¹⁰¹ Stanier, R. S., JHS lxxiii (1953), 69–71.Google Scholar

¹⁰² Boon, G. C., op. cit. (note 46), 56.Google Scholar

¹⁰³ CIL viii. 28065; 13. 5416, 7.Google Scholar

¹⁰⁴ CIL v. 5262Google Scholar; Duncan-Jones, R., CIL (note 90), 30–1, 160. 468.Google Scholar

¹⁰⁵ Columella, de re rust., 3. 3. 8.

¹⁰⁶ Wacher, J., op. cit. (note 21), 230–3, 338–43, 362–6Google Scholar. For hypothetical, but plausible, figures for the cost of providing a city's public buildings see MacMullen, R., Roman social relations (New Haven, 1974), 144–5.Google Scholar

¹⁰⁷ Dalginross–Macdonald, G., PSAS lxxiii (1938–1939), 253Google Scholar. High Rochester–Bruce, J. C., Handbook to the Roman Wall (2nd ed.; Newcastle, 1884), 301Google Scholar. Beckfoot–Robinson, J., Trans. Cumb. West. Arch. Soc.¹ v (1881), 145Google Scholar. Piercebridge–Keeney, G. S., Trans. Archit. Arch. Soc. Durham Northumb. ix (1939–1943), 60Google Scholar; Keeney, G. S., ibid. x (1946–1954), 306.Google Scholar

¹⁰⁸ The recovery of an iron pipe-collar (Manning, W. H., Catalogue of Romano-British ironwork in the Museum of Antiquities, Newcastle upon Tyne (Newcastle, 1976), 40. 151, fig. 24. 151Google Scholar) and the presence of a fountain (Lap. Sept., 170, 234) show that it was probably supplied by an aqueduct.Google Scholar

¹⁰⁹ Richmond, I. A., The Roman fort at South Shields: a guide (Newcastle, 1953), 11.Google Scholar

¹¹⁰ ‘Military aqueducts in Roman Britain’, Arch. Journ. forthcoming.

¹¹¹ cf. Richmond, I. A. & McIntyre, J., PSAS lxxiii (1938–1939), 122, 134, for the resulting eavesdrips caused by the action of dripping water.Google Scholar

¹¹² Bosanquet, R. C., Arch. Ael.² xxv (1904), pl.Google Scholar

¹¹³ Thompson, F. H., Roman Cheshire (Chester, 1965), 52.Google Scholar

¹¹⁴ e.g. Newstead, R., Journ. Chester Arch. Soc.² xxxiii (1939), 50–1.Google Scholar

¹¹⁵ Birley, E. & Charlton, J., Arch. Ael.⁴ xi (1934), 192, pl. xxviiiGoogle Scholar; Birley, E. & Keeney, G. S., ibid. xii (1935), pl. xxii.Google Scholar

¹¹⁶ The latrine, and presumably also the bath-house, of the large courtyard-house in Insula VIII, was evacuated by its own drain conveying waste to the river Severn (J. P. Bushe-Fox, Third report on the excavations on the site of the Roman town at Wroxeter Shropshire 1914 (Research Report of the Society of Antiquaries 4; London, 1916), 12, 14–5).Google Scholar

¹¹⁷ e.g. Waste from Site XIVN emptied into a sump (Ashby, T., Archaeologia lx (1906–1907), 126)Google Scholar, whilst waste from the building underlying XIXS presumably did the same (Ashby, T. et al. , Archaeologia lxii (1910–1911), 441).Google Scholar

¹¹⁸ e.g. A drain in Insula XXVIII emptied into a sump (Hope, W. H. St J., Archaeologia lxi (1908–1909), 201)Google Scholar. Waste from the industrial processes–probably tanning–carried out in Insulae XIX (Hope, W. H. St J. & Fox, G. E., ibid. lvi (1898–1899), 235–7) and XXXIVGoogle Scholar (Hope, W. H. St J., ibid. lx (1906–1907), 445–9), was presumably carried off in wooden flashes, which will again have been constructed as a private initiative.Google Scholar

¹¹⁹ Arch. Journ. xl (1883), 319Google Scholar; Venables, E., ibid. xli (1884), 320Google Scholar. Cf. Richmond, I. A., ibid. ciii (1946), 36.Google Scholar

¹²⁰ Duncan, P. M., Trans. Essex Arch. Soc. i (1858), 210–28Google Scholar; Hull, M. R., Roman Colchester (Research Report of the Society of Antiquaries 20; London, 1958), 84, 87–91Google Scholar. The city must, however, have lacked a complete network of stone drains, for at least one major road was flanked by a gulley (Dunnett, B. R. K., Trans. Essex Arch. Soc.³ iii (1971), 39).Google Scholar

¹²¹ Frere, S. S., Ant. Journ. xlii (1962), 154Google Scholar; Richardson, K. M., Archaeologia xc (1944), 85–6.Google Scholar

¹²² Webster, G., Trans. Bristol Glos. Arch. Soc. lxxviii (1959), 53Google Scholar; Brown, P. D. C. & McWhirr, A. D., Ant. Journ. xlvii (1967), 196Google Scholar; Brown, P. D. C. & McWhirr, A. D., Ant. Journ. xlix (1969), 231.Google Scholar

¹²³ Aerial photographs of Caistor suggest the presence of road gutters or drains (Joseph, J. K. St, JRS li (1961), 132, pl. x.1Google Scholar; Wacher, J., op. cit. (note 21), 233Google Scholar). The presence of a timber-lined sump by the Forum (S. S. Frere, op. cit. (note 54)), which dates from c. 150-160 or later, suggests that any drainage-system was of late date–possibly as late as the road gutters at Silchester, which are third-century (Boon, G. C., op. cit. (note 46), 89).Google Scholar

¹²⁴ Whitwell, J. B., The Church Street sewer and an adjacent building, The Archaeology of York, 3.1 (York, 1976).Google Scholar

¹²⁵ e.g. Threipland, M., Arch. Camb. cxiv (1965), 131Google Scholar; Boon, G. C., ibid. cxiii (1964), 22.Google Scholar

¹²⁶ e.g. Mumrills (Macdonald, G., PSAS lxiii (1928–1929), pl.Google Scholar), Balmuildy (Miller, S. N., The Roman fort at Balmuildy (Glasgow, 1922), pl. lviii),Google Scholar and Birrens (Christison, D. et al. PSAS xxx (1895–1896), pl. i).Google Scholar

¹²⁷ Buchanan, M., PSAS xxxix (1904–1905), pl.Google Scholar; MacIvor, I. et al. , ibid. cx (1978–1980), 237, 241.Google Scholar

¹²⁸ Stevenson, G. H. & Miller, S. N., PSAS xlvi (1911–1912), fig. 3.Google Scholar

¹²⁹ e.g. Biddle, M. & Quirk, R. N.. Arch. Journ. cxix (1962), 151Google Scholar; Biddle, M., Ant. Journ. l (1970), 312.Google Scholar

¹³⁰ e.g. Down, A. & Rule, M., Chichester excavations I (Chichester, 1971), 9Google Scholar; Down, A., Chichester excavations II (Chichester, 1974), 42Google Scholar; Down, A., Chichester excavations III (Chichester, 1978), 1Google Scholar. At least one drain is known (ibid., 151–2)Google Scholar, but equally not all roads had drainage gullies (Down, A. & Rule, M., op. cit., 27).Google Scholar

¹³¹ RIB 1060, 430.Google Scholar

¹³² The maintenance of discipline could be a problem (Messer, W. S., Classical Philology xv (1920), 158–75). Vitellius punished legio XIII Gemina for its defeat at the first battle of Bedriacum by drafting it to build amphitheatres at Cremona and Bononia (Tacitus, Histories, 2. 67).CrossRefGoogle Scholar

¹³³ Vegetius, 2. 11. Cf. Macmullen, R., Soldier and civilian in the later Roman empire (Cambridge, Mass., 1963), 23–48Google Scholar; Harvard Studies in Classical Philology lxiv (1959), 214–17. Pliny, for example, was advised by Trajan to approach the governor of Lower Moesia for a surveyor (Ep., 10. 42).Google Scholar