¹ For studies of the architectural school see A. Rivet (ed.), The Roman Villa in Britain (1969). By contrast M. Todd, The Coritani (1973), groups villas into three classes. The third class includes ‘isolated aisled halls’ (p. 77).

² This is, for example, the starting-point of Carandini's social interpretation of Sette Finestre, A. Carandini et al, Sette Finestre: una villa sciavistica nellétruria romana (1985).

³ The most important study to adopt this argument is Wallace-Hadrill, A., ‘The social structure of the Roman house’, PBSR lvi (1988), 43–97. For a more art historical interpretation along similar lines see E. Gazda (ed.), Roman Art in the Private Sphere (1991).Google Scholar

⁴ op. cit. (note 3), 77.

⁵ This definition is similar to that most recently discussed by E. Scott, A Gazetteer of Roman Villas in Britain (1993), 2–7. I have adopted more of a minimal definition, excluding items of more dubious cultural significance such as tiled roofs. Similarly the cultural significance of wall-painting and architectural flourishes such as a niche needs careful interpretation.

⁶ The classic text for this remains C. Stevens, Sidonius Apollinaris and His Age (1933).

⁷ This argument was proposed by Stupperich, R., ‘Some fourth-century British mosaics’, Britannia xi (1980), 289–302.CrossRefGoogle Scholar It has recently received further support from Ling, R., ‘Brading, Brantingham, and York: a new look at some fourth-century mosaics’, Britannia xxii (1991), 147–57.CrossRefGoogle Scholar

⁸ For discussion of these arrangements see Wallace-Hadrill, op. cit. (note 3). The atrium house is traditionally seen as the earlier Republican form of house that was superceded by the peristyle house. However the atrium house is now seen as continuing to be developed in the imperial period. See Dwyer, ‘The Pompeian atrium house in theory and practice’, in Gazda, op. cit. (note 3), 25–48.

⁹ This discussion of the basic layout of a provincial Roman house is based on that of Rebuffat, R., ‘Maisons à pèristyle d'Afrique du Nord: répertoire de plans publiés’, MEFR lxxxi (1969), 659–724.CrossRefGoogle Scholar

¹⁰ For U-shaped panels see Rebuffat, op. cit. (note 9), Althiburos 2, Thrysdrus 1, Utica 2, and Uthina I. Ling, op. cit. (note 7), 153 has proposed to identify such a panel at Brading, but, although it is an unusually wide border, it is not broad enough for dining couches.

¹¹ For the social context of patronage, R. Saller, Personal Patronage in the Roman Empire (1982). The architectural setting of the salutatio at Pompeii is described by Wallace-Hadrill, op. cit. (note 3). The setting of the later Roman salutatio is described by Ellis, S., ‘The Palace of the Dux at Apollonia and related houses’, in Barker, G., Reynolds, J. and Lloyd, J. (eds), Cyrenaica in Antiquity, BAR S236 (1985), 15–25.Google Scholar

¹² For the impact of décor on Roman receptions see the papers collected in Gazda, op. cit. (note 3).

¹³ B. Cunliffe, Excavations at Fishbourne. I. The Site (1971).

¹⁴ Ward-Perkins, J.B., ‘The Roman villa at Lockleys, Welwyn’, Antiq. Journ. xviii (1938), 339–76.CrossRefGoogle Scholar

¹⁵ ibid., 345.

¹⁶ For example in Rebuffat's survey, op. cit. (note 9), Banasa 7, Banasa 10, Volubilis 2, Volubilis 11, Volubilis 16, Volubilis 20, Volubilis 23, Timgad 2, Thrysdrus 5. Outside Africa mention could be made of the site of Stobi where the Theodosian Palace, the House of Psalms, and the House of the Fuller all have this arrangement – J. Wiseman, Stobi: A Guide to the Excavations (1973).

¹⁷ The Theodosian Palace in Stobi is a notable example – Wiseman, op. cit. (note 16), 44–7.

¹⁸ A clear case for the influence of wall-painting is made by B. Bergmann, ‘Painted perspectives of a villa visit: landscape as status and metaphor’, in Gazda, op. cit. (note 3), 49–70.

¹⁹ Ditchley – Radford, C. Raleigh, ‘The Roman villa at Ditchley, Oxon.’, Oxoniensis i (1936), 26–69Google Scholar; Boxmoor – Neal, D., ‘The Roman villa at Boxmoor: interim report’, Britannia i (1970), 156–72CrossRefGoogle Scholar; Newport – D. Tomalin, Newport Roman Villa (1977); Sparsholt – Wilson, D. (ed.), ‘Roman Britain in 1972’, Britannia iv (1973), 271–323.Google Scholar

²⁰ Walthew, C., ‘The town house and villa house in Roman Britain’, Britannia vi (1975), 189–205.CrossRefGoogle Scholar

²¹ K. Branigan, The Roman Villa in South-West England (1975), 46.

²² There is little systematic literature on the stibadium. First-century examples are discussed by E. Prima Ricotti, ‘The importance of water in garden triclinia’, in E. MacDougall (ed.), Ancient Roman Villa Gardens (1987), 135–84. Some idea of the room's derivation is mentioned by Lavin, I., ‘The House of the Lord’, Art Bulletin xliv (1962), 1–27. The most important source for domestic architecture is the publication of a house in Argos: G. Åkerström-Hougen, The Calendar and Hunting Mosaics of the Villa of the Falconer at Argos (1974).Google Scholar

²³ Lavin, op. cit. (note 22), associates the apse with the basilica. For other views of its development see N. Duval, ‘Les maisons d'Apamée et l'architecture “palatiale” de l'antiquite tardive', Colloque Apamee de Syrie (1984), 447–70, and Ellis, S., ‘The end of the Roman house’, AJA xcii (1988), 565–76.CrossRefGoogle Scholar

²⁴ A large number of sigma tables have been found at Salamis in Cyprus, and their catalogue provides a convenient reference point to their various forms: Roux, G., ‘Tables chrêtiennes en marbre découvertes à Salamine’, Salamine de Chypre IV (1973), 133–96.Google Scholar

²⁵ W. Ritter et al, Die Wiener Genesis (1895), folio 17 verso. Other manuscripts cited are shown in K. Weitzmann, Early Christian Manuscript Illumination (1977).

²⁶ Verulamium – R.E.M. Wheeler and T.V. Wheeler, Verulamium: A Belgic Town and Two Roman Cities (1936), 96; London – R. Merrifield, The Roman City of London (1965), 246; Aldborough – D. Neal, Roman Mosaics in Britain (1981), 38–9.

²⁷ Aldborough – Neal, op. cit. (note 26); York – Frere, S.S. (ed.), ‘Roman Britain in 1976’, Britannia viii (1977), 382Google Scholar; Box - Hurst, H., ‘Excavations at Box Roman villa 1967–7’, Wilts Arch. & Nat. Hist. Mag. lxxxi (1987), 19–51.Google Scholar

²⁸ For the Mediterranean parallels see Argos – Åkerström-Hougen, op. cit. (note 22), and K. Dunbabin in W. Slater (ed.), Dining in a Classical Context (1991). The British examples are discussed in more detail below.

²⁹ Histria – N. Duval, ‘L'archéologie chrêtienne en Roumanie à propos de deux livres récents de I. Barnea’, Rev. Arch. (1980), 313–40; Corinth - Daux, G. (ed.), ‘Chronique des fouilles 1965 – Corinthe Lechaion’, BCH lxxxix (1966), 766–70; Djemila, House of the Stuccos - Blanchard-Lemée, M., Maisons à mosaïques du quartier central de Djemila (Cuicul) (1975), 181–95; Carthage, House of the Horses – J. Salomonson, La mosa'ique aux chevaux de l'antiquarium de Carthage (1965). The latter was re-surveyed by the present author in 1988.Google Scholar

³⁰ Wilson, D. (ed.), ‘Roman Britain in 1972’, Britannia iv (1973), 314–15.Google Scholar

³¹ Åkerström-Hougen, op. cit. (note 22).

³² G. Meates, Lullingstone Roman Villa. I. The Site (1979).

³³ The house is published by F. Berti, Mosaici antichi in Italia: Aemilia-Ravenna I (1976). The interpretation offered here is argued in S. Ellis, ‘Power, architecture, and decor: how the late antique aristocrat appeared to his guests?’, in Gazda, op. cit. (note 3), 117–34.

³⁴ For example the Maison du Triclinos at Apamea – J. Balty, La grande mosaíque de chasse du triclinos (1969).

³⁵ As suggested by Toynbee, J., ‘Apollo, beasts, and seasons; some thoughts on the Littlecote mosaic’, Britannia xii (1981), 1–6.CrossRefGoogle Scholar For the layout of the house see Frere, S.S. (ed.), ‘Roman Britain in 1988’, Britannia xx (1989), 315–17.Google Scholar

³⁶ For the Palace of Theodoric see Berti, op. cit. (note 33). The identification of such rooms as ‘ceremonial’ or grand dining halls is discussed in detail by Ellis, op. cit. (note 33). The parallels with the imperial palaces are from Krautheimer, R., ‘Die Decanneacubita in Konstantinopel’, R. Quart. Suppl. iii (1966), 195–9.Google Scholar

³⁷ op. cit. (note 35).

³⁸ Y. Thébert, ‘Private life and domestic architecture in Roman Africa’, in P. Veyne (ed.), A History of Private Life. I. From Pagan Rome to Byzantium (1987), 357–64 with allusions to Meleager. Ellis, op. cit. (note 33), with allusions to Meleager and Bellerophon, including the Ravenna mosaic.

³⁹ RCHM Dorset. I. West (1952), 150.

⁴⁰ op. cit. (note 7), 300.

⁴¹ A. Bulleid and D. Horne, ‘The Roman house at Keynsham, Somerset’, Archaeologia xxv (1926), 109–38.

⁴² The mosaic is discussed at length by Stupperich, op. cit. (note 7), 293–6.

⁴³ The grand dining-room and other examples are discussed by Ellis, op. cit. (note 33). For the Mediana house see S. Drca et al., Mediana (1979).

⁴⁴ Clifford, E., ‘The Roman villa, Witcombe, Gloucestershire’, TBGAS lxxiii (1954)Google Scholar; Wilson, D. (ed.), ‘Roman Britain in 1969’, Britannia i (1970), 294–5.Google Scholar

⁴⁵ The most notable attempt to identify Iron Age social structure in villa architecture is that of J. Smith, ‘Villas as the key to social structure’, in M. Todd (ed.), Studies in the Romano-British Villa (1978), 149–85, and ‘Halls or yards? A problem of villa interpretation’, Britannia ix (1978), 351–8Google Scholar. For a critique of Smith see Clarke, S., ‘The social significance of villa architecture in Celtic north-west Europe’, Oxford Journ. Arch, ix. 3 (1990), 337–53.CrossRefGoogle Scholar