¹ All abbreviations are standard, as used by M. Kaser in Das römische Privatrecht. D. stands for Digest; C. for Justinian's Code; G. for Gaius' Institutes. Laum followed by a number refers to the section of Latin inscriptions in volume n of Laum, B., Stiftungen in der griechischen und römischen Antike (1914)Google Scholar. The following works are cited by author and date only: Duncan-Jones, R. P., The Economy of the Roman Empire (2nd ed., 1982)Google Scholar; Grosso, G., I legati nel diritto romano (2nd ed., 1962)Google Scholar; di Salvo, S., Il legato modale nel diritto romano (1973)Google Scholar; Vitrano, F. Messina, ‘La convertibilità del modo eretto su legato o fedecommesso nel diritto romano classico e giustinianeo’, Studi Riccobono III (1936), 97–110Google Scholar; Amelotti, M., Il testamento romano (1966)Google Scholar; D'Ors, A., Epigrafia juridica de la España romana (1953)Google Scholar.

² On the social role of munificence, see esp. Veyne, Paul, Le pain et le cirque (1976), ch. 4Google Scholar (on the Roman empire); Andreau, J., ‘Fondations privées et rapports sociaux en Italie romaine (Ier-IIIe s. ap. J.-C.)’, Ktema 2 (1977), 157–209Google Scholar; Bruck, E. F., Über römisches Recht im Rahmen der Kulturgeschichte (1954), ch. 2Google Scholar.

³ Duff, P. W., Personality in Roman Law (1938), ch. 3, esp. 86 ffGoogle Scholar.; for more general discussion of these concepts, R. Feenstra, ‘Le concept de fondation du droit romain classique jusqu' à nos jours: théorie et pratique’, RIDA 3 (1956), 245–63; v. also von Carolsfeld, L. Schnorr, Geschichte der juristischen Person 1 (1933, repr. with addenda, 1969)Google Scholar.

⁴ Neugebauer, O. and van Hoesen, H. B., Greek Horoscopes (1959), 97Google Scholar.

⁵ In Greek inscriptions philotimia is commonly used to express this: TAM 11. 905; 1G IX. 1107; SIG (3rd ed.) 850, etc.

⁶ This is suggested by the evidence of inscriptions relating to foundations and sportulae collected by Duncan-Jones (1982). See also the excursus to Section 11 below.

⁷ Peter Brown, The Making of Late Antiquity (1978), 36.

⁸ By this I do not mean, of course, that bequests operated entirely outside politics. Clearly it was of advantage for a local political career if one's father had been notably munificent. Nor were fathers with dynastic ambitions blind to this: see, e.g., Philostratus, vitae 548 (on Herodes Atticus).

⁹ Pliny, ep. 1.8. 10 and 5. 11. The predominance of references by Ulpian to ‘imperator noster cum divo patre’ in passages of D. 50, 12, de pollicitationibus, may suggest a very late second-century origin. Cf. Buckland, Textbook (3rd ed., 1963), 458. On pollicitatio in general see, e.g., Garnsey, P., JRS 61 (1971), 116–29Google Scholar with further references.

¹⁰ Ulp., D. 39. 5. 19 pr.

¹¹ The relative proportions of legacies and pollicitationes in the total of munificence can be conveniently assessed using the lists in Duncan-Jones (1982).

¹² Lists of munera are given in D. 50. 4, de muneribus et honoribus. Duties such, for example, as maintenance of roads or buildings might be involved. On excusatio from munera see Millar, F., JRS 73 (1983), 76–96Google Scholar.

¹³ Objects changed over the years, particularly when munera came to occupy an important place in the local economy (which is not only in the late empire). For example, when the building of roads becomes a public munus, it is hardly likely that legacies for road building will continue to be left.

¹⁴ CIL 11. 4514; III. 6835; VIII. 1495; XIV. 350; Scaev., D. 33. 2. 17, D. 33. 1. 21. 3; Pap., D. 31. 77. 33; Marc., D. 33. 1. 24; Mod., D. 33. 1. 6, D. 33. 2. 16.

¹⁵ CIL v. 1978; III. 6998; x. 5654, 5657, 5853; VIII. 1495; XIV. 2827; Marc., D. 33. 1. 23.

¹⁶ CIL v. 5262; x. 5056, 6328; VIII. 1641; XIV. 350; XI. 5272; II. 1174; Scaev., D. 34. 1. 20. 1.

¹⁷ CIL v. 4015, 5262; XII. 4393; XIV. 2793.

¹⁸ CIL v. 5262; XI. 1602; XIII. 4132; XIV. 2934.

¹⁹ CIL x. 3851; Cels., D. 31. 30.

²⁰ CIL v. 6522; IX. 5074, 5075; XIV. 2978, 2979; Scaev., D. 32. 35. 3.

²¹ CIL 11. 3664; v. 5128.

²² A few cases where no object is specified are found: Scaev., D. 31. 88. 8; D. 32. 101 pr.; Ulp., D. 30. 71. 5. The juristic texts that provide for individuals to forfeit property to a municipality if they fail to comply with stated conditions also specify no object for the town's use of the property: Scaev., D. 32. 38. 5; D. 33. 2. 34 pr.; Pap., D. 36. 1. 59 pr. Epigraphic evidence involving forfeiture, however, (which is mostly from one town to another) normally envisages that the conditions of use are to attach to the beneficiary of the forfeit as well: CIL XI. 1436; XIV. 367; 431; 2795, etc.

²³ Pliny, ep. 5. 7. 1, ‘nec heredem institui nee praecipere posse rem publicam constat’, is not included here since clearly it is a case of incapacity to take per praeceptionem (which presupposed capacity to be instituted heir: G. 2. 218; cf. Voci, DER 1 (2nd. ed., 1967), 421 n. 83; and n. 33 below). Equally, it should not be taken as implying by omission that ordinary legacies were possible; that issue is not to the point in the letter.

²⁴ This work is cited henceforth as Ulp., E. The standard edition is by Schulz, F., Die epitome Ulpiani des Codex Vaticanus Reginae 1128 (1926)Google Scholar. The text is also in FIRA 11. 261 ff.

²⁵ Mitteis, RP (1908), 377.

²⁶ Voci, DER 1. 421.

²⁷ See e.g., Der kleine Pauly, s.v. Trebiae.

²⁸ But for Suetonius as a precise user of technical terms see A. Wallace-Hadrill, Suetonius (1983), 20, etc. Yet the fact that Suetonius wrote in the time of Hadrian, when the first gradual steps towards the assimilation of legacies and fideicommissa were being taken, may make us wonder whether the difference would then have been regarded as very significant.

²⁹ Ulp., E. 22. 5; D. 38. 16. 3. 6; D. 40. 3. 1–2; D. 38. 3. 1. 1.

³⁰ CIL x. 5056 (= FIRA 3. 55c).

³¹ The precise wording of the text (‘legavit ut’) is important here since the Roman jurists distinguished between conditional legacies and modal legacies (legacies sub modo). In modal legacies the beneficiary acquires the object legated (or a right to acquire it) immediately on the death of the testator, whereas in conditional legacies the object is acquired only once the condition has been satisfied (Kaser, RP 1. 259 ff.; admittedly this statement applies strictly only to suspensive conditions, but in classical law resolutive conditions are rare for strict obligations). The beneficiary under a modal legacy is supposed to use it in the prescribed manner, but since ownership passes to him in any case there are difficulties in providing secure remedies. The point is discussed fully below.

³² Voci, DER 1. 421.

³¹ Towns could not inherit, but for other exceptions to the rule see Mitteis, RP, 378 n. 12. It is as well to point out here that in Roman law there is a sharp distinction between inheritance and legacy and so also between relative capacities to acquire them: a legatee acquires a single object; an heir is a ‘universal successor’ who acquires not only the estate, but the duty to pay the legacies, and liability for debts of the estate. It was consequently necessary to circumscribe the right to be heir more than that to be legatee.

³⁴ cf. n. 31.

³⁵ AE 1926, 143 (Sinuessa); CIL x. 3851 (Capua, colonia); CIL XI. 720 (Bononia, imperial colonia); CIL XI. 5745 (Sentinum); CIL x. 1416 (Herculaneum). I am grateful to Michael Crawford for pointing out to me the evidence for ownership of land in Cisalpine Gaul by Italian communities (Arpinum and Atella, both municipia) under the Republic: Cicero, ad fam. 13. 11. 1; 13. 7. 1–3). As he suggests (Coinage and Money under the Roman Republic (1985), 340), this property was most likely acquired by bequest.

³⁶ Gnomon of the Idios Logos 18 (FIRA 1. 469 ff.). Admittedly, this refers only to fideicommissa hereditatis. But I think the argument in the text can still stand, since this must have introduced new difficulties.

³⁷ The evidence could perhaps be explained in another way: while it was not until Nerva possible to leave legacies to towns in the sense that an obligation to pay them thereby arose and that obligation became actionable, it could be argued that our pre-Nervan evidence reflects cases where legacies left were in fact paid by the heirs in spite of the absence of an obligation. (I owe this point to a discussion with Prof. Crook.) This is the sort of position envisaged in Pliny, ep. 5. 7. None the less, I am inclined, particularly on the basis of lav., D. 35. 1.39. 1, and on that of arguments produced from the statistical material used in the excursus to this section, to believe that given the extent of pre-Nervan munificence in Italy a legacy to a municipium must have been actionable before Nerva.

³⁸ If the SC is to be dated to A.D. 117, it is of course scarcely Hadrianic; although Trajan died in August of that year, Hadrian returned to Rome from the East only in 118.

³⁹ Voci, DER 1. 421 n. 85.

⁴⁰ The question whether towns were really regarded as incertae personae is in fact more doubtful than is usually assumed: the term is rarely found in that context. I hope to investigate this question at a later date.

⁴¹ Pliny, ep. 10. 75. 2.

⁴² Gnomon 18.

⁴³ CIL VIII. 26528b: ‘… M. Aurelio … pagus Thugg. caelesti beneficio eorum auctus iure capiendorum legatorum d.d.p.[p.]’

⁴⁴ RE XVIII. 2 (esp. cols. 2327 ad fin.–2331), s.v. pagus (Kornemann). For other discussions of the question, Thompson, L. A., Latomus 24 (1965), 150–4Google Scholar; J. Gascou in ANRW 10. 2 (1982), 201, 207.

⁴⁵ Kornemann, ibid.

⁴⁶ RE Suppl. VII, cols. 1567–71, s.v. Thugga (Windberg).

⁴⁷ ibid.

⁴⁸ On the mechanism of beneficia see F. Millar, The Emperor in the Roman World (1977), 420 ff. Cf. also Saller, R., Personal Patronage under the early Empire (1982), 41 ffCrossRefGoogle Scholar.

⁴⁹ Gai., D. 30. 73. 1; Ulp., D. 30. 32. 2.

⁵⁰ G. 2. 195.

⁵¹ P. Oxy. IV. 705 of A.D. 202.

⁵² The referral to Severus is therefore not to be explained as an exception due to his presence in Egypt: he had left the province by A.D. 202.

⁵³ For discussion of the imperial supervisory role over endowments, a question that is largely avoided here, see K. M. T. Atkinson in RIDA 1962, 261–89, esp. 285 ff., who argues that it goes back to Augustus, using CIL III. 7124, an endowment of Vedius Pollio at Ephesus, and its parallel Greek text which states that the διάταξις was συνφυλαχθεῖσα by Augustus. I am not inclined, however, to follow her argument (287 ff.) that imperial involvement was brought about by fiscal interest, in the shape of fines payable to the fisc; it seems more probable that involvement was requested as an aid to securing endowments from misuse. J. H. Oliver, ‘The Ruling Power’, TAPS 43 (1953), 963 ff., discusses a procedure of registering endowments with provincial or imperial authorities which appears to be limited to Greek cities. Since the form taken is generally the prescription of a penalty for interference (often of twice the value of the endowment), this procedure forms an exception to the view expressed in the text that imperial involvement tended to confirm existing, rather than establish new, sanctions. P. Veyne, Le Pain et le cirque (1976), 731 n. 9, also mentions imperial protection of foundations: infringement would be an impiety towards the emperor. (He cites Dunant, C. and Pouilloux, J., Recherches sur l'histoire et les cultes de Thasos (1954–1957) 11. 78Google Scholar. Unfortunately vol. II has not been accessible to me.)

⁵⁴ The term cautio denotes a contractual obligation assumed in order to guarantee performance of an act, whether already protected by law or (as here) not. The usual method was to use the standard verbal contract (stipulatio) and to take security that the act would be performed.

⁵⁵ Paul, D. 22. 6. 9. 5 (lib. sing, de iuris et facti ignorantia).

⁵⁶ Ulp., D. 36. 3. 6. 1 (lib. 6 fideicommissorum). In the text there is some oddity about ‘intercedat’ (Lenel, Pal. col. 925) but the substance seems acceptable. On repromissio in general, Kaser, RZ, 336.

⁵⁷ Paul, D. 30. 122 pr. (lib. 3 regularum), ‘Towns can also be left by legacy objects which bring them esteem (honos) or embellishment (ornatus). Examples of embellishment are what has been left for the forum, theatre or stadium; of esteem, what has been left for games, hunting, drama, circus, for division among the citizens or for a dinner. Moreover money left for maintaining the infirm (whether old people or children) is said by jurists to contribute to the esteem of a town.’

Ulp., D. 30. 32. 2 (lib. 20 ad Sabinum), ‘If something which contributes to the embellishment (ornatus) or profit (compendium) of a town has been left to a part of the town, then the legacy is certainly owed.’

⁵⁸ Marc, D. 30. 117 (lib. 13 institutionum). In the texts cited in the previous note, Ulpian's ornatus and compendium are matched by Paul's honos and ornatus. The fact that Marcian's list is rather different is not in itself problematic, since the Digest (in spite of Justinian's boast to the contrary: Const. Tanta 15) contains a mass of divergent and sometimes contradictory statements. My grounds for going on to argue against the genuineness of ‘sive quid aliud’ in the text of Marcian are therefore not belief in total consistency in the Digest but belief that, given certain basic facts about the law concerning variation of endowments, the position expressed by Marcian is unlikely to reflect classical law. The point recurs (see nn. 130, 131 and the text there): it is misguided to force fluid or tentative or particular texts into universal harmony.

⁵⁹ On interference with the texts, Voci, DER 1. 424 n. 96; di Salvo (1973), 164 n. 225.

⁶⁰ Astolfi, R., Studi sull'oggetto dei legati in diritto romano III (1979), 104 ffGoogle Scholar.

⁶¹ Pap. lust., D. 50. 12. 13. 1 (lib. 2 de constitutionibus). The language of the text is odd: it speaks first of donationes, then gives an example concerning legacy; gifts with in followed by the accusative are inelegant; the last sentence (‘esse enim tolerabilia quae vetus consuetudo comprobat’) is mysterious, its ‘enim’ particularly unusual (we might expect ‘autem’). For views on interpolation: Beseler, , Beiträge III (1913), 67Google Scholar; Albertario, RIL 60 (1927), 608; Longo, , Labeo 18 (1972), 39 ffGoogle Scholar.

⁶² CIL II. 4514. On this inscription see Amelotti (1966), 24, 132; d'Ors (1953), 420–2. I translate dicta here, following d'Ors. Hubner (editor in CIL) suggests lecta. The question at issue here is in any case not affected.

⁶³ By this I do not mean that acceptance necessarily involves a positive decision about economic viability. The point will be discussed below.

⁶⁴ Marcel., D. 34. 2. 6. 2 (lib. sing, responsorum).

⁶⁵ SHA, Antoninus Pius 12. 1Google Scholar; Marcel., D. 28. 4. 3; Kunkel, Herkunft, 213; RE IXa, col. 570 (Mayer-Maly).

⁶⁶ The mixed language of the text is notable: the wording is that of a fideicommissum, yet it is described by both petitioner and jurist as a legacy.

⁶⁷ Voci, DER 1. 424 n. 96.

⁶⁸ The case of Pliny provides an interesting comparison (CIL v. 5262 = Laum, 85). Here the legacy to Comum is intended in the first instance to provide alimenta for Pliny's freedmen and subsequently to be spent on a dinner for municipes. Some similar arrangement could apply in Scaev., D. 34. 1. 20. 1, in which case Voci's argument would be even weaker.

⁶⁹ Paul, D. 22. 6. 9. 5.

⁷⁰ Water was after all a valuable commodity in those parts; valuable enough to make possible a fideicommissum of water: Ulp., D. 34. 1. 14. 3.

⁷¹ Mod., D. 33. 2. 16 (lib. 9 responsorum).

⁷² The figures in Duncan-Jones (1982) show that in spite of great fluctuation there were ‘normal’ cost levels for items.

⁷³ On dilapidated buildings: several SCC from the SC Hosidianum of A.D. 44 onwards affirm that buildings are not to be demolished; there is an exception from this in favour of a town in Ulp., D. 30. 41. 5. Other texts showing concern for the state and appearance of the city: Jul., D. 43. 8. 7; Ulp., D. 43. 8. 2. 17; D. 1. 16. 7. 1; SHA, Hadrian 18. For further details, B. Ward-Perkins, From Classical Antiquity to the Middle Ages (1984), 12 f.

⁷⁴ Pliny, ep. 10. 70. 2.

⁷⁵ CIL XIII. 4132 (= Laum, 96).

⁷⁶ The normal rate of interest on foundations (which usually had their funds invested in land) was 6 per cent: Duncan-Jones (1982), 133 ff., although large foundations with a capital over 100,000 sesterces may have had a rate closer to 5 per cent.

⁷⁷ Mrozek, S., ‘Zur Frage der tutela in römischen Inschriften’, Acta antiqua Academiae Scientiarum Hungaricae 16 (1968), 283–8Google Scholar.

⁷⁸ See costs, nos. 11433–1160 in Duncan-Jones 1982).

⁷⁹ ibid., nos. 1143c and d.

⁸⁰ ibid., no. 1143a.

⁸¹ Mickwitz, G., ‘Economic rationalism in Graeco-Roman agriculture’, English Historical Review 52 (1937), 577–89CrossRefGoogle Scholar.

⁸² Cicero, ad Att. 6. 2. 5.

⁸³ Pliny, ep. 7. 18. 1.

⁸⁴ 7. 18. 2. Similar methods are found in CIL x. 5853 (Laum, 23) and CIL XI. 419 (Laum, 55).

⁸⁵ The stipulatio is the standard Roman verbal contract. Although formal, it is flexible in the sense that it can be used to encompass any kind of obligation. Here a stipulatio poenae would be a possibility: the contract would set a penalty to be paid by the town if it failed to perform its obligation (cf. Kaser, RP I. 519 ff.; R. Knütel, Stipulatio poenae (1976), 45 ff., etc.).

⁸⁶ On Pliny in Bithynia, Sherwin-White, A. N., The Letters of Pliny: a historical and social commentary (1966), 527Google Scholar; Duncan-Jones (1982), 304–5. For further examples of misuse of funds, mainly at Ephesus, see Jones, C. P., JRS 73 (1983), 116–25Google Scholar.

⁸⁷ Periodic payment: Marc, D. 33. 1.23 (divisiones to decurions on testator's birthday); D. 33. 1. 24; Scaev., D. 33. 1. 20. 1; Paul, D. 35. 2. 3. 2; FIRA 3. 118 with Arangio Ruiz 383 n. 4 and Amelotti (1966), 24; di Salvo (1973), 162 n. 219.

⁸⁸ Ex reditu clauses: (e.g.) CIL x. 5809; XIV. 2795; 11. 3167; x. 3851. Similarly, ex usuris clauses: III. 6998; v. 1978; x. 107; 5657.

⁸⁹ On the other hand satisfactory treatment of property under a usufruct depended also on taking a cautio (see n. 54), here the cautio usufructuaria; and on problems with cautiones see below.

⁹⁰ There are examples of usufructs in the context of foundations, but not with towns as beneficiaries. For instance, Mod., D. 34. 1. 4; the inscription of Iunia Libertas (with F. de Visscher in Studi Solazzi (1948), 542–53).

⁹¹ Gai., D. 7. 1. 56; D. 33. 2. 8; Pap., D. 31. 66. 7; Mod., D. 7. 4. 21; Macer, D. 35. 2. 68 pr., all with P. G. Stein, ‘Generations, lifespans and usufructs’, RIDA (1962), 347 ff.

⁹² Kaser, RP 1. 259. See also n. 31.

⁹³ For example, the validity of the legacy was not at all affected by either initial or supervening impossibility of the modus: Voci, , DER II. 620Google Scholar ff.

⁹⁴ Mitteis, RP, 200.

⁹⁵ Mitteis, RP, 198; di Salvo (1973), 186.

⁹⁶ On cautio see n. 54.

⁹⁷ Kaser, RP 1. 259; Grosso (1962), 469.

⁹⁸ The cautio would be more than an initial remedy if were possible to obtain a condictio incerti for it. This is disputed (for references see ‘Prohibitions and perpetuities’ in SZ 102 (1985), 255 n. 121). In none of the texts under consideration, however, does there appear to be any hint of one. Texts such as Paul, D. 22. 6. 9. 5 seem depend on the non-availability of a condictio in those circumstances. The most interesting text in this connection is Ulp., D. 35. 3. 3. 10 in which Pomponius' view that a condictio for a cautio is possible, in the form a condictio interponendae satisdationis gratia, is approved by Ulpian. The remedy appears, therefore, to have been early known in this ‘disguised’ form in which clearly a set sum for satisdatio would be sought. On the other hand, satisdatio is not normal for public bequests Ulp., D. 36. 3. 6. 1), so the condictio would have to be for the much less tangible repromissio. The question cannot be covered thoroughly here. I hope to deal with subsequently.

⁹⁹ ‘Legavit uti’ shows that it is a modus. The use of condicio in the text is untechnical: di Salvo (1973), 103 ff.

¹⁰⁰ Beseler, , Beiträge IV (1920), 290Google Scholar.

¹⁰¹ Pernice, , Labeo III. 1, 47Google Scholar.

¹⁰² But see n. 98.

¹⁰³ For further discussion, SZ 102 (1985), 252 ff.

¹⁰⁴ di Salvo (1973), 214 ff., 321 ff.; Pernice, , Labeo III. 1, 39Google Scholar; cf. Mitteis, RP, 196 n. 5.

¹⁰⁵ As Pomponius says later in D. 33. 1. 7, enforcement by a co-heir would be achieved by an action for division of the estate (iudicium familiae erciscundae).

¹⁰⁶ See SZ 102 (1985), 254 ff.

¹⁰⁷ On conditional ademptio: Grosso (1962), 468; Pernice, 47. In containing this clause this text is clearly exceptional, presumably the reason the questioners quote it in the first place. But cf. CIL v. 5134 (Laum, 84).

¹⁰⁸ There appears to be a lacuna after this sentence, since facts are assumed in the responsum which are not stated in the quaestio (cf. Lenel, Pal. ad loc). Mommsen suggests the addition of ‘invitis heredibus praedia possedit et’.

¹⁰⁹ Biondi, La compensasione nel diritto romano (1927), 260 ff.: he also expunges ‘et invitis heredibus possessione adprehensa’ and inserts ‘non’ before ‘restituendos’.

¹¹⁰ I do not, however, mean to imply by this that multiple bequests to towns were uncommon: Pliny is only one (unusually generous) example; others may be found in Duncan-Jones (1982), e.g. nos. 468 + 646 + 653 + 654.

¹¹¹ On modus as fideicommissum: for the literature SZ 102 (1985), 284 n. 224. Mod., D. 33. 2. 16 (quoted in Section III) is an interesting example. The confused terminology of the text is notable. It is probably because the modus is treated as a fideicommissum that the responsum is concerned with the voluntas testatoris, and that the heirs have more say in variation than otherwise.

¹¹² Suetonius, , Tiberius 31. 3Google Scholar.

¹¹³ Mitteis, RP, 198 n. 12.

¹¹⁴ Val., D. 50. 8. 6.

¹¹⁵ Call., D. 50. 10. 7 pr.

¹¹⁶ Val., D. 50. 8. 6.

¹¹⁷ Marc, D. 33. 1. 23 and 24.

¹¹⁸ J. H. Oliver (op. cit., n. 53).

¹¹⁹ F. Millar (op. cit., n. 48), 341–55.

¹²⁰ Millar, 507 ff., 528 ff., etc.

¹²¹ Suetonius, Tiberius 31.3.

¹²² Call., D. 50. 10. 7 pr.

¹²³ CIL XII. 4393; XIV. 2934; a fine has to be paid to the fisc in three cases cited by J. H. Oliver (op. cit., n. 53) (his numbers 1, 5 and 6).

¹²⁴ CIL XIV. 2793, 2795; 11. 4514.

¹²⁵ Messina Vitrano (1936), 97–110; Beseler, , Beiträge II (1911), 49Google Scholar; TR 10 (1930), 199; SZ 45 (1925), 487. The text goes on to discuss cases where the incidence of the Lex Falcidia causes insufficiency of funds. That discussion begins ‘et ideo’, decidedly odd since no apparent logical connection links that sentence with the first (quoted here). Similarly, the last sentence cannot be connected at all with the sentence that precedes it.

¹²⁶ Messina Vitrano, 107.

¹²⁷ Grosso (1962), 473; di Salvo (1973), 165 n. 229.

¹²⁸ Grosso, ibid.

¹²⁹ Call., D. 50. 10. 7 pr. (lib. 2 de cognitionibus).

¹³⁰ Messina Vitrano, 108 ff.

¹³¹ There are plenty of examples of building: see, for instance, Marcel., D. 34. 2. 6. 2 where there is no objection to using the money for an ‘opus novum’. Since Marcellus was a member of the consilium of Antoninus Pius, and later that of M. Aurelius (see n. 65), his evidence is contemporary with or, more likely, later than this rescript. On more general grounds for treating this as a particular rather than a general case, see n. 58.

¹³² Millar, loc. cit., n. 48.

¹³³ Macer, D. 50. 10. 3. pr. (lib. 2 de officio praesidis).

¹³⁴ cf. Millar, 451–2.

¹³⁵ Much imperial legislation on destruction and dereliction of buildings is found in SCC. The best known is perhaps the SC Hosidianum of A.D. 44, whose principles were restated in both the SC Volusianum (A.D. 56) and the SC Acilianum (A.D. 122).

¹³⁶ These include Paul, D. 30. 122 pr., one of the texts purporting to list possible objects for bequests to towns (see n. 57) and also Mod., D. 33. 2. 16, ‘the case of the illicit spectacle’ which envisages only that a licit spectaculum be chosen instead. Other cases: Scaev., D. 33. 2. 17; D. 33. 1. 21. 3, etc.

¹³⁷ Biondi, , Il diritto romano cristiano II (1952), 279Google Scholar; di Salvo (1973), 165 n. 229.

¹³⁸ SHA, Hadrian 19. 2–4Google Scholar; cf. 2. 1; 7. 12; 14. 10; 20. 13.

¹³⁹ Early Republican instances of variation to avoid games are listed by Duncan-Jones (1982), 149. Later examples are CIL XI. 5276 (pre-A.D. 100) and X. 1491 (post-A.D. 100). Duncan-Jones, 136 ff. points out that there are only three large foundations for games attested in Italy, and that this is because of attempts to divert such legacies to more useful public purposes. Cf. also the tentative inscription CIL V. 7637 (Laum, 62).

¹⁴⁰ SIG (3rd ed.), 850. While this is really an example of pollicitatio (ἐπαγγελία), the principles of gifts favoured and gifts discouraged are likely to be the same where bequests are concerned.

¹⁴¹ Messina Vitrano, 106 ff.; di Salvo (1973), 165 n. 229.

¹⁴² Constantine, C. 1. 2. 1. (321). We cannot now enter on the details of the changes forced by Christianity. Much had been left in a non-Christian empire for pagan purposes, or for purposes (such as venatio) of which the church did not approve. Equally, Christian morality favoured alms-giving rather than the pagan glories of conspicuous consumption (P. Veyne, Le pain et le cirque, 51 ff., etc.). The church began by affecting the amount and purposes of bequests to towns; it ended by stifling the largesse almost entirely. Cf. Ward-Perkins (op. cit., n. 73), 14 ff.; and on churches, esp. 65 ff.

¹⁴³ Brown, Peter, The Making of Late Antiquity, 49Google Scholar.