[HOME]

Arsinoe II

 

Arsinoe II1, daughter of Ptolemy I and Berenice I2, born c. 3163, married thrice.

She first married Lysimachus, king of Thrace and Macedon4, in 300/299 as at least his third marriage5, by whom she had three sons, Ptolemy (here identified with Ptolemy Nios), Lysimachus and Philip6. The marriage was terminated by the death of her husband at the battle of Corupedium in Feb 2817, after which she commanded the garrison at Cassandreia till winter 281/808.

She second married her half-brother Ptolemy Ceraunus, king of Macedon9, in early 280 at Cassandrea as probably his second marriage10. The marriage was terminated by his murder of her two younger sons shortly thereafter11, after which she fled to Samothrace12. By him she had no children.

She third married her full brother Ptolemy II, king of Egypt13, between 280 and 272, probably c. 273/214. She had no children by Ptolemy II but at some point after her death, probably in the late 260s, he had the children of Arsinoe I legally declared to be her children15.

Arsinoe II was incorporated in the dynastic cult with Ptolemy II in year 13 (Mac.) = 273/2 or 14 (Mac.) = 272/1 as the Sibling Gods, Qeoi Adelfoi16. She was victor in all three events for harnessed horses probably in the 127th Olympics, summer 27216.1. She died on the night of the new moon in Pachons year 15 (Eg.) = 25 July 270 or = 1/2 July 26817. She was honoured, probably from year 16 (Mac.) = 270/69 or year 17 (Mac.) = 269/8, by a priestess in the dynastic cult at Alexandria, the canephore ("basket-bearer")18.

Arsinoe II held titles as queen of Egypt, possibly posthumously19, as follows20:

Throne Name     xnm-jb-n-MAat mr-nTrw21
Nomen             'rsjnAt

[1] PP VI 14491. Gr: Arsinoh FiladelfoV. Ý

[2] Full sisterhood of Ptolemy II given in Pausanias 1.7.1; parentage of Ptolemy II in Pausanias 1.6.8. Ý

[3] Inferred from her marriage to Lysimachus shortly after the battle of Ipsus in c. 300 (Plutarch, Demetrius 31), making her at least a teenager, and the start of Ptolemy I's liaison with Berenice I in c. 317. Ý

[4] Pausanias 1.10.3. Ý

[5] Plutarch Demetrius 31, Justin 24.3. She is not explicitly named by Plutarch, but the chronology of her younger sons, aged 16 and 13 at their death in 281/0 according to Justin, forces a marriage in 298 or before. For the statement of Pausanias 1.10.3 that her half-sister Lysandra, who appears to have married Lysimachus' son Agathocles in c. 292, already had children by him at the time that Arsinoe II married Lysimachus, see discussion under Lysandra. Ý

[6] Justin 24.3. It is possible, though highly unlikely, that Arsinoe I was her daughter. Ý

[7] See discussion under Ptolemy Ceraunus. Ý

[8] Justin 24.2. Ý

[9] Justin 24.2, 24.3. Ý

[10] See Ptolemy Ceraunus. Ý

[11] Justin 24.3. Ý

[12] Justin 24.3. W. W. Tarn (JHS 46 (1921) 155) suggested that she went to Egypt from Samothrace after the final failure of her son Ptolemy's attempts to gain the Macedonian throne, and married Ptolemy II shortly thereafter. This makes sense to me, but is still essentially speculative. Ý

[13] Pausanias 1.7.1.

Diodorus 10.31.1, in listing famous incestuous marriages, includes Zeus and Hera and Ptolemy and Berenice. The Ptolemy involved is not further identified, but the only possible sibling pair of Ptolemy and Berenice was Ptolemy XI and Berenice III, although it is sometimes suggested that Ptolemy I and Berenice I were half-siblings. Surely Diodorus did not have either pair in mind, but rather the marriage of Ptolemy II and Arsinoe II, which scandalised the Hellenistic world. Ý

[14] Universally agreed termini set as (a) her marriage to Ptolemy Ceraunus at Cassandrea in 281/0 and (b) CCG 22183 (the Pithom Stele) line 15 records a visit to Heroopolis on 3 Thoth year 12 = 2 November 274 (if Egyptian year 12 is based on coregency accession) or = 1 November 272 (if Egyptian year 12 is based on true accession).

R. A. Hazzard, The Imagination of a Monarchy: Studies in Ptolemaic Propaganda 89f., proposes a tighter terminus post quem by accepting the official explanation for this first incestuous marriage, that Ptolemy II was following the example of Zeus and Hera (Theocritus Idyll 17.131, Plutarch, Moralia 736e), which he reasonably argues implies that the couple must have been deified before the marriage. The dynastic cult first introduces the couple as gods in year 14 (Mac.) = 272/1, hence Hazzard concludes that they must have been deified in year 13 or 14 (Mac.). In view of the Pithom stele, he concludes that the marriage occurred in year 13 (Mac.) = 273/2. This date is consistent with a mention of Arsinoe I in KAI 43, a Cyriote inscription dated to year 11 (Mac.), which is explicable if the marriage occurred after that date.

Hazzard's argument requires that the retrospective reference to year 12 (Eg.) in the Pithom stele must be based on accession-based regnal years, rather than a coregency-based regnal year. However, it is generally held that the retrospective references in the Pithom stele use coregency-based dates, and certainly the events recorded in year 16 seem to require that dating. If so, the marriage must have occurred before November 274, which apparently conflicts with Hazzard's argument. But there is also an indication that the deification may have occurred as early as 274.

There is one other way to resolve this difficulty whlie retaining Hazzard's argument. E. Grzybek, Du calendrier macédonien au calendrier ptolémaïque: problèmes du chronologie hellénistique 103ff., has argued that the Pithom stele (CCG 22183) shows Arsinoe II as alive in year 16, while the Mendes stele (CCG 22181) dates her death to year 15. Grzybek has explained this by proposing that the Mendes death date of year 15 is referenced to the year of Ptolemy II's accession, on the death of Ptolemy I, while the Pithom stele record for year 16 is referenced to the year of his coregency with Ptolemy I. But Hazzard's analysis of the date of the marriage, if completely correct, requires that the reference basis for the dates on the Pithom stele also changed from Ptolemy II's accession to his coregency. If it is agreed that the Pithom stele shows Arsinoe II as being alive in year 16, then on Hazzard's system the change must have occurred between the entries for year 12 and year 16, i.e. either Egyptian year 12 was directly followed by year 15 or year 13 was followed by year 16.

Neither proposal is consistent with Grzybek's analysis of the death date on the Mendes stele, which requires that there was an Egyptian year 15 and that it was calculated based on the true accession date of Ptolemy II. Also, other dated records exist for Egyptian years 13, 14 and 15 (listed in R. A. Hazzard, Phoenix 41 (1987) 140, 157). In isolation, the simplest way out of this difficulty is to suppose that the Pithom record for year 16 should not be interpreted as showing that Arsinoe II was alive in that year. In this case, we can assume that both the Pithom and Mendes stele were using accession-based dates for all entries before year 21. However, despite the assertion of M. Minas (discussed below) that the events encompassed under the year 16 entry could not have taken place within a single year, there seems no good reason to reject Grzybek's analysis on this point. The other way I see to rescue Hazzard's analysis is to argue that the scribes compiling the Pithom and Mendes stelae, in year 21, were using sources that did not use the same basis for dating the same events. The Theban and Elephantinean evidence discussed below is consistent with this position.

For the argument of A. M. Honeyman, JEA 26 (1940) 57, that Arsinoe II was married to Ptolemy in or before 278/7, see discussion under Arsinoe I. For the argument of E. Bresciani et al., EVO 26 (2003) 33, that the Satis graffito dates the marriage to 30 Mesore year 5 = 25 November 278, see discussion of the Satis graffito below. Ý

[15] Schol. Theocritus 17.128. G. H. Macurdy (Hellenistic Queens, 121) translates the relevant passage as "He married his sister Arsinoe and he had the children of the first Arsinoe legally called those of his sister, for the latter died without bearing him children."

The length of the period between her death and the adoption is unknown. However, its effect is clear: it removed all taint from the children of a traitor, in effect relegitimising them. One consequence was that it placed Ptolemy III back into the line of succession, which suggests that it could even have followed the rebellion of Ptolemy Nios. However, the data from the new Posidippus epigrams suggests that it happened earlier. Posidippus, Hippika AB 80, 82 indicate that Berenice Phernophorus won victories at the Nemean and Isthmian games as a child (i.e. between about 7 and 14), and that she was accompanied at the latter by her father. These victories require that she had legitimate status at the time, so the adoption must have occurred earlier. Since Arsinoe II and Ptolemy II were certainly married by 272, the very latest possible date is about 259.

If Ptolemy Nios was himself a son of Arsinoe I, as has been suggested, then the adoption must have occurred very shortly after the death of Arsinoe II, since it would have been a prerequisite of his elevation. However, this relationship is almost certainly ruled out by his omission from the list of Arsinoe I's children in Schol. Theocritus 17.128. Hence the adoption would have posed a political threat to him by elevating his "brothers" into the line of succession. This suggests that it was a factor in his rebellion, which in turn suggests a date in the late 260s.

A second circumstance pointing to the same general region is that Ptolemy II's mistress Bilistiche won victories in the Olympics of 268 and 264. She was clearly much loved by the king, but, if A. Cameron, GBRS 31 (1990) 287 has correctly interpreted AB 127, another epigram attributed to Posidippus, as a lampoon on her victory, she was not popular at court. One likely reason is that she was in a position to seek to raise any children of hers to the throne. Even if this was not her intention she was open to the charge -- particularly since Ptolemy II's own mother, Berenice I, may well have followed such a course. The adoption would have had the effect of shielding her from such charges, since it would have made it clear to all that Ptolemy II did not regard any children she may have had as being in the running.

R. A. Hazzard, Imagination of a Monarchy 44-46 etc., has argued that Ptolemy II conducted a major overhaul of his regime and his image around 263/2. If so -- and the thesis seems plausible if not all of Hazzard's arguments for it are equally so -- then perhaps the adoption was part of this overhaul. G. H. Macurdy, Hellenistic Queens, 40, suggested that Ptolemy II originally intended to marry Berenice Phernophorus to Ptolemy Nios. If so, then this also points to a date in the late 260s, when she would have started to become of marriageable age.

While none of these arguments are decisive, they suggest a date in the late 260s for the adoption. Ý

[16] First reference: pHibeh 2.199, year 14 (Mac.). From the titulary of the dynastic priesthood, we can be sure that they were first introduced into the cult in that year, with the nauarch Kallikrates of Samos as the first eponymous priest of the Qeoi Adelfoi.

Actual deification may have occurred up to two years earlier. Posidippus, Hippika AB 74, is an epigram celebrating the victory of Kallikrates in the quadriga for colts in the Pythian Games, and his dedication of a statue to the Qeoi Adelfoi. The original editors of the Posidippus epigrams argued that this victory must postdate the introduction of the deities into the dynastic cult, and hence dated it to the Pythian Games of 270. P. Bing, GBRS 43 (2002/3) 243 at 250f., argued rather than the victory was in 274 and that Kallikrates erected the statue in part in thanks for the honour of being the first eponymous priest in the cult. But CCG 22183 (the Pithom Stele) records that Arsinoe II and Ptolemy II were already married in Thoth of year 12 (Eg.), which is normally interpreted as a coregency-based date, i.e. November 274. If, as R. A. Hazzard, The Imagination of a Monarchy: Studies in Ptolemaic Propaganda 89f. reasonably argues, deification preceded marriage, then deification must have occurred in 274 or earlier, and Kallikrates could well have dedicated his statue in the year of his victory, as the poem seems to imply. Ý

[16.1] Posidippos, Hippika AB 78. The number of the Olympiad is not mentioned by Posidippos, but must be between her return to Egypt and her death. As noted by P. Bing, GBRS 43 (2002/3) 243 at 253 n. 23, only the 126th and 127th Olympiads are possible. The 127th is proposed here since only this one occurred after her deification and marriage. The three harnessed races as listed in pOxy 17.2082 (imaged here) were the quadriga and the pair for horses, and the quadriga for colts. Against this, L. Moretti, Olympionikai, 136 n. 542, lists the Athenian Glaukon son of Eteocles as victor in the quadriga for horses (Pausanias 6.16.9) in 272, but he admits the date is speculative, and only presents an argument that it must be before 268. Ý

[17] Mendes Stele (CCG 22181): Pachon year 15 (Eg.).

The debate on the date of Arsinoe II's death is the most involved in Ptolemaic chronology.

Callimachus, in his hymn to the deceased Arsinoe partly preserved in pBerol 13417A, has been held to state that her death coincided with a full moon. Assuming year 1 (Eg.) = 285/4 (coregency dating), the combination gives a date of 9 July 270. This analysis was first established by R. Pfeiffer, Kallimachosstudien 1ff and has generally been repeated since. However, it was challenged by E. Grzybek, Du calendrier macédonien au calendrier ptolémaïque: problèmes du chronologie hellénistique 103ff., who argued that the Pithom stele (CCG 22183) shows Arsinoe II being alive in year 16 (Eg.), to greet the return of an expedition which Ptolemy II has sent to tropical Africa. He interprets this to mean that the Pithom stele must be counting years from the association of Ptolemy II in coregency while the Mendes stele is counting them from his accession to sole rule, i.e. that Mendes really used a year 1 (Eg.) = 283/2 for its year 15 entry while Pithom used a year 1 (Eg.) = 285/4 for its year 16 entry. On this basis, she died between 26 June and 25 July 268.

As to the exact date, Grzybek notes that the Mendes stele specifies that the (monthly?) anniversary of her death be celebrated on the 10th of the month, and proposes that the four days of anointing described in the stele actually represents the 4-day ceremony of the opening of the mouth, meaning that she died on 6 or 7 Pachons = 1 or 2 July 268. Grzybek notes, however, that this is a new moon date. He reanalyses the scholia on and the language of pBerol 13417A to show that the poem should be understood to refer to a new moon, not a full moon. Finally, he looks for confirmation of this by correlating the date to the date of the Greek Arsinoea festival, celebrated on 6 Loios (= 27 Mesore in year 36 = 16 October 250) according to pCairZen 3.59312) and reconstructing the Macedonian calendar for these years accordingly.

H. Hauben, CdE 67 (1992) 143, further points out that this date results in Arsinoe dying very shortly before Ptolemy Nios becomes associated as coregent, in the year before the postulated change of the Egyptian year number from an accession base to a coregency base, and also very shortly before the usual dating of the Decree of Chremonides (SIG3 I 434, M. M. Austin, The Hellenistic World from Alexander to the Roman Conquest 94ff. (49)). This decree, dated 9 Metageitnion (Ath.) in the archonship of Peithidemos, states that Ptolemy II, in supporting the alliance announced in the decree, was following the policy of his ancestor (i.e. Ptolemy I) and his sister (i.e. Arsinoe II) in supporting the freedom of Athens from Macedonian rule. Since the archonship is Peithodemos is usually dated to 268/7, a date of 268 for the death of Arsinoe II allows her mention in the Decree to be interpreted as indicating her active involvement in building the anti-Macedonian alliance. Although not conclusive, this all seems rather satisfactory, and the set of coincidences listed by Hauben looks like a good circumstantial argument in favour of Grzybek's position.

However, the date of the archonship of Peithidemos is itself not completely certain (H. Hauben, CdE 67 (1992) 143 at 162). Although 268/7 is the date most strongly favoured by modern scholarship, other dates have been proposed at various times, ranging from 270/69 to 265/4. Most recently, S. Byrne, MeditArch 19 (2006) [in print], notes that the date of the decree gives the equation 9 Metageitnion = 9 Prytany II, would ordinarily be taken as proof that this year was an ordinary year. Assuming that the Athenian calendar was regulated by the Metonic cycle at this time. 268/7 and 265/4 would both be intercalary years, so Byrne argues that the correct date is 269/8.

The Decree's statement that Ptolemy II was following the policy of his sister implies that she was no longer instrumental in setting policy at the time of the decree. But this need not mean that she was dead. Whatever the date of Peithidemos, the mention of Ptolemy I shows that the mention of Arsinoe II merely means that she had been involved in forming an anti-Antigonid alliance with Athens at another time. After her brief marriage to Ptolemy Ceraunus, she had spent an unknown amount of time on Samothrace and her son Ptolemy had attempted to gain the throne of Macedon, which was ultimately won by Antigonus II. It is perfectly possible, for example, that the Decree of Chremonides is referring to her opposition of Antigonus II during her time on Samothrace. Thus, unfortunately, the correct date of the Decree of Chremonides is essentially irrelevant to determining her death date.

Grzybek's analysis of the year of her death has been widely, almost generally, rejected by Egyptologists (though classicists seem only dimly aware of the debate). But it seems to me that none of the arguments made against him are solid, let alone conclusive, and that some, if examined closely, actually tend to favour Grzybek's view.

H. Cadell (in H. Melaerts (ed.) Le culte du souverain dans l'Égypte ptolémaïque au IIIe siècle avant notre ère 1) notes:

(a) that the Mendes and Pithom stelae were both erected in two major temples that were close to each other at roughly the same time, in or shortly after year 21, and both include references to the opening of a major new canal: they therefore should be based on the same calendrical system.

But the point is that the dates in question, in both stelae, are retrospective. It is generally agreed (though not universally -- see below) that the Egyptian regnal year was moved from an accession-based count to a coregency-based count at the end of year 16, so that year 16 was followed by year 19. Hence retrospective dates before year 19 might be adjusted to the coregency-based count, or not, without affecting current dates. The fact that the two cities used the same system in year 21 does not necessarily imply that all scribes in both cities chose to apply it retroactively.  

(b) that the events associated by Grzybek with year 16 in the Pithom stele are difficult to encompass in a single year, and the date should therefore not be understood as covering all of them.

This is a matter of opinion. More to the point, even if granted, this objection only has validity if it can be shown either that the narrative does not proceed in the same temporal direction, or that the events cover so much time that they imply Arsinoe was living even on Grzybek's chronology. I.e., even if the objection is granted for the sake of argument, Grzybek's point would still be valid if the expedition returned in year 17, though not if it returned in year 20.  It would be necessary to show that the logic of the inscription allows the return to be dated before year 16 or after year 18. Grzybek's analysis of the data (E. Grzybek, Du calendrier macédonien au calendrier ptolémaïque: problèmes du chronologie hellénistique 72f.) does not appear to show any discontinuities where the narrative goes backwards in time, or any particular reason to extend the events over more than 2 years.

(c) that the Arsinoea, being celebrated on 27 Mesore, is not likely to be understood by the Egyptians as a feast of the death day of Arsinoe II celebrated on the 10 of the month.

This is fair enough as far as it goes, but it doesn't go very far. It only proves that the Egyptian feast of the death day of Arsinoe II was not the Arsinoea. It does not prove that the 10th of the Egyptian month was not the day of the Egyptian festival of her death. 

(d) that pSorb 2440, dated to Audnaios of year 18 (Mac.) = 268/7 names a canephore, Berenice daughter of Andromachus, showing that this office must have been established at or before the start (=c. 25 Dystros) of that year, i.e. before c. Feb/March, meaning that Arsinoe II was already dead by then.

But Audnaios is only two months before Dystros, i.e. almost at the end of year 18, and several months after Arsinoe's imputed death date. Cadell asserts that such an office would only be instituted at the start of a Macedonian regnal year, but provides no reason to believe this. Even assuming -- which is not proved -- that the canephorate was in fact established as a posthumous priesthood, it seems to me perfectly possible that the first canephore was only named for the remainder of the year of Arsinoe's death.

(e) that p dem Bryce = pEhev. 12, whose date is not preserved but must be early, names an unplaced canephore Eukleia daughter of Aristodikos, who may well be placed therefore in year 17 (Mac.) = 269/8, i.e. before the date calculated by Grzybek.

Unfortunately, this is the only part of the dating formula that survives. E. Lüddeckens, Ägyptische Eheverträge 27, restores "Ptolemy son of Ptolemy", which would ensure that the papyrus dates before year 26, and the available space would appear to eliminate later Ptolemies or the period of coregency with Ptolemy Nios, but it is not clear that the form "Ptolemy son of Ptolemy Soter" can be definitely excluded.

While in theory a canephore may have been replaced partway through the year, e.g. due to death, we should prefer to assign the papyrus to a year without a canephore that is otherwise known, i.e. to years 16, 17, 32 or 37. In favour of an earlier year, the canephore of year 20, Berenice, was also a daughter of Aristodokos. Hence, the balance of probabilities is that this papyrus does indeed show that the canephorate existed before the date that Grzybek asserts for Arsinoe II's death, but the case is far from certain, and later dates cannot yet be excluded.

However, as noted above, and by L. Koenen in A. Bulloch et al. (eds) Images and Ideologies 25 at 56 even before the announcement of pSorb 2440, we do not actually know that the position of canephore was created for a posthumous cult. It could well have actually been instituted towards the end of the lifetime of Arsinoe II to reflect her divinity. While we know of eponymous priests but no canephores for 273/2, 272/1 and 271/0 (years 13-15 (Mac.)), the changes to the royal cult were in their infancy at this point. Certainly, Cleopatra III did not see any need to wait for her death before her priestesses were added to the dynastic cult.

The cult of Arsinoe had been preceded by a cult for her sister Philotera that was certainly posthumous, and that was not just Greek but also Egyptian. The creation of the cult of Arsinoe was shortly followed by a major reform of the Egyptian clergy, which resulted in Nesisti-Pedubast, the High Priest of Memphis (an office that may even have been resurrected for the purpose), becoming the head of a cult that was funded by significant tax revenues allocated by the state. It hardly seems likely that such a major reform was occasioned by her death, nor that it was delayed until she died. Rather, it seems quite possible that it was already well in train when she died, and her death merely added impetus to the nascent cult. That impetus may nevertheless have been very significant. On Grzybek's analysis of the date, as given above, it would appear possible that elements of Egyptian funerary rites were incorporated into hers, although we know from a fragment of Callimachus (pBerol 13417A) that Arsinoe's body was eventually cremated.

In other words, what Cadell has shown is that, if Arsinoe II died in 270, then the canephorate was a posthumous institution, but if she died in 268 then it very probably was not. However, unless we can show independently that the canephorate was a posthumous cult, the date of its establishment cannot legitimately be used to argue the date of her death.

A related objection had been raised by G. Hölbl, Tyche 7 (1992) 117 at 120. He noted that line 21 of the Pithom stele describes the foundation of a temple of Philadelphos in, for Grzybek, year 16, which was dedicated by the priests of Atum. According to Hölbl, this would imply that Arsinoe II had been individually deified in an Egyptian cult before her death, which raises the question of why it was necessary, in the Mendes stele, to propagate her cult through all the temples of Egypt.

L. Criscuolo, Aegyptus 71 (1991) 282 at 286, also makes points (a) and (b), and amplifies point (b) by noting that according to Grzybek's analysis, together with a correct interpretation of the text, the royal couple should have attended the installation of an Apis bull, a Mnevis bull and another unidentified sacred animal in the same year 16, and not merely, as Grzybek states, paid court to these animals. However, the records for Apis bulls under Ptolemy II cited by Criscuolo (D. J. Thompson, Memphis under the Ptolemies, 284 ff.) show an Apis year 22 for Apis of Ta-net-Merwer in Ptolemy II year 6 and Apis year 3 for Apis of Ta-Renenut I in year Ptolemy II year 32 (recte 33 -- see H. Brugsch, ZÄS 22 (1884) 110 No. 1), with apparently one bull -- Apis of Wadjet-iyti -- in between. An installation in year 16 would imply that the previous bull lived at most 10 years, which is very improbable. If this event is therefore not really connected to year 16, neither can any of the other events listed, and Grzybek has no basis for arguing that the Pithom stele shows Arsinoe II alive in that year.

M. Minas, in Fs Winter 203, also makes points (a), (b) and (c), and rejects the linkage between the festival date of 10 Pachons and the death date of Arsinoe proposed by Grzybek as arbitrary. She further notes that the standard analysis of Ptolemy II's Egyptian regnal years shows a transition from accession based counting to coregency based counting at the end of (accession-based) year 16, which was therefore followed immediately by (coregency-based) year 19, but that Grzybek's interpretation requires that the year 16 date on the Pithom stele be a coregency-based count.

Minas nevertheless accepts Grzybek's reanalysis of pBerol 13417A as implying Arsinoe II's death on the new moon, while rejecting the logic that led Grzybek to perform this reanalysis. This leaves us in the somewhat unpleasant position of accepting that Grzybek arrived at good conclusions for bad reasons, but is not necessarily wrong. Accordingly she redates the death of Arsinoe II to the new moon of 30 Pachon = 25 July 270, the date accepted here for the coregency-based dating.

E. Kosmetatou, AfP 50 (2004) 18 at 34, notes that on Grzybek's chronology Arsinoe II died very shortly before the celebration of the Olympic Games of 268, at which Ptolemy's mistress Bilistiche won the prize for the quadriga for colts. She argues that no mistress would have been granted this royal privilege while Arsinoe II still lived, and that there was not sufficient time for Bilistiche to have developed a team; this is much more believable if she died in 270.

A last argument on the topic, that does not seem to have been noticed outside numismatic circles, concerns the so-called "Arsinoe era". A sequence of octadrachms, tetradrachms and decadrachms minted at Alexandria and bearing the portrait of the deified Arsinoe II carry control marks from A to W, AA to WW, AAA and BBB, together the earliest coins in the series which are unmarked. J. N. Svoronos, Die Münzen der Ptolemäer, nos 408-519 and 937-961 interpreted these control marks as corresponding to dates in an era beginning with the death of Arsinoe II. On this system, assuming A to mark 270/69, the year following Arsinoe II's death in 270 (commemorated by the unmarked coins), W marks 247/6, the last year of Ptolemy II. It is also the last year of a series of bronze coins with similar markings. BBB, the end of the decadrachm series, also marks the year 221/0, then held to be the last year of Ptolemy III. These coincidences appeared to prove the correctness of the interpretation, and hence of the assumption that the base year was 271/0. That is, the Arsinoe era proves that Arsinoe II died in 271/0.

This analysis was refuted by H. Troxell, ANSMN 28 (1983) 35, building on an observation by H. B. Mattingly that there was some sharing of dies between the octadrachms and the tetradrachms, but not with the decadrachms. Troxell first did a stylistic analysis of the dies. She was able to establish four groupings and to place coins of each type in the groupings. She found that the decadrachm control marks were considerably in advance of the octadrachm/tetradrachm sequences, i.e. decadrachms A-C formed one group while octadrachm/tetradrachms A-Q shared die styles with decadrachms O-BB, and so on. She also compared these coins with similar coins from Ptolemaic Phoenician mints bearing letters that were certainly regnal year dates, and showed that the die styles of these coins corresponded to the Alexandrian coins in such a way that the control marks would have to increment at a faster than annual rate. These two observations are sufficient to show that these control marks cannot be used as an annual count. It further follows that they cannot be used to substantiate 271/0 (or any other year) as the date of death of Arsinoe II.


Central to the discussion is the nature and date of the change in Ptolemy II's Egyptian regnal year. The standard analysis of this (see A. E. Samuel, Ptolemaic Chronology 26ff.) runs as follows: The highest Egyptian date for Ptolemy I is year 21 = 285/4 (p dem BM 10525 = Epeiph year 21 = August/September 284). However, stele iBucheum 3 states that a Buchis bull born on (Egyptian) year 14 Payni 19 of Ptolemy I died on Mecheir 25 year 13 of Ptolemy II aged 20 years, 8 months and 13 days, implying 22 completed years for Ptolemy I. This shows that in year 13 (Eg.) of Ptolemy II his reign was dated from 283/282, i.e. that he dated from the death of Ptolemy I. Further, pEleph 5 is dated to year 2 (Eg.) of Ptolemy II, while pEleph 3 and pEleph 4 are both dated to year 41 (Mac.) of Ptolemy I, and all three papyri are in Greek and concern members of the Macedonian community. Hence year 2 must refer to an accession-based regnal year. Thus, Ptolemy II certainly began his reign with an accession-based count that appears to have lasted at least up to year 13.

But double-dated documents, or documents naming eponymous priests for known Macedonian years, dating after (Egyptian) year 21 of Ptolemy II, such as pdem Phil 510b dated to year 22 Loios 16 (Mac.) = year 21 Epeiph 12 (Eg.), show that by this time the Macedonian year number was sometimes one in advance of the Egyptian year number, meaning that the Egyptian year count was based on the coregency date by this time. Hence iBucheum 3 was held to show that the Egyptian dating system only changed from an accession basis to a coregency basis after year 13 (accession-based) = year 15 (coregency-based) and before year 21. Pestman further argued that the transition happened during or after year 16, which was therefore followed by year 19 (see P. W. Pestman, Chronologie Ëgyptienne d'après les textes démotiques 22), claiming that the only data for years 17 and 18 were demotic ostraca recording receipts for NHb and NHT taxes from Thebes and Elephantine that should be reassigned to Ptolemy III. E. Grzybek, Du calendrier macédonien au calendrier ptolémaïque: problèmes du chronologie hellénistique 118ff. similarly proposed that they should be reassigned to Ptolemy I, while L. Koenen, Eine agonistiche Inschrift aus Ägypten und frühptolemäische Königsfeste, 43, suggested that the ostraca dated to years 17 and 18 were simply errors.

However, B. Muhs (Fs. Pestman 71), analysing the prosopographical data for individuals named in these ostraca, has persuasively argued that they best fit the reign of Ptolemy II after all, since this match consistently gives the most reasonable lengths for the active careers of individuals who can be traced over a number of years. According to Muhs, it would follow that, at least in Thebes, the Egyptians recognised years 17 and 18 of Ptolemy II. Since this gives us a complete sequence of Egyptian years from 13 to 21, and we already have a complete accounting of years 1-13, we are left without an identifiable transition point between the two systems for counting Egyptian regnal years.

Muhs proposes to reconcile the difficulty by questioning the accuracy of the standard interpretation of iBucheum 3. He proposed that the unorthodox writing of the age of the Buchis bull (as three lines: 10+1/5/4) indicates that the scribe had a problem with the calculation, meaning that this datum can be discarded as a proof of an accession-based Egytian year count. (It seems to me that uncertainty about the basis for the regnal year could explain the scribe's confusion.) In Muhs' view, since the taxation ostraca complete a record of Egyptian year dates for Ptolemy II, we must conclude that the Egyptians always dated the reign of Ptolemy II from the start of the coregency.

This reconstruction would destroy the theoretical basis of Grzybek's argument, leaving us to with the challenge of finding another explanation for the discrepancy which led to it. Moreover, the fact that the highest Egyptian date we have for Ptolemy I is in year 21 leaves room for the theoretical possibility that Ptolemy II's Egyptian years were accounted from the same year.

However, it seems to me unlikely that Muhs is right.

The existence of pEleph 5, dated to Egyptian year 2 of Ptolemy II and not discussed by Muhs, must show an accession-based dating system at that time, otherwise we are forced to suppose that some Greeks dated according to Ptolemy I (pEleph 3 and pEleph 4, dated to year 41 = 284/3 = coregency-based year 2 of Ptolemy II) while others dated according to Ptolemy II at the same time and in the same place.

Additionally, we have a series of related demotic papyri which cover the transition from Ptolemy I to Ptolemy II:

While the procedures are not entirely clear, it seems that a purchase tax ("tithe") of 2.5 kite was due on completion of the sale of a house, with property taxes ("encyclion" in Granville's translation) equivalent to 2 kite (=1 stater) per annum thereafter, but that these could be paid cumulatively triennially. B. P. Muhs, Tax Receipts, Taxpayers, and Taxes in Early Ptolemaic Thebes 66ff, regards both as purchase taxes: the 2.5 kite tax as a purchase tax paid to the state, since the receiving officials hold positions in the Theban bureacracy and the army, while the 10% encyclion is a purchase tax paid to the Temple, since the receiving officials hold positions in the temple and phyle hierarchy. (S. P. Vleeming in J. H. Johson (ed.), Life in a Multicultural Society 343, notes that pdem BM 10529, 10530 and 10535 are the only known examples of this type of payment.)

While I agree that the payments were made to different authorities, I do not agree that they are both purchase payments. First, the 10% encyclions were paid at least a year after the house purchase. Second, the theory does not explain why Teinti pays these taxes at least twice on the same houses. Rather, the 2.5 kite must be seen as a purchase tax paid to the state and the encyclion as a property tax, or leasing fee, the amount of which was based on the purchase price.

If this is correct then the interval between pdem BM 10537 and pdem BM 10530 should be 3 years, not 1, favouring an accession-based date for pdem BM 10530.

It is notable that the data supporting Muhs' analysis all comes from a specific class of taxation ostraca, the NHb and NHT taxes. The Theban ostraca are listed in great detail in his formal publication, B. P. Muhs, Tax Receipts, Taxpayers, and Taxes in Early Ptolemaic Thebes 33ff. While the years before year 7 are only sparsely documented, three of these receipts do appear to show the use of a year number of Ptolemy II while he was still coregent:

Additionally, odem Louvre 85 is a receipt for NHb taxes of year 3 dated 12 Choiak year 4, and odem PLBat 26.4 is a receipt for NHb taxes of year 3 dated 11 Tybi year 4.

However, these dates do not necessarily represent a regnal year number. The taxes in question were apparently state capitation taxes administered through tax farmers, a Ptolemaic innovation. Later on, such receipts were dated according to an Egyptian financial year based on 1 Mecheir, which is generally supposed to derive from the alignment of the Macedonian regnal year at the time of a major tax reform, c. year 21. This suggests that the year number in these tax receipts is also a financial year -- in effect, that the Egyptian financial year was an evolution of an earlier system in which tax receipts given by tax farmers were based on Ptolemy II's accession to the coregency and on the Macedonian regnal year.

In support of this notion one might note two NHb receipts dated to years 30 and 33 (odem Louvre 1424 and 87). Muhs (Tax Receipts, Taxpayers, and Taxes in Early Ptolemaic Thebes 31 n. 168) is unable to explain these dates, though he accepts the readings as correct. On the theory that they are dates for a tax year derived from the Macedonian regnal year, these receipts could easily date from the reign of Ptolemy I.

On this hypothesis, the NHb and NHT taxation ostraca need not have any bearing on the regulation of the Egyptian civil year of Ptolemy II.

Nevertheless, in addition to Apis stele IM 4177, whose assignment to year 5 of Ptolemy II is highly doubtful, I have found three pieces of evidence that arguably support Muhs' thesis.

In summary, in favour of a shift from accession-based to coregency-based Egyptian years, we have, as noted, iBucheum 3, pEleph 5, and the pdem BM documents from year 21 of Ptolemy I to year 9 of Ptolemy II. On the other hand, we may have records for all of Ptolemy II's years, even though some of it may be using a financial year system.

How to explain this data?

In light of the uncertainty of the king named in the Medinet Habu graffiti, the only evidence against the standard model, that an accession-based year 16 was followed by a coregency-based year 19, remains the assignment of tax ostraca to Ptolemy II, discussed above. Since these may well reflect a financial year rather than a regnal year, it is very possible that the standard model is still in fact correct. However, there is another possible solution: that the calendrical transition for the Egyptian year count was not rigorously or uniformly applied.

This idea was suggested in passing by S. R. K. Glanville, Catalogue of Demotic Papyri in the British Museum I xix. A. E. Samuel, Ptolemaic Chronology 27 n. 56 dismisses it, saying that "once the order were issued, there would only be the interval required for the news to get throughout the country before the new system were followed everywhere". But this makes the assumption that an order was actually issued. At the time Samuel wrote this, it was naturally assumed that the Egyptian and Macedonian regnal year bases changed at the same time. But it now seems that the Macedonian year count was changed well over a decade before the earliest traceable date for a change in the Egyptian year count. That is, no order was issued to change the Egyptian year count when the Macedonian year count was changed. Granted that there is abundant evidence of a major administrative reform taking place at about year 16/19, during the interim there are no double-dated documents or any other evidence to suggests that any steps were taken to control the Egyptian regnal year: it was apparently not relevant to the Alexandrian government.

A non-uniform change from an accession-based year count to a coregency-based year count, would allow us to reconcile all this data. If correct, the unusual numeration of iBucheum 3 might be an attempt to have it both ways on the age of the bull (i.e. its 8 or 10 years depending on which system you mean for the death date). Most significantly, we would have no reason to object to a coregency-based year 16 reference occurring on the Pithom stele.

Whatever the solution to the problem of Ptolemy II's Egyptian regnal years, I think there is no good reason to doubt that there was a change from an accession-based system to a coregency-based system. However, the question of whether the coregency-based system was universally applied retroactively is separate and distinct.

 
Even though Grzybek's view has not been widely accepted, my own view r