The account on the verso of P.Oxy. III 521 was first mentioned in the introduction to P.Oxy. III 521, but it has never been fully published. There, Bernard Grenfell and Arthur Hunt mentioned the entries in lines 6 and 8 in order to strengthen their supposition about the list of objects on the recto. The list of objects is written along the fibres, and a kollesis is visible on the left hand side. Also, «to the left of the list on the recto are a few letters of the ends of lines of an account in a different hand» (P.Oxy. III p. 257). The account on the verso is written across the fibres.

The few letters on the recto run across the kollesis, and so does the account on the verso: the kollesis itself does not generally correspond to the place where the papyrus is broken; the break occurs immediately before or after it². Running over the kollesis, the remains of an account on the recto indicate that a single roll was used for different documents concerning the temple’s administration. The document published as P.Oxy. III 521 recto was therefore not intended to be sent to officials, but for internal use only. The same applies to the verso. Also, the roll was used as a whole. The single documents follow one another in space and time, beginning on the recto and ending on the verso. There is a Y-shaped hole on the level of lines 15-16, but the papyrus is otherwise well preserved.

To sum up, the account is on the verso, and is written across the fibres. It is part of a roll containing several documents concerning the temple’s administration. These documents were for internal use only, and they were discarded after they had served their purpose. Also, the account was written at a later date than the list of objects on the recto.

The account provides us with some new insights into temple administration, and it informs us about the birthday celebrations of emperors. First, it lists a large amount of expense for birthday celebrations of deified emperors. Second, it is the first internal temple document mentioning bankers and money transactions. Third, it is the first temple account using brackets.

The birthdays of the emperors, either reigning or already dead at the time, were celebrated in Sebasteia, probably also named Kaisareia, as well as in temples of the Egyptian gods, where they had been integrated into the temple cult as sunnaoi³. Sometimes, monthly celebrations of an emperor’s birthday can be found under the name Cεβαсτὴ ἡμέρα⁴. Here, we are concerned with the birthday celebrations of divus Nerva and divus Vespasian. Similar entries in two accounts of the temple of Soknopaios at Soknopaiou Nesos in the second century (SPP XXII 183 [later than 14-22 August, AD 138?]; P.Louvre I 4 [later than AD 166]) do not come close to the 500 dr. of our account.

To my knowledge, there is no other documented account of a temple, mentioning bankers transferring money from private people or officials to a temple for certain celebrations⁵. These transactions are pinned to specific dates put between brackets⁶. The temple to which this papyrus belongs, is probably the Kaisareion in Oxyrhynchus, because of the large amount paid for Vespasian’s birthday celebration⁷.

The recto is dated to the second century AD. The verso gives no terminus post quem for the papyrus fragment, but only for the account, which is AD 98, the death of Nerva. The situation of banks and bankers in Oxyrhynchus in the second century and a possible location of the temple might even allow us to date the account to the middle of the second century AD (see below, comm. on lines 9, 12 and 15).

5.8 x 17.2 cm          Mid-second century AD

]κ̣αι χωρα[

].η´ φεοδ̣[

] αс ̅λβ̅ ουс[

]δογμάτω(ν) (δρ.) η´ λόγω̣ν̣[

5 ]Ἀππιανοῦ (δρ.) σ´ ὑπογ̣[

]γενεсίω(ν) θεοῦ Νέρουα . [

]. ου (δρ.) σ´ ὑπογραφῆc.[

γενεс]ίω(ν) θεοῦ Οὐεс(παсιανοῦ) (δρ.) φ´[

] καὶ ὑπὸ Διδύμου ` ούκ ´ φ[

10 ]ου τοῦ οἰκ( ) ετ̣ι̣κ̣ω̣ λο[

]υ̣ Οὐλπίου Kλέωνoc [

].̣ υ̣χ() καὶ Ἱέρακο(c) τρα(πεζιτῶν) (θ[

] .I) (δρ.) Γφξ · διὰ δὲ τῆc. [

τ]ρ̣α̣(πεζίτου) (Ἁθ(ὺρ) ̅δ) (δρ.) ση´[

15 Διδ]ύμο(υ) τρα(πεζίτου) (ὁμ(oιῶc) Ἁθ(ὺρ) ̅δ) (δρ.) σ [

]. ναρίο(υ) τρα(πεζίτου) (Χο(ιὰκ) ς) (δρ.) [

]ο̣υ τοῦ πρoκεχειριсμ[ένoυ

]. ηс Τίτου Φλαουίου [

]υсηс Kλα̣υδίαс Ἀπολλω[

20 ]ο̣c (Θὼθ ̅κ̅ζ) (δρ.) ση´[

]. c καὶ Τιβερίου Kαίсαρo[c

]…. κ̣αὶ̣ ̣ επ̣ροc̣[

. . . . .

7 ϋπoγραφηс      12 ϊερακο      14 (α̅θ ̅δ) pap.      15 (ομ α̅θ ̅δ) pap.      16 (Χο ς) pap.      20 (Θὼθ κ̅ζ̅) pap.

(…) Appianos 200 dr. (…) birthday of divus Nerva (…) 200 dr. for the subscription (…) birthday of divus Vespasian 500 dr. (…) and by Didymos not (…) the household (?) account (…) Ulpius Kleon (… NN) and Hierax, bankers (…) 3560 dr. (…) through (…) banker, Hathyr 4^th, 208 dr. (…) Didymos, banker, also Hathyr 4^th, 200 dr. (…) -narios, banker, Choiak 6^th, (…) dr. (…) treated previously (…) Titus Flavius (…) Claudia Apollo- (…) Thoth 27^th, 208 dr. (…) and of Tiberius Caesar (…).

6 θεοῦ Νέρουα . [. Grenfell and Hunt (P.Oxy. III p. 257) additionally read αδ after Νέρουα, but I cannot see it on the papyrus. There is only a small illegible trace following the α of Νέρουα. It seems that the papyrus was in a better condition then. The title θεόc for emperors need not indicate that these emperors had already been deified. See Price (1984) and – especially for Egypt – Whitehorne (1992) 424-434, followed by Pfeiffer (2010) 281-283. For the possibility of using this title as a translation of the Egyptian nṯr «god» after the deification of an emperor, see Whitehorne (1991) 23-24. White-home’s distinction is based on the word order: θεόc + name stands for divus + name, while name + θεόc normally implies that the title had already been used during an emperor’s lifetime (Whitehorne [1992] 423-424), most likely in an Egyptian priestly milieu. Therefore Nerva and Vespasian had already been deified when the account was written. This gives a terminus post quem, i.e. Nerva’s death (AD 98).

8 (δρ.) φ´. Grenfell and Hunt (p. 257) read ω (800 dr.), but there are traces of an upward and a downward line in the middle of the letter. Therefore I prefer φ to ω. Even then, 500 drachmai is an unusually high amount of money for the birthday of Vespasian in line 8, for it exceeds all other amounts of temple expenditure for imperial or divine birthdays. The amounts for the birthday celebrations from two other lists of temple expenditure amount to only 8 percent of what appears in our list. See SPP XXII 183, 51-53 (later than August 14-22, AD 138?), with BL II.2 167 for line 53, indicating that the entry is for the deified emperors; P.Louvre I 4, 35-36 (later than AD 166). Both accounts are from the temple of Soknopaios in the village of Soknopaiou Nesos. A third list from the Arsinoite nome, BGU II 362 (AD 215/216), lists the total amount of about 100 drachmai for the birthday celebrations of the emperor Caracalla shortly before his death (fr. ii, p. X 9-19; around 20 percent of the figure appearing in our list), while only 58 drachmai are mentioned for the deified emperor Septimius Severus (fr. ii, p. XI 8-14; around 10 percent of the list), showing that the amount for deified emperors may even be lower than for the living emperor. However 500 dr. were spent for the κῦφι for the god Soknopaios in Soknopaiou Nesos according to SPP XXII 183, 54-55 and P.Louvre I 4, 33. Soknopaios is the main god of the temple and even for his sunnaos 96 drachmai were spent for κῦφι. It was an Egyptian temple of the first order with some temples under its control (Rübsam [1974] 154-172). The only celebrations for emperors mentioned in both accounts are for their birthdays. The 500 drachmai for the κῦφι of the main god are per annum, and so is the amount for the birthday celebrations (γενεοίων is a genetivus relationis, see Mayser, Grammatik II.2, § 83.3 [192-194]; in P.Louvre I 4, 35, γενεсί[oιс] should probably be changed to γενεсί[ων]). SPP XXII 183, 51-53 uses the genetivus relationis, and it is the only parallel so far. On the one hand the word order is different, but it does not necessitate a change in the case. On the other hand the context is completely the same. Both documents have these lines in the λόγοс ἀργυρικόc, while the dative is used in the λόγοс сιτικόс (SPP XXII 183, 68-69 and 86-87; P.Louvre I 4, 50-51 and 67-68; for this part, see also SB VI 9199, 10 [mid-second century AD]). From these amounts, it seems likely that the temple in question should have been a place dedicated to the worship of the emperors, i.e. a Sebasteion/Kaisareion. A Hadrianeion is so far excluded due to the facts that: a) such temples were erected especially for the worship of Hadrian; b) other emperors were worshipped there only if integrated as sunnaoi at a later date (Blumenthal [1913] 322). Vespasian and Nerva reigned before Hadrian. Therefore the temple should be a Sebasteion/Kaisareion. Still, 500 drachmai for only a single celebration seems to be high. Perhaps it is the entry for the whole year, and the birthday was celebrated monthly (Blumenthal [1913] 339-340).

9 Διδύμου. There are many men bearing the name Δίδυμοс at Oxyrhynchus in the second century AD. This Δίδυμοс, however, was expected to give 500 drachmai, as the οὐκ above the line indicates, and he is not styled τραπεζίτηс as the Didymos in line 15. The Diydmos in line 9 was either a wealthy private person, or a cultic official. Supposing that specifications of persons were given in an internal account only if deemed necessary, he may have been the magistrate responsible for the cult of the emperors in Oxyrhynchos. This would include his responsibility for the Sebasteion (see P.Oxy. XLII 3072, 5 [ca. AD 197-200]) and the Hadrianeion (see P.Oxy. LXXI 4827, 1-3 [ca. Sept. – Dec., AD 173]). If a high priest of the Hadrianeion could fulfill this magisterial function, this man could also be identified with the Δίδυμοс, also called Διδυμίων, of P.Oxy. VIII 1113, 3-6 (26 Jan. – 24 Feb., AD 203) with BL VIII 241 (this entry suggests the identity with the Didymos of P.Oxy. XLIX 3492, 9 [AD 161-169]) and of P.Oxy. XLIX 3492, 9. This would imply a date in the middle of the second century AD.

10 τοῦ οἰκ( ) ετ̣ι̣κ̣ω̣ λο[. The sequence ετ̣ι̣κ̣ω̣ remains unclear to me. The kappa is above the line as in the other abbreviations. It is, therefore, likely that it also is a sign for abbreviation. Still, I would consider two other readings. One could read τοῦ οἰκετικῶ λο[, to be corrected into τοῦ οἰκετικοῦ λο[. If λό[γου is read, then this list will either be considered a household account (of a different sort than SPP XXII 183 and P.Louvre I 4), or parts of the list are somehow connected to a household account. If λογ[ιсτηρίου is read, then there is some connection to a financial department of the temple administration. A model of temple administration is given by Stead (1984) 1050. Or perhaps one should read τοῦ οἰκ( ) ἐπακολο[.

12 Ἱέρακο(c) τρα(πεζίτου). The amount was transferred from an unknown private account to the account of the temple in question. On banks and bankers in Ptolemaic and Roman Egypt, see above, footnote 5. A Ἵεραξ appears as ἐπιτηρητήc of ἡ ἐπὶ τοῦ πρὸc Ὀξυρύγχων πόλει Cαραπείου τράπεζα in P.Oxy. XXXIV 2722, 7-9 (30 March, AD 154), so far, the only bank attested in Oxyrhynchus between 154 and 201 (see Bogaert [1994] 79 and 91). Whether it is possible to identify this Hierax with the one appearing here is doubtful.

14 (Ἁθ(ὺρ) ̅δ). On the use of brackets, see below, Appendix 1.

15 Διδ]ύμου τρα(πεζίτου). I do not think that this banker is the Didymos of line 9. There is a bank owner called Didymos in P.Oxy. XXXIV 2726, 24-25 (II AD); see Bogaert (1995) 155. If they are the same person and if the same holds for both Hierax (see comm. on line 12), then P.Oxy. XXXIV 2726 and P.Oxy. III 521 verso are to be dated to around AD 154.

18 Τίτου Φλαουίου [. A soldier in Oxyrhynchus (SB XX 14071, 4-5 [25 June – 24 July, AD 77]) and a prefect of Egypt (P.Oxy. I 34 verso, iii, 1 [2 October, AD 127]) bear this name.

20 The traces at the beginning do not hint at the abbreviation for τραπεζίτηс. Therefore the 208 drachmai are not given through a banker.

Appendix 1: on the use of brackets

The brackets were put around the date by the scribe (see also 12-13, 15, 16 and 20). A short look at a few documents of the second century which also have brackets may suggest that these are used as signs of deletion. First, the bracketed entries are not counted to the sums of some lists. See SB XIV 11657 (AD 138/139); P.Lond. I 119 (p. 140; later than AD 143/144?); P.Berl. Leihg. I 8 (ca. AD 163); P.Bour. 42 (AD 166/167); P.Berl. Leihg. I 2 verso (later than 6 February, AD 168); P.Petaus 39 (ca. AD 182-187). Second, the entry either comes up a second time in the list, or the name was already on the list. See P.Mich. VI 381, with BL III 316 (AD 151-200); P.Lond. II 254 verso (p. 230; II AD). Furthermore, one list concerning catoecic land displays brackets and corrections inserted above the entries. See P.Vind. Bosw. 16, 12 and 16 (II/III AD); the same applies to P.Lond. I 119, 74-75 (p. 140; later than AD 143/144?), with BL I 220 for line 75.

There are, however, five documents that display two formal characters, indicating otherwise⁸. First, there are brackets around parts of the texts and second, there are strokes for erasure between these brackets. Supposing that an official or a scribe would not use two different signs for the same purpose in the same part of the document, brackets are used to indicate something else.

The evidence becomes stronger if we allow for documents which have brackets around parts of the texts and strokes for erasure in other parts of the texts⁹. If the above supposition is slightly altered, i.e. instead of «in the same part of the document», we choose «in the same document», the evidence points to a different understanding of the two signs. The documents listed comprise documents that seem to indicate that brackets are signs of deletion. Based on formal characteristics and their use alone, the counter evidence, however, suggests that one does not explain the brackets as signs of deletion, but otherwise.

I shall therefore give a list of all documents I could classify. These documents are dated to the second century and they have the same formal characteristic, namely brackets. The list will give groups of the different uses of brackets I have encountered in the documents. The classification is based on the official versus private distinction and, additionally, on the context of each document.

1) Documents emanating from officials:

– 1.1) To group entries: BGU II 476, 2-3 and 7-10 (AD 164/165?); 561, 3-19 passim (II AD)¹⁰.

– 1.2) To indicate the addenda: P.Lond. I 119,84-94 passim (p. 140; later than AD 143/144?), with BL I 221 for 84 and 89¹¹; P.Ryl. II 215, 17-18 and 22-24 (II AD), with BL VIII 295 for 23; 216, 6-13 passim, 86-87 and 103-104 (AD 176-225), with BL VII 173 for 7 and 10; CPR VII 8 (II/III AD), with BL VIII 107¹².

– 1.3) To indicate additional information necessary for other administrative tasks: BGU IX 1891, 32, 463 and 464 (3 December, AD 133); BGU IX 1894, 131-133 and 142-143 (ca. AD 157); P.Ross. Georg. II 28, 5 (later than AD 163/164); P.Bour. 42, 281-282 (AD 166/167); SB XIV 11710, fr. A 18 and 26; fr. B 1 and 11 (AD 171/172); P.Mich. IV 223 (August, AD 172)¹³; IV 224 (later than 16 November, AD 173); IV 225 (later than 26 March, AD 175); P.Cair. Mich. 359, 1546, 1631, 1636, 1646, 1653, 1854, 1865, 1872, 1876, 1881 and 1937 (later than 25 May, AD 175); P.Mich. IV 362, 6 and 28 (AD 175); SB XII 11115, 1, 5, 14 and 18 (AD 183/184); P.Petaus 69, 30 and 37 (ca. AD 186/187); BGU II 494, 3; VII 1625, 15 (both II AD).

– 1.4) To indicate private dealings: P.Col. II 1 recto 6, ii, 24-25; iii, 15; ix, 7 (AD 157?)¹⁴.

– 1.5) To indicate unaccomplished tasks: P.Oxy. XXIV 2409, 8-9 (late II AD), with BL XI 162 for 8 (see P.Oxy. XXIV, p. 158).

– 1.6) Probably added by another official: P.Oxy. VI 957 (AD 122/123), with BL I 330; P.Lond. I 119, 74-75 (p. 140; later than AD 143/144?), with BL I 220 for 75; P.Berl. Leihg. II 28,16 and 18 (ca. AD 170); P.Vind. Bosw. 16, 2, 3, 12, 16 and 17 (II/III AD).

– 1.6.1) To indicate that a document was added to a τόμοс сυγκολλήαμοс and when it was added: P.Oxy. LX 4060, 39, 68, 81, 95 and 120 (June, AD 161); BGU II 482, 9 (II AD), with BL I 438 for 9.

– 1.6.2) To indicate where some information is to be found: P.Laur. I 2, 10 (16 July, AD 148), with BL VIII 160; P.Oslo III 84, 1 (AD 138-161), with BL III 125 for 1; P.Petaus 127, 6, 15 and 27 (ca. AD 182-187).

– 1.6.3) To indicate misplaced entries: P.Lond. I 119 (p. 140), 2, 34 and 107-108 (p. 140; later than AD 143/144?), with BL I 219 for 2, BL I 221 for 107, and BL VII 81 for 108a; 109 B, 96 (p. 151; mid-second cent. AD). The entries on 9-10, 16-17, 28, 41, 51, 118 and 132-133 have no further remarks as to where they should be put. Therefore assigning those lines under this section is highly hypothetical¹⁵.

2) Private documents:

2.1) To indicate the addenda: P.Mil. Vogl. VII 308, 164 (AD 151-200).

Occurrences in P.Petaus and P.Lond I 119 show that the use of brackets is manifold even in the same official’s office or the same document. Therefore it is not possible to generalise as to the meaning of the brackets except to say that brackets are not signs of deletion.

The verso discussed also displays brackets. The document belongs to the temple administration. If it is allowed to transfer the use of brackets from the civil to the temple administration, I would suggest reading the brackets as signs indicating additional information, possibly for the purpose of checking.

Appendix 2: relation to the recto

The recto is a list of objects, which belonged to a temple in Oxyrhynchus. The verso helps to identify the temple, which is the Sebasteion/Kaisareion of Oxyrhynchus, a special place for the worship of the emperors.

Another interesting feature revealed by the verso is connected to the nature of the list. The verso is a list of the amounts of money given to the temple by private persons, four times through a banker. The dates in brackets provide the necessary information for checking the entries, a task done by another temple scribe. Independently of the argument given in the introduction, this is one reason to see the list on the verso as an internal document of the temple administration. Furthermore the simple names, without mentioning the patronym or the function of the persons involved, except for the bankers who transferred the money, may also indicate the internal use of the document. These two reasons confirm the ratiocination concerning the internal character of the list of objects on the recto (see the introduction). The list of cultic objects on the recto is the first Greek internal list documented so far¹⁶. All other Greek lists of such objects, as far as I know, record either inventory lists for official dealings with the state or offerings to a temple. An even more precise date for the recto can be given (see the introduction), i.e. the first half of the second century AD.