¹ So for example de Ste. Croix, G. E. M. in his influential paper, ‘Why were the early Christians persecuted?’, Past and Present 26 (1963), 6–38CrossRefGoogle Scholar = Finley, M. I. (ed.), Studies in Ancient Society (1974), 210–49Google Scholar.

² On this see especially Fowden, G., Empire and Commonwealth: Consequences of Monotheism in Late Antiquity (1993).Google Scholar

³ So for example Turcan, R., ‘Le culte impérial au IIIe siècle’, ANRW II.16.2 (1978), 996–1084Google Scholar.

⁴ The most complete collection is Knipfing, J. R., ‘The libelli of the Decian persecution’, HTR 16 (1923), 345–90CrossRefGoogle Scholar, which provides texts and translations of forty-one certificates; four of these have more recent editions (Knipfing nos 35–6 = P. Mich. III.157–8; Knipfing nos 38–9 = P. Hamb. 61 a and b), and three others have since been published (PSI VII. 78, SB VI.9084, P. Oxy. XLI.2990).

⁵ There is a new edition of the letters by G. F. Diercks (Corpus Christianorum 3B, 1994), while Clarke, G. W., The Letters of St. Cyprian (4 vols, 1984–1989)Google Scholar provides essential commentary; in my references I follow the numbering of Clarke. On the De Lapsis, see M. Bévenot, Cyprian: De Lapsis and De Ecclesiae Catholicae Unitate (1971).

⁶ The first letter (HE 6.41–2 and 44) was written to Fabius of Antioch and dealt with the problem of the lapsed; its date is suggested by the roughly contemporary letter of Cornelius of Rome to Fabius (HE 6.43.3–22), written after the synods of Rome and Carthage in April–May A.D. 251 but before Fabius' death later that year (HE 6.46.4). See further P. Nautin, Lettres et écrivains Chrétiens des IIe et IIIe siècles (1961), 143–56 and 163–5. The second (HE (6.40) was written in response to a certain Germanus; its date is indicated by the fact that the persecution of Valerian was underway (HE 6.40. 1), and had been for some time (HE 7.11.2–19); see further Pietras, H., ‘Lettera pros Germanon di Dionigi Alessandrino’, Gregorianum 71 (1990), 573–83Google Scholar.

⁷ Grégoire, H., Orgels, P., and Moreau, J., ‘Les martyres de Pionios et de Polycarpe’, Bulletin de l'Académie royale de Belgique, classe des lettres et des sciences morales et politiques 47 (1961), 72–83Google Scholar, argue strongly for a date under Marcus Aurelius; H. Musurillo, The Acts of the Christian Martyrs (1972), xxviii–xxix and lvii, argues that the text is late and ‘obviously embroidered’.

⁸ See especially Barnes, T. D., ‘Pre-Decian Acta Martyrum’, JTS 19 (1968), 509–31CrossRefGoogle Scholar, at 529–31; Fox, R. Lane, Pagans and Christians (1986), 460–8Google Scholar; and Robert, L., Le Martyre de Pionios, prêtre de Smyrne (1994), especially 1–9Google Scholar. Lane Fox argues that Pionius wrote most of the text while in prison, and that an editor shortly thereafter made short additions at the beginning and end (ibid., 468–72); cf. Robert, ibid., 49–50. Other martyr acts may also recount events under Decius, e.g. the Acts of Carpus, Papylus, and Agathonikê, but the difficulties in employing them as independent sources of evidence are such that I have not attempted to use them here.

⁹ Potter, D. S., Prophecy and History in the Crisis of the Roman Empire: A Historical Commentary on the Thirteenth Sibylline Oracle (1990), 141–2Google Scholar, 147, and 258–9, argues that the main author of this text, working in Syria in A.D. 253, compiled ll. 1–88 from earlier material but composed ll. 89–154 himself. He was thus responsible for the brief and uninformative account of Decius' death (ll. 100–2), but for his reign combined two different texts, one positive (ll. 81–3) and the other negative (ll. 84–8); the latter contains the reference to persecution (ll. 87–8): ‘and immediately there will be robberies and murders of the faithful [pistôn] on account of the former king’; the reading pistôn is Wilamowitz' emendation for the manuscript's piptôn. Because this is the only explicitly Christian reference in the poem, it has often been taken as a later interpolation, but Potter argues that these lines are more likely a contemporary Christian composition.

¹⁰ The best discussions are Molthagen, J., Der römische Staat und die Christen im zweiten und dritten Jahrhundert (1970), 61–84Google Scholar; Clarke, op. cit. (n. 5), 22–39, with his two earlier studies, ‘Some observations on the persecution of Decius’, Antichthon 3 (1969), 63–76CrossRefGoogle Scholar, and ‘Two measures in the persecution of Decius? Two recent views’, BICS 20 (1973), 118–23Google Scholar; Pohlsander, H. A., ‘The religious policy of Decius’, ANRW II.16.3 (1986), 1826–42Google Scholar; and Potter, op. cit. (n. 9), 40–3 and 261–8. Also important are Lieserung, E., Untersuchungen zur Christenverfolgung des Kaisers Decius (1933)Google Scholar; Alföldi, A., ‘Zu den Christenverfolgung in der Mitte des 3. Jahrhunderts’, Klio 31 (1938), 323–48CrossRefGoogle Scholar = Studien zur Geschichte der Weltkrise des 3. Jahrhunderts nach Christus (1967), 285–311; Frend, W. H. C., Martyrdom and Persecution in the Early Church (1965), 404–13Google Scholar; Sordi, M., ‘La data dell'editto di Decio e il significato della persecuzione anticristiana’, Rivista di storia della Chiesa in Italia 34 (1980), 451–61Google Scholar and The Christians and the Roman Empire (1986), 100–7; and Lane Fox, op. cit. (n. 8), 450–62. I shall also refer to Keresztes, P., ‘The Decian libelli and contemporary literature’, Latomus 34 (1975), 761–81Google Scholar, reprinted in his Imperial Rome and the Christians (2 vols, 1989).

¹¹ Senatorial recognition of Decius occurred probably in September A.D. 249: Peachin, M., Roman Imperial Titulature and Chronology, A.D. 235–284 (1990), 30–2Google Scholar. Among the first Christians to be martyred as a result of the decree were Fabian of Rome and Babylas of Antioch (Eus., , HE 6.39.1Google Scholar and 4), whose deaths according to Western tradition took place on 20 and 24 January respectively. In order for the edict to be known in Antioch by late January, it cannot have been issued much later than the beginning of that month. Most scholars favour a date in December or early January, although Potter, op. cit. (n. 9), 261–3, makes a good case for the autumn of A.D. 249.

¹² Cyprian speaks of an edictum (Laps. 27) or edicta (Ep. 55.9.2), while Novatian refers more vaguely to edicta vel leges (ap. Cypr., , Ep. 30.3.1Google Scholar) and the Roman confessors to leges (Ep. 31.3 and 5). The Passio Pionii uses the Greek word diatagma (3.2), while Dionysius of Alexandria uses prostagma (Eus., , HE 6.41.1Google Scholar and 10); both are regular translations of the Latin edictum: H. J. Mason, Greek Terms for Roman Institutions: A Lexicon and Analysis (1974), 36 and 81. The papyri use the less technical phrases ta prostechthenta or takeleusthenta. On edicts in general, see Millar, F., The Emperor in the Roman World (1977), 252–9Google Scholar; the fact that the Egyptian certificates have virtually identical wording suggests that it was posted in public for individuals to copy: Millar, ibid., 255–6.

¹³ All the extant papyrus certificates contain some variation on the phrase ‘I have always and without interruption sacrificed to the gods and now in your presence in accordance with the edict's decree I have sacrificed, poured a libation, and tasted of the sacred victims’. Similarly, the confessors in Rome refer to lips defiled by accursed food (ap. Cypr., , Ep. 31.7.1Google Scholar), and Cyprian himself praises those who refused to offer sacrifices or taste of them (Laps. 2 and 28); the Passio Pionii describes how the neokoros Polemon ‘sought out the Christians and summoned them to sacrifice and eat defiled food’ (3.1; cf. Robert at 2.4). Cyprian at one point mentions thurificati (Ep. 55.2.1), presumably people who had only offered incense; whether this was a special dispensation for the poor or simply a local variation is unknown.

¹⁴ The papyrus petitions are addressed ‘to those chosen for the sacrifices’: twenty-five of the petitions from Theadelphia are certified by two individuals, presumably local commissioners, and signed by the town secretary, while one from Arsinoe is certified by a town councillor (Knipfing, op. cit. (n. 4), no. 25). Cyprian mentions in Carthage a board of five leading citizens in association with the magistrates (Ep. 43.3.1), but in Capsa only one magistrate (Ep. 56.1.1); an imprisoned confessor in Rome refers to magistrates (Ep. 21.1.1). In Smyrna, the commission consisted of ‘the neokoros Polemon and those associated with him’ (Pass. Pion. 3.1). This variation suggests that local authorities could either appoint a commission or assume the duties themselves.

¹⁵ Pionius, after persisting in his refusal to sacrifice, is held in prison until the proconsul arrives; it is the latter who orders him burned alive (Pass. Pion. 19–20). In Carthage there were people who confessed once before the magistrates and a second time before the proconsul (Cypr., , Ep. 38.1.2Google Scholar); others in Capsa withstood the first interrogation but not that before the proconsul (Cypr., , Ep. 56.1Google Scholar). Dionysius mentions trials before the governor (ap. Eus., , HE 6.41.18, 21, 23Google Scholar), although other instances of Christians being killed sound more like lynchings (e.g. HE 6.41.15 and 42.1). On punishments, see Clarke, op. cit. (n. 5), I, 35–6.

¹⁶ On the universal applicability of the edict, see further Molthagen, op. cit. (n. 10), 62–3, and Clarke, op. cit. (n. 5), I, 26–8. There is no direct evidence for the Jews, but if they had been required to sacrifice, we would expect some trace of this in the record; in fact, there is none whatsoever. This silence is so striking that we must assume an exemption. The only explicit mention of Jews in any of the sources is in the Passio Pionii, which depicts them as part of the anti-Christian mob (3.6; cf. 4.2); although this text has an obvious anti-Jewish tendency, it would hardly have depicted them as allies of the gentiles if they were actually in the same plight as the Christians. As a parallel, a passage in the Jerusalem Talmud suggests that the Jews of Caesarea were specifically exempted from the requirements of Diocletian's fourth edict: Abodah Zarah 5.4, 44d; cf. Lieberman, S., ‘The martyrs of Caesarea’, Annuaire de l'Institut de Philologie et d'Histoire Orientates et Slaves 7 (1939–1944), 395–446Google Scholar, at 403–4. It is likely that both Diocletian and Decius followed the long established principle of making allowances for the ancestral traditions of the Jews.

¹⁷ FGrH 100 F 22; see also F 26. Schwartz, J., ‘A propos des ch. 4 à 6 du “De Mortibus Persecutorum”’, in Fontaine, J. and Perrin, M. (eds), Lactance et son temps (1978), 91–102Google Scholar, argues that the brief account of Lactantius also derives from Dexippus. For other historians who covered Decius' reign, see Potter, op. cit. (n. 9), 70–94.

¹⁸ Zos. 1.21–3; Zon. 12.19–20. Since their accounts of Decius' accession, final battle, and death agree in several details with each other and against other accounts, they probably derive from the same source; since their version of Decius' death differs from that of George Syncellus, this source was probably not Dexippus.

¹⁹ Vict. 28.10–29.5; Eutr. 9.4; Epit. de Caes. 29; on the Kaisergeschichte, see most recently Burgess, R. W., ‘On the date of the Kaisergeschichte’, CP 90 (1995), 111–28Google Scholar. All three works note the place of Decius' birth, record the elevation of his son as Caesar, and agree in their accounts of his death, while Victor and Eutropius both mention his building activities. Ammianus Marcellinus and the Historia Augusta may have provided fuller examples of this tradition, but unfortunately neither is extant for the reign of Decius; against the idea that this was a deliberate omission by the author of the HA, unwilling to handle the delicate topic of Christian persecutions, see R. Syme, Emperors and Biography: Studies in the Historia Augusta (1971), 200–2.

²⁰ For general studies, see Salisbury, F. S. and Mattingly, H. B., ‘The reign of Trajan Decius’, JRS 14 (1924), 1–23Google Scholar; Wittig, K., RE 15 (1931), 1244–84Google Scholar; Fronza, L., ‘Studi sull'imperatore Decio’, Annali Triestini 21 (1951), 227–45Google Scholar and 23 (1953), 311–33 (non vidi); PIR ² M 520; and most recently the brief account of Potter, op. cit. (n. 9), 40–4.

²¹ Eutropius (9.4) and the Epitome de Caesaribus (29.1) name the village, while Victor (29.1) simply says near Sirmium; note also Sib. Or. 13.83: ‘emerging from the Dacians’. The Chronicon Pascale (505 Dindorf) says he was sixty when he died, the Epitome de Caesaribus (29.4) fifty; as Syme, op. cit. (n. 19), 197, observes, the former fits much better with his career.

²² On his earliest coins and inscriptions he appears as ‘Cos. II’; since he does not previously appear in the consular lists, his first consulship must have been suffect; its date is a guess based on his probable age and his known consular command in A.D. 234.

²³ PIR ² H 136; cf. Syme, op. cit. (n. 19), 197: the Etruscan connection is indicated not only by the name ‘Etruscilla’ and her son's name ‘Etruscus’, but more reliably by her younger son Hostilianus' second name ‘Perpenna’ (Epit. de Caes. 30.3, if reliable) and the very rare nomen ‘Cupressenia’, elsewhere attested only for Cupressenus Gallus, suffect consul in A.D. 147.

²⁴ Moesia Inferior: CIL III.12519Google Scholar, 13724, and 13758; Tarraconensis: PIR ² D 28 with AE 1951.9; see further Gerov, B., ‘Zur Identität des Imperators Decius mit dem Statthalter C. Messius Q. Decius Valerinus’, Klio 39 (1961), 222–6CrossRefGoogle Scholar, and Syme, op. cit. (n. 19), 192 and 196.

²⁵ of Antioch, John (FHG IV.597–8Google Scholar F 148) presents a wildly different version: Philip, after a battle with ‘Scythians’, was marching with his army to Byzantium when Decius led a revolt in Rome and declared himself emperor; after Decius won over his envoys, Philip fled to Beroia where he was assassinated. Although defended by Dušanić, S., ‘The end of the Philippi’, Chiron 6 (1976), 427–39Google Scholar and by Sordi, op. cit. (n. 10, 1986), 98–9, this account has been forcefully rejected by Pohlsander, H. A., ‘Did Decius kill the Philippi?’, Historia 31 (1982), 214–22Google Scholar; see most recently Potter, op. cit. (n. 9), 254–8.

²⁶ Rescripts, : CJ 4.16.2Google Scholar and 10.16.3; for other laws in Decius' name, see CJ 6.30.4, 7.32.3, 8.53.3 (March–April 250); 5.12.9 (June 250); 3.22.2 and 6.58.3 (December 250). The baths are attested by Eutropius (9.4) and Victor (29.1); see now La Follette, L., ‘The baths of Trajan Decius on the Aventine’, in Rome Papers, JRA Supplementary Series II (1994), 6–88Google Scholar.

²⁷ So Dexippus, FGrH 100 F 22; on the location of Abritus, see Ivanov, T., ‘Borne milliaire avec le nom d'Abritus’, Archéologie (Sofia) 23.3 (1981), 48–53Google Scholar. Eusebius noted in his Chronicle that Decius died at Abritus (Schöne 180–1), whence Prosper Tiro, Cassiodorus, and Jordanes (Get. 103 with Mommsen ad loc). Zosimus (1.23) and Zonaras (12.20), on the other hand, locate the battle on the Tanais (!), Victor (29.4) on the far side of the Danube, while Eutropius (9.4) and the Epitome de Caesaribus (29.3) simply in barbarian territory. The latter reports that his body was lost in the marshes, a fact also known to Ammianus (31.13.13; cf. 5.15); see further Potter, op. cit. (n. 9), 278–82.

²⁸ The tradition that Philip was a Christian is recorded by Eusebius, (HE 6.34Google Scholar; cf. 36.3) and defended by some modern scholars, e.g. Sordi, op. cit. (n. 10), 96–9, but Pohlsander, H. A., ‘Philip the Arab and Christianity’, Historia 29 (1980), 463–73Google Scholar, has raised cogent objections to it; it is possible, however, that Philip had an interest in Christianity without actually being a Christian himself: G. Bowersock, Roman Arabia (1983), 125–7. Pohlsander, op. cit. (n. 25), has also shown that reports of Decius' hostility to Philip are unreliable, and so dismisses Eusebius' interpretation of his decree. Yet he overlooks the ‘prediction’ in the Thirteenth Sibylline Oracle that immediately after Decius' accession ‘there will be robberies and murders of the faithful on account of the former king’ (ll. 87–8), which suggests that a version of Eusebius' story was current during or shortly after Decius' reign; hence it is possible that Philip's apparent interest in the Church may have led some Christians to interpret Decius' perceived hostility as a reaction to his predecessor: see Clarke, op. cit. (n. 5), I, 25, and Potter, op. cit. (n. 9), 267–8.

²⁹ The History of the Decline and Fall of the Roman Empire (ed. Bury, J. B.) vol. II (1909), 121Google Scholar.

³⁰ Knipfing, op. cit. (n. 4), no. 3. Petesouchos was one of the forms under which the crocodile god Sobek was worshipped in the Arsinoite nome: Thissen, H.-J., Lexikon der Ägyptologie 4 (1982), 994Google Scholar; Moeris was a village that had been incorporated into the metropolis of Arsinoe: A. Calderini and S. Daris, Dizionario dei nomi geografici e topografici dell'Egitto greco-romano, s.v.; we may assume that the priestess of its chief deity was locally both well known and respectable. Knipfing, op. cit. (n. 4), 359–61, points out that a few of the names in the Oxyrhynchus papyri (nos 4 and 33) have possible Christian associations, although such cases are neither definite nor numerous.

³¹ Keresztes, op. cit. (n. 10), II, 59; contra, Molthagen, op. cit. (n. 10), 64 and 70–2.

³² Lane Fox, op. cit. (n. 8), 455; contra, Potter, op. cit. (n. 9), 262 n. 174.

³³ All the datable papyri were issued between 12 June and 14 July A.D. 250, while Cyprian's first explicit reference to libelli occurs in a letter written probably in July 250 (Ep. 20.2.2 with Clarke ad loc.); Clarke suggests there may also be an allusion in Ep. 15.3.1, dating to May 250.

³⁴ Keresztes, op. cit. (n. 10), II, 59–64, who believes that the decree from the start affected Christians alone, argues that the authorities, instead of continuing to insist that recalcitrant Christians sacrifice, decided to give them the option of receiving a document that falsely stated they had sacrificed; the point of these official fictions was ‘to make the demoralization of the Christian Church even more deep’. Such a strategy assumes on the part of Decius a sophistication and a knowledge of Christian doctrine that I find unbelievable.

³⁵ See n. 30. Both Keresztes, op. cit. (n. 10), II, 61, and Lane Fox, op. cit. (n. 8), 456, suggest that the priestess of Petesouchos was a secret or former Christian, or had Christians in her family; although this hypothesis cannot be excluded, it seems strained.

³⁶ We do hear of people who sacrificed once but were then ‘tested’ again: Cypr., , Ep. 24.1.1Google Scholar, Laps. 13. The circumstances are unknown, but presumably involved either bureaucratic confusion or a particular animosity towards Christians.

³⁷ Cypr., , Epp. 22.1.1Google Scholar and 39.2.1 with Clarke, op. cit. (n. 10, 1969), 63–8.

³⁸ Cypr., , Ep. 10.2.2Google Scholar (incendium persecutionis), 12.2.2 (persecutionis tempestas), and 17.1.1 (infestatio persecutionis); cf. Epp. 43.7.2, 55.6.1, 66.4.1, Laps. 5 and 15. Dionysius ap. Eus., , HE 6.40.2Google Scholar and 41.1 (diôgmos). There is no explicit mention of a persecution in the body of the Passio Pionii, but the redactor has dated it to tbe time of ‘the persecution under Decius’ (2.1).

³⁹ Clarke, op. cit. (n. 10, 1973); he shows that ordinary Christians as well as clergy were arrested from an early date and argues that the use of the plural edicta (see n. 12) is rhetorical; see also Molthagen, op. cit. (n. 10), 67. The earlier decree of Maximinus against the clergy (cf. Eus., , HE 6.28Google Scholar), frequently alleged as a precedent for Decius, probably did not exist: see e.g. Clarke, G. W., ‘Some victims of the persecution of Maximinus Thrax’, Historia 15 (1966), 445–53Google Scholar and Sordi, op. cit. (n. 10), 92.

⁴⁰ See especially de Ste. Croix, G. E. M., ‘Aspects of the “Great” Persecution’, HTR 47 (1954), 75–114CrossRefGoogle Scholar.

⁴¹ Cypr., , Ep. 5.2.1Google Scholar; see further Molthagen, op. cit. (n. 10), 65, and Pohlsander, op. cit. (n. 10), 1839–40.

⁴² Eusebius' view that Decius was hostile to Christianity because of his hatred of Philip is discussed at the beginning of this section. Cyprian says that Cornelius of Rome became bishop when there was an emperor who ‘would receive with much greater patience and forbearance the news that a rival emperor was raised against him than that a bishop of God was established in Rome’ (Ep. 55.9.1). Although Frend, op. cit. (n. 10), 405, takes this as an actual remark of Decius, it is clear that Cyprian meant it not as a quotation but as a characterization of Decius and the menace that he posed to a new bishop: Clarke, op. cit. (n. 5), III, 178–9. The method of portraying character through invented thoughts or speech was highly developed in ancient rhetorical theory: see e.g. [Cic., ], Rhetorica and Herennium 4.63–5Google Scholar.

⁴³ For an example of this attitude in non-Christians, see Eus., , HE 7.11.9Google Scholar.

⁴⁴ So, for example, Frend, op. cit. (n. 10), 405; Molthagen, op. cit. (n. 10), 73–8; Clarke, op. cit. (n. 5), I, 23, and especially Pohlsander, op. cit. (n. 10), 1829–31.

⁴⁵ Alföldy, G., ‘Die Krise des Imperium Romanum und die Religion Roms’, in Eck, W. (ed.), Religion und Gesellschaft in der römischen Kaiserzeit: Kolloquium … F. Vittinghoff (1989), 53–102.Google Scholar

⁴⁶ Gerov, op. cit. (n. 24); cf. PIR ² M 520 and Peachin, op. cit. (n. 11), 239–52. ‘Traianus’ is consistently found in the documentary evidence: of the ninety-one variations of his name listed by Peachin, only ten omit it; in contrast, it does not appear in any literary text except Pass. Pion. 23, and an allusion in Sib. Or. 13.83 (see Potter ad loc). On Trajan's reputation, see Syme, op. cit. (n. 19), 220 and, in general, 89–113.

⁴⁷ Mattingly, H., Sydenham, E. A., Sutherland, C. H. V., The Roman Imperial Coinage IV. 3 (1949), 107–50Google Scholar; the emperors depicted are Augustus, Vespasian, Titus, Nerva, Trajan, Hadrian, Pius, Marcus Aurelius, Commodus, Severus, and Alexander. Mattingly's attempt to attribute this series to the mint of Milan is now rejected: Elks, K. J. J., ‘Reattribution of the Milan coins of Trajan Decius to the Rome mint’, Numismatic Chronicle 12 (1972), 111–15Google Scholar. K. E. T. Butcher has recently suggested that this series of portraits may have simply been a form of compensation to these emperors for the overstriking of their issues: ‘Imagined emperors: personalities and failure in the third century’, JRA 9 (1996), 514–27Google Scholar, at 522–3 n. 15; yet it must be more than coincidence that only deified emperors appear.

⁴⁸ There exists a letter from Decius and his son Herennius Etruscus to Aphrodisias in which they thank the city for honouring the establishment of their rule and making ‘the proper sacrifice and prayers’: J. Reynolds, Aphrodisias and Rome (1982), no. 25. Although some scholars have connected this with the decree on sacrifices, e.g. Frend, op. cit. (n. 10), 406, it instead records the long-established custom that a city would celebrate the accession of a new emperor and send to him ambassadors to announce that fact in the hope of winning his favour: Millar, op. cit. (n. 12), 410–20; hence the confirmation of privileges and the mention of ambassadors at the end. More relevant is an inscription from Aquileia recording the public restoration of a statue of Neptune ‘by order’ of an emperor whose name has been erased but whom the editor identifies as Decius: Brusin, J. B., Inscriptions Aquileiae I (1991), no. 326Google Scholar. If this identification is correct, the inscription would date to the very end of A.D. 250, and would further attest an interest in the cult of traditional gods.

⁴⁹ AE 1973.235; the designation of Decius as ‘cos. III’ dates it to the first half of A.D. 251, pace W. H. C. Frend, The Rise of Christianity (1984), 319, who implies that it was erected near the beginning of his reign. See further Babcock, C. L.., ‘An inscription of Trajan Decius from Cosa’, AJP 83 (1962), 147–58Google Scholar; Marelli, U., ‘L'Epigrafe di Decio a Cosa et l'epiteto di “Restitutor Sacrorum”’, Aevum 58 (1984), 52–6Google Scholar; and Lane Fox, op. cit. (n. 8), 453. Babcock plausibly suggests that Decius may have sponsored the rededication of the temple in which the inscription was found.

⁵⁰ As correctly noted by Sordi, op. cit. (n. 10), 104. and 107 n. 17.

⁵¹ C. Sourvinou-Inwood, ‘What is polis religion?’, in O. Murray and S. Price (eds), The Greek City from Homer to Alexander (1990), 295–322, at 302; see further eadem, ‘Further aspects of polis religion’, Annali, Sezione di archeologia e storia antica: Istituto Universitario Orientale di Napoli 10 (1988), 259–74Google Scholar. 482–501; Rives, op. cit. (n. 53), 51–63 and 85–96.

⁵² Sourvinou-Inwood, op. cit. (n. 51, 1990), 295–8.

⁵³ Rives, J. B., Religion and Authority in Roman Carthage from Augustus to Constantine (1995), 5–12 and 28–51.Google Scholar

⁵⁴ Price, S., Rituals and Power: The Roman Imperial Cult in Asia Minor (1984), 62–77Google Scholar; Fishwick, D., The Imperial Cult in the Latin West (2 vols, 1987–1992), II, 482–501Google Scholar; Rives, op. cit. (n. 53), 51–63 and 85–96.

⁵⁵ For the latter, see Snyder, W. F., ‘Public anniversaries in the Roman Empire’, YCS 7 (1940), 225–317Google Scholar, and Herz, P., ‘Kaiserfeste der Principatszeit’, ANRW II.16.2 (1978), 1135–200Google Scholar.

⁵⁶ Arval rites: Scheid, J., Romulus et ses frères. Le collège des Frères Arvales, modèle du culte public dans la Rome des empereurs (1990), 290–383Google Scholar; Cyrenaican inscriptions: Reynolds, J. M., ‘Vota pro salute Principis’, PBSR NS 17 (1962), 33–6Google Scholar, and ‘Notes on Cyrenaican inscriptions’, PBSR NS 20 (1965), 52–4Google Scholar.

⁵⁷ Epp. 10.35, 36, 100 and 101; Epp. 10.52, 53, 102, and 103 deal similarly with the prayers and oaths made on the emperor's dies imperii (cf. Sherwin-White ad Ep. 10.35).

⁵⁸ Pertinax, : BGU II.646Google Scholar = Hunt, A. S. and Edgar, C. C., Select Papyri II: Non-Literary Papyri, Public Documents (1934), no. 222Google Scholar; see also P. Oxy. VII.1021 = Select Papyri II no. 235, on Nero; P. Oxy. LV.3781, on Hadrian; Sijpesteijn, P., ‘Edict of C. Calvisius Statianus’, ZPE 8 (1971), 186–92Google Scholar = SB XII.10991, on Avidius Cassius; SB 1.421, on Maximus, the son of Maximinus Thrax; P. Oxy. LI.3607, on Gordian I and II. Note also IG II².1077, a decree from Athens for celebrations in honour of Geta's accession to imperial status, prompted by the decree of an imperial legate.

⁵⁹ It is not clear whether the key unit was the person or the household. A number of the Egyptian certificates were filed by households: two brothers and their wives (Knipfing, op. cit. (n. 4), no. 2), a man with his son and daughter (no. 4), and women with one or more children (nos 7, 30–2, 34–6). In one case the petitioner explicitly states that he performed the sacrifice for his wife, two sons, and daughter (no. 33). According to Cyprian, some men made their entire families sacrifice, while others took it upon themselves to do so on their behalf (Ep. 55.13.2); he refers to parents who carried their infants and small children to the altar (Laps. 9) and to a man who forced his wife to sacrifice (Ep. 24.1.1), and relates an anecdote about a baby girl whose nurse fed her bread dipped in sacrificial wine (Laps. 25). It thus seems that although the head of a household could fulfill the terms of the edict on behalf of the entire family, it was relatively common for all members of the household to take part individually. In what follows, ‘individual’ means either the single person or the single household.

⁶⁰ See especially Wissowa, G., Religion und Kultus der Römer (2nd edn, 1912), 398–400Google Scholar; for the suspension of business, see Varr., , Ling. 6.30Google Scholar and 6.53, and ap. Macr., , Sat. 1.16.19Google Scholar.

⁶¹ In both the Greek and the Italic traditions there was also extensive private worship by individuals, and Sourvinou-Inwood, op. cit. (n. 51, 1988), 264–7, has cogently argued that in Greek religion at least the individual was the basic cultic unit. But this was not a part of public cult as such.

⁶² Eretria: SIG ³ 323 = Sokolowski, F., Lois sacrées des cités grecques, supplément (1962), no. 46, ll. 6–7Google Scholar; Lampsacus: Sokolowski, F., Lois sacrées de l'Asie Mineure (1955), no. 8Google Scholar, ll. 18–19 and 23–4, with further references ad loc; see also LSAM no. 81 and OGIS 6 and 219.

⁶³ See Wissowa, op. cit. (n. 60), 423–6, and at RE IV A (1931). 942–51Google Scholar.

⁶⁴ ILS 108; cf. Fasti Amiternini (Degrassi no. 25) on 3 September, and especially Res Gestae 9.2: ‘privatim etiam et municipatim universi cives unanimiter continenter apud omnia pulvinaria pro valetudine mea supplicaverunt’.

⁶⁵ OGIS 219 = I. Ilion 32, with the emendation of L. Robert, ‘Un décret d'Ilion et un papyrus concernant des cultes royaux’, in Essays in Honor of C. Bradford Welles, American Studies in Papyrology I (1966), 175–211, at 183–92.

⁶⁶ SIG ³ 695 = LSAM no. 33, 11. 43–5 and 86–90; see also 11. 7–10.

⁶⁷ Veyne, P., ‘Les honneurs posthumes de Flavia Domitilla et les dédicaces grecques et latines’, Latomus 21 (1962), 49–98Google Scholar, at 71–5; see further Price, op. cit. (n. 54), 112.

⁶⁸ OGIS = I. Ephesos I a. 21.

⁶⁹ Plin., , Ep. 10.100Google Scholar: ‘nova [vota] certante commilitonum et provincialium pietate suscepimus’; cf. Epp. 10.35, 36 and 101.

⁷⁰ In Fink, R. O., Hoey, A. S., and Snyder, W. F., ‘The Feriale Duranum’, YCS 7 (1940), 1–222Google Scholar, at 190–202.

⁷¹ See in general Herz, op. cit. (n. 55), 1189–93; Price, op. cit. (n. 54), 110–14; Fishwick, op. cit. (n. 54), ll, 528–32.

⁷² Since the officials who tried Christians often demanded that they make an offering to the emperor, it has often been thought that such offerings were legally required. Yet the first and most detailed account of this practice shows that it was simply a way of proving that the accused was not a Christian: Plin., , Ep. 10.96.5–9Google Scholar with Millar, F., ‘The imperial cult and the persecutions’, in den Boer, W. (ed.), Le culte des souverains dans l'empire romain, Fondation Hardt, Entretiens 19 (1973), 145–65Google Scholar, at 152–5; see further Fishwick, D., ‘Pliny and the Christians: the rites ad imaginem principis’, AJAH 9 (1984), 123–30Google Scholar, and op. cit. (n. 54), II, 533 with n. 352.

⁷³ Clarke, op. cit. (n. 5), I, 25, tentatively connects it with the imperial vota; the suggestion of Alföldi, op. cit. (n. 10), 333, that the occasion was a dies Imperii is rightly criticized by Frend, op. cit. (n. 10), 407. The only reference to the emperor is in the Passio Pionii (8.4), where the neokoros Polemon, after his other attempts have failed, suggests to Pionius to ‘sacrifice at least to the emperor’; the concessive phrasing of this remark hardly suggests that the emperor was the chief focus of the decree. Robert, op. cit. (n. 8), sees the whole thing as ‘a matter of loyalty’, and considers Polemon as the neokoros of an imperial temple in which the emperor was associated with the Nemeseis (at Pass. Pion. 3.1 and 8.4). But none of the imperial temples in Smyrna seems to have been jointly dedicated to the Nemeseis: Price, op. cit. (n. 54), 258; nor does neokoros, when used as a personal rather than a civic title, necessarily imply imperial cult.

⁷⁴ The only contemporary evidence for Philip's millennial celebrations is his coinage: Mattingly et al., op. cit. (n. 47), 70–89. One common type has on the reverse the legend ‘Saeculares Augg.’, with images of wild animals; variants show the she-wolf and the twins, and a column inscribed ‘Cos. III’. Another series has on the reverse ‘Saeculum Novum’ with the depiction of a temple and a cult statue, probably that of Roma. Some bronze medallions with the same legend show a similar temple, with a scene of sacrifice before it, while others show chariot races: Cohen, H., Description historique des monnaies 5 (1885), 139–9Google Scholar nos 12–14, and 170 no. 82. Later references occur in the Chronicle of Eusebius (Schöne 180–1), Aurelius Victor (28.1), Eutropius (9.3), HA Gordiani (33.1), and Orosius (7.20.2), most of whom say only that Philip celebrated the anniversary with splendid games. The evidence thus does not suggest that Philip's millennial celebrations had a significant cultic component, although they no doubt involved prayers and sacrifices.

⁷⁵ It is significant that Decius' coinage does not continue the theme of a saeculum novum.

⁷⁶ P. Giss. 40; among the numerous studies, see especially Bickerman, E., Das Edikt des Kaisers Caracalla in P. Giss 40 (1926)Google Scholar; Stroux, J., ‘Die Constitutio Antoniniana’, Philologus NS 42 (1933), 272–95Google Scholar; Wilhelm, A., ‘Die Constitutio Antoniniana’, AJA 38 (1934), 178–80CrossRefGoogle Scholar; d'Ors, A., ‘Estudios sobre la “Constitutio Antoniniana', V: Caracalla y la unificacion del imperio’, Emerita 24 (1956), 1–26Google Scholar; Wolff, H., Die Constitutio Antoniniana und P. Giss. 40 (1976)Google Scholar; K. Buraselis, Theia Dôrea: Studies on the Policy of the Severans and the Constitutio Antoniniana (1989, in Greek).

⁷⁷ P. Oxy. IV.658 = Knipfing, op. cit. (n. 4), no. 4, in Knipfing's translation.

⁷⁸ P. Oxy. L.3565, in the translation of J. R. Rea.

⁸⁰ I am here developing an observation first made by Potter, op. cit. (n. 9), 43: ‘when they sacrificed they would obtain a certificate … just as they received a receipt from the tax-collectors whenever they paid their taxes: indeed, the procedure in this edict appears to parallel the process of tax collection very closely’.

⁸⁰ See especially Knipfing, op. cit. (n. 4), 346–54, and Clarke, op. cit. (n. 5), I, 30–2.

⁸¹ R. S. Bagnall and B. W. Frier, The Demography of Roman Egypt (1994), 1–30.

⁸² P. Oxy. XLVI.3295, in the translation of J. R. Rea; see further Montevecchi, L., ‘Ricerche di sociologia nei documenti dell'Egitto greco-romano. VI: Denunce di nascita dei greco-egizi’, Aegyptus 27 (1947), 3–24Google Scholar, and Schulz, F., ‘Roman registers of birth and birth certificates’, JRS 32 (1942), 78–91Google Scholar and 33 (1943), 55–64. On death certificates, see Sijpesteijn, P. J., ‘A document concerning registration of deaths’, ZPE 52 (1983), 282–4Google Scholar and L. Casarico, Il controllo della popolazione nell'Egitto romano I: he denunce dimorte (1985).

⁸³ P. Batav. I.14, in the translation of P. J. Sijpesteijn. On the laographia, see further S. L. Wallace, Taxation in Egypt from Augustus to Diocletian (1938), 116–34.

⁸⁴ See in general P. A. Brunt, Roman Imperial Themes (1990), 324–46. There is a land-declaration made during a census in Arabia: N. Lewis, The Documents from the Bar Kochba Period in the Cave of Letters: Greek Papyri (1989), no. 16, with Cotton, H. M., ‘Another fragment of the declaration of landed property from the province of Arabia’, ZPE 99 (1993), 115–21Google Scholar; the census in Judaea is attested by Luke (2:1–5), while Ulpian's discussion of what should be included in census returns (Digest 50.15.3–4) is clearly meant to have general application.

⁸⁵ Oaths of loyalty constitute another parallel, although more partial. Although everyone in the Empire was required to swear such an oath, the process seems to have been a collective one: one inscription is headed ‘senatus et populus Go[nobariensium] in ea ver[ba iuraverunt]’: González, J., ‘The first oath pro salute Augusti found in Baetica’, ZPE 72 (1988), 113–27Google Scholar. The text of an oath of allegiance to Augustus from Gangra in Paphlagonia (ILS 8781) is put in the first person singular (‘I swear by Zeus’), but the heading is again collective, ‘the oath taken by the inhabitants of Paphlagonia’. There is certainly no evidence of a procedure for individual certification.

⁸⁶ Notably Molthagen, op. cit. (n. 10), 63 and 73–5, following Lieserung, op. cit. (n. 10), 37–43; see further Pohlsander, op. cit. (n. 10), 1838. Both Knipfing, op. cit. (n. 4), 354, and Alföldi, op. cit. (n. 10), 329, pointed out problems with this interpretation.

⁸⁷ On supplicationes, see above at n. 63.

⁸⁸ On religious competition, see North, J., ‘The development of religious pluralism’, in Lieu, J., North, J. and Rajak, T. (eds), The Jews among Pagans and Christians (1992), 174–93Google Scholar; on the relative decline in importance of civic cults, see Rives, op. cit. (n. 53), 173–249. I do not mean by this that civic cults lost meaning and importance, only that they suffered from unrestricted competition; for their continued vitality, see e.g. MacMullen, R., Paganism in the Roman Empire (1981), 126–30 and Lane Fox, op. cit. (n. 8), 574–85Google Scholar.

⁸⁹ We may assume that Dexippus did not report it, since there is nothing in George Syncellus that could have come from him: George took from Eusebius his account of the persecution (683–704 Dindorf; cf. Eus., , HE 6.39–44Google Scholar), and then followed Dexippus for Decius' death and the beginning of Gallus' reign; if Dexippus provided any information on the persecutions, we might expect George to have made some use of it. We have no way of knowing about other third-century historians, although we may note that Zosimus, who probably used a source other than Dexippus (above at n. 18), is also silent: cf. Syme, op. cit. (n. 19), 200. But since Zosimus also passes over the persecutions of Valerian and the Tetrarchs, his silence may be due more to policy than ignorance.

⁹⁰ Capitoline Triad in Carthage: Cypr., Laps. 24: ‘Unus ex his qui sponte Capitolium negaturus ascendit’; Ep. 59.13.3: ‘conpelluntur adhuc insuper lapsi ut linguis atque ore quo in Capitolio ante deliquerant sacerdotibus convicium faciant’. In Rome too the Capitol may have been the scene of the sacrifices: see Cypr., , Ep. 8.2.3Google Scholar and 21.3.2 with Clarke ad loc., while a remark of Pacian (Ep. 2.3) suggests that in the fourth century Novatianists used Capitolini as a derogatory term for Catholics. Nemeseis in Smyrna: Pass. Pion. 6.3 and 7.2.

⁹¹ Some scholars have suggested that Decius insisted on the traditional state gods of Rome: Frend, op. cit. (n. 49), 320, claims that ‘sacrifice was made on the capitols to the specifically Roman gods and the emperor's genius, rather than to local gods’; this is clearly untrue in the case of the Nemeseis, who were important in Smyrna but never received public cult in Rome; as for the emperor's genius, see above at n. 73. There is also a striking passage in the Passio Pionii where the proconsul, frustrated with Pionius' insistence on praying to God, says ‘we all honour the gods and heaven and the gods in heaven; do you look to the air? Sacrifice to it’ (Pass. Pion. 19.10). His suggestion implies very strongly that any deity would do, even the god of the Christians, just so long as a sacrifice was performed. It thus seems best to accept at face value the texts of the Egyptian papyri, which refer simply to ‘the gods’: so already Knipfing, op. cit. (n. 4), 353, and rightly stressed by Millar, op. cit. (n. 72), 159–60, and Clarke, op. cit. (n. 5), I, 25–6.

⁹² Gordon, R., ‘The veil of power: emperors, sacrificers, and benefactors’, in Beard, M. and North, J. (eds), Pagan Priests (1990), 199–231, at 207Google Scholar; since Gordon sees the sacrificial system as deeply implicated in the social control exercised by local élites, it is interesting to note that Decius' edict effectively levelled out distinctions of status in favour of cultic homogeneity.

⁹³ Acta Cypriani I: ‘eos qui Romanam religionem non colunt, debere Romanas caeremonias recognoscere’; that something like this was in the actual edict is suggested by the fact that the prefect of Egypt used a broadly similar phrase to Dionysius of Alexandria: Eus., , HE 7.11.7Google Scholar.

⁹⁴ Lact., , Mort. Pers. 34.3Google Scholar: ‘nostra iussio extitisset, ut ad veterum se instituta conferrent’; cf. 34.2: ‘tanta eosdem Christianos voluntas invasisset… ut non illa veterum instituta sequerentur’. Palace and army: Lact., Mort. Pers. 10 and Div. Inst. 4.27.4–5; third edict: Eus., , HE 8.6.10Google Scholar and MP praef. 2; fourth edict: Eus., , MP 3.1Google Scholar. See in general de Ste. Croix, op. cit. (n. 40); on the importance of traditional cult acts in the policy advocated by Porphyry, see Digeser, E. D., ‘Lactantius, Porphyry and the debate over religious toleration’, JRS 88 (1998), 129–46Google Scholar.

⁹⁵ Eus., , HE 8.14.8–9Google Scholar and 9.4.2–3; Lact., , Mort. Pers. 36.4–5Google Scholar.

⁹⁶ See, e.g., Fowden, op. cit. (n. 2), 80–90.