THE ACCESSION OF ARTAXERXES I

JULIA NEUFFER
Washington, D.C.

Introduction
This article is principally a reexamination of the source
data relevant to the accession date of the Persian king
Artaxerxes I, and especially a study of a double-dated
papyrus from Egypt that was, until a few years ago, the only
known ancient document assigning an approximate date to
that event.
The "first year" and, therefore, the other years of his
reign have long been known in two calendars. According to
Ptolemy's Canon, which is fixed by eclipses, and according
to certain double-dated papyri from Egypt (to be discussed
below), his year I in the Egyptian calendar was the 365-day
year beginning on Thoth I, the Egyptian New Year's Day
(that is, December 17), 465 B.C. In the Persian reckoning
(in the Babylonian calendar, which was adopted by the
Persian kings), his first year was the lunar year beginning
in the spring, with Nisanu (Jewish Nisan) I, approximately
April 13, 464, several months later than the Egyptian year.
Postdating and Antedating. This Persian reckoning means
that his reign must have begun before Nisan I, 464, because
the Babylonian-Persian method was to postdate all reigns.
That is, when a new king succeeded to the throne the scribes,
who had been dating all kinds of documents by the day and
month "in the z ~ s [ort whatever] year of King X," would
begin using the new dateline "in the accession year [literally,
The equivalents of Persian dates in this article are taken from
the reconstructed calendar tables in Richard A. Parker and Waldo
H. Dubberstein, Babylonian Chronology, 626 B.C.-A.D. 75 (Providence,
R. I., 1956), hereinafter abbreviated: PDBC (1956).
 ACCESSION OF ARTAXERXES I 61

"beginning of the reignJJ]of King Y," and would wait until

the next New Year's Day to begin dating "in the year I of
King Y."
Does Ptolemy 's Canon, then, similarly indicate that Arta-
xerxes came to the throne before December 17, the Egyptian
New Year of his year I ? No. Detailed checking of the source
data has shown that the Canon uses two methods. In its earlier
portion, which lists Babylonian and early Persian kings, it uses
the postdating method (called by some the "accession- year
method"). But in its latter portion, which lists the Seleucids
and the Roman emperors, it antedates the reigns. That is, it
counts as "year I" the year in which a king came to the
throne, as if he had been reigning since the first day of the
year. By this method, commonly used in Egypt, a scribe
would begin dating in the king's "year I" as soon as he came
to the throne, and the first New Year's Day would begin
"year 2.')
If Ptolemy's Canon dated Artaxerxes in this way, it would
indicate that he came to the throne after December 17, 465.
Since the Canon used both methods, and source data for
the later Persian kings are insufficient, the Canon does not
help in determining whether Artaxerxes came to the throne
before or after Thoth I.
Ancient Documents. Thousands of ancient documents from
the period of the Persian Empire written on clay tablets-
letters, deeds, contracts, business accounts-have been found,
mostly in Babylonia. Many of them carry datelines in the
day, month, and year of the king. Thus it is often possible

a On postdating and antedating, see Edwin R. Thiele, The Mysteri-

ous Numbers of the Hebrew Kings (rev. ed.; Grand Rapids, Mich.,
19651, P. 17.
3 Because of this uncertainty, a conclusion had to be held in
abeyance in the present author's "A Study of Ptolemy's Treatment
of the Babylonian and Persian Regnal Years" (unpublished Master's
thesis, S.D.A. Theological Seminary, Andrews University, 1947) as
to the method used in Ptolemy's canon t o number the regnal years
of Artaxerxes I.
62 JULIA NEUFFER

to determine, from the month and day of the last tablet in

one reign and the first dated in the next, the approximate
date of the accession.
But because no contemporary tablets have been found
dated in Artaxerxes' accession year or in his father's last
year, his accession could formerly be dated only approximately
by the only known contemporary dated document, a papyrus
from Egypt.
A Double-dated Papyrus. This was one among a number
of Jewish papyri written in Aramaic found on the Nile island
of Elephantine a t Syene (modern Aswan). Jewish soldiers in
the Persian army in Egypt lived here in a garrison town with
their families, spoke Aramaic, and had their own temple.
They dated by their lunar calendar; but on documents they
used double dates, in their own lunar calendar and in the
Egyptian solar calendar. Many of these papyri can be dated
exactly in our calendar because a month and day in the
shifting lunar calendar can synchronize with the Egyptian
month and day in the fixed 365-day calendar only once in
twenty-five years. Thus the papyrus A P 6, with a double
date, can be fixed to January 213, 464. Its double-year date,
in the year 21 of Xerxes and the accession year of Artaxerxes,
has been interpreted to indicate that Xerxes had died and
Artaxerxes had succeeded him very recently, probably in
December, 465.
A Tablet Formerly Used. A tablet from Ur, published in
1949 (designated UET IV, 193) apparently indicated that
Xerxes was still living in late December. Written in the 13th
year of Artaxerxes, it was, as described by its editor, a

See the list of these in PDBC (1956), pp. 11-24.

5 A. E. Cowley, Aramaic Papyri of the Fzfth Century B.C.(Oxford,
1923)~No. 6, pp. 15-18. This papyrus, hereinafter designated as
A P 6, was Papyrus B in its initial publication by Sayce and Cowley
in 1906.
6 So Parker and Dubberstein in their first edition (Chicago, 1942)~
p. 15. This edition (626 B.c.-A.D. 45), is hereinafter referred to as
PDBC (1942).
 ACCESSION OF ARTAXERXES I 63

"rearrangement of land parcels," mentioning the "first

arrangement: Kislimu [the month of Kislev], 21st year of
Xerxes." According to this, Xerxes died after the 1st of
Kislimu, which began about December 17, 465 (thus, inci-
dentally, coinciding almost exactly with the Egyptian month
of Thoth in that year). This seemed a t first to settle the
question, but not for long.
A Hellenistic Tablet With an Exact Date. By the time
Parker and Dubberstein brought out the 1956 edition of
their Babylonian Chronology, another clay tablet from
Babylonia had come to light, an unpublished astronomical
text of the Hellenistic period (designated LBART No.
*141g) mentioning the murder of Xerxes in the month of
Abu (Jewish Ab), on the q t h ? (or any day from the 14th
to the 18th; the number is broken). If the writer of this
tablet, about 150 years (or more ?) after the event, had correct
information, Xerxes died approximately August 4-8, 465.
In the absence of any contemporary evidence, this has
been accepted by Parker and Dubberstein in their 1956
edition, and by others. Figulla, the editor of the above-
mentioned Ur tablet, in which he had read "Kislimu, in the
year 21 of Xerxes," decided that the partly broken word
which he had taken as "Kislimu" must have been something
else if Xerxes was dead some months earlier. Actually, the
original may have read "Kislimu," but since no one knows
what the entire word was, this text is eliminated as evidence.

7 H. H. Figulla, ed., Ur Excavations : Texts, IV (London, 194g),

NO. 193, P. 15.
8 Late Babylonian Astronomical and Related Texts, A. J . Sachs, ed.
(Providence, R. I., 1955), No. * 1419. This tablet, hereinafter desig-
nated LBART No. * 1419, is merely described briefly, in this volume
of Hellenistic texts, as listing certain eclipse dates; for the incidental
mention of a date for the death of Xerxes (not mentioned in LBART),
see PDBC (1956), p. 17, citing Sachs. Since this tablet was described
in a book issued twelve years ago but still remains unpublished, there
is no point in awaiting its publication in order to use it a t least tenta-
tively, though it can hardly be evaluated since details of its contents,
date, provenience, and general accuracy are not yet available.
64 JULIA NEUFFER

Thus we are left with two dated documents: (I) the con-
temporary papyrus AP 6, which has been taken to indicate
that the accession was still recent in January; (2) the Hellenis-
tic tablet LBART No. *141g, which dates the death of
Xerxes five months earlier. Can they be reconciled? An
examination of the papyrus and of the historical accounts
relating or mentioning the death of Xerxes furnishes clues
to a harmonious interpret ation. This study, comprising two
main parts, will examine first the historical and chronological
records, then papyrus A P 6.

Ancient Historical Accounts

The Oldest Historical Account. Even earlier than the
Hellenistic tablet that dates Xerxes' death is a historical
narrative of his murder, produced by Ctesias, a Greek physi-
cian at the court of Artaxerxes I1 (grandson of Artaxerxes I),
about 65 years after Xerxes' death. Ctesias lived in Persia,
knew the language, and had access to the official archives
and to the accounts preserved by the royal family. His
Persica is extant only in a summary by Photius (9th century
A.D.).
Ctesias tells the story as follows: Artabanus, a very powerful
courtier, with the aid of an influential palace chamberlain,
assassinated Xerxes, then procured the death of Darius, the
older son and heir, by accusing him to Artaxerxes, the younger
son. Thus Artaxerxes reigned with the support of Artabanus.
But later the powerful Artabanus decided to put his young
protCgC out of the way and take the throne. He made the
mistake of enlisting the help of Megabyzus, a brother-in-law
of Artaxerxes. When Megabyzus told the king everything-
the plot against him, the murder of Xerxes, and the false
accusation against Darius-Artaxerxes asserted himself, and
Artabanus was put to death. There followed a battle with
the partisans of Artabanus in which three of his sons were
killed. Then the Bactrians revolted under their satrap,
 ACCESSION OF ARTAXERXES I 65

another Artabanus, but after two battles they submitted. 9

Later Ancient Writers. Others (here cited in chronological
order) mention Xerxes' murder, and several tell essentially
the same story as Ctesias, with some differences, mostly
on minor points.
Aristotle (4th century B.c.) makes a casual allusion-by
way of illustration, not as historical narrative-to the murder
of Xerxes by "Artapanes, " who feared punishment for
having hanged Darius.
What may or may not be the next historical statement is
the one found on the above-mentioned tablet (LBART
No. *141g) from the Hellenistic period-late 4th century
or possibly even later-which says that Xerxes was killed
on Abu 14 (-18 ?), approximately August 4-8, 465. l1 Unfor-
tunately, the date and the text of this tablet are not available
since it remains unpublished.
Manetho, an Egyptian priest (3d century B.c.) whose
history of Egypt, in Greek, is now lost, included Artabanus
among the Persian rulers of Egypt, giving him a seven-month
reign. At least he did so if the Epitome of his history, compiled
soon afterward in the form of king lists, reflects accurately his
historical account. l2
Diodorus of Sicily (late 1st century B.c.) tells the story of
the murder of Xerxes by Artabanus, captain of the king's
9 Ctesias, Persica (Summary by Photius), 29-3 I (Brussels, I 947,
pp. 33-35). A year or two later came a revolt in Egypt, led by Inarus,
in which the Athenians aided the Egyptians, and which lasted about
five years (Ctesias, op. cit., 32-36).
, (Loeb ed., pp. 448, 449).
lo Aristotle, Politics, v. 8. 14; 131~ b 38
11 A. J. Sachs, cited in PDBC (1956). p. 17. This text is listed a-
mong the historical sources, not because i t presents an account of the
event, but because it is not a contemporary dated document but a
statement made by a writer a century and a half afterward, if not
later.
12 Manetho, Aegyptiaca (Epitome), Fragment 70, from Africanus,
as preserved by Syncellus (Loeb ed., pp. 174, 175). In footnote I
(see also facsimile on P1. 111),reference is made to a papyrus fragment
of this Epitome, a copy from the 5th century A.D., independent of
Africanus; this also lists [Artalbanus between Xerxes and Artaxerxes.
66 JULIA NEUFFER

bodyguard, who then offered Artaxerxes the help of the

guard in punishing Darius whom he accused of having
committed the crime. When "he saw his plan was prospering"
he decided that the time had come to kill Artaxerxes also.
Calling his sons together, he attacked and slightly wounded
Artaxerxes, whereupon the latter dealt him a fatal blow,
and then "took over the kingship." Diodorus places the death
of Xerxes, after a reign of more than 20 years, in the Athenian
year of the archonship of Lysitheus (which ran from mid-
summer 465 to midsummer 464) and in the Roman year
(January-December, 465) of the consulship of Lucius Valerius
Publicola and Titus Aemilius Mamercus; that is, in the
second half of 465. Apparently it was two years later (46312)
that Artaxerxes, "who had just recovered the throne, first
of all punished those who had a part in the murder of his
father and then organized the affairs of the kingdom to suit
his own personal advantage." l3
Trogus Pompeius, sometimes called Gnaeus Pompeius
Trogus (1st century B.C. to 1st century A.D.),gives a similar
account, as transmitted in extracts by Justin (3d century
A.D.).This narrative says that Artabanus, fearing a struggle
for the throne among the nobles, plotted to seize the throne
himself. Upon learning of this treachery Artaxerxes, being
only a boy, feared Artabanus and his seven sons. He therefore
ordered out the troops for review. As Artabanus presented
himself the young king asked the commander to exchange
corselets with him, since his own was too short. While Arta-
banus was thus unarmed, Artaxerxes ran him through with
a sword and ordered the arrest of the sons. l4
Two other Greek historians mention Artabanus. Nepos
Diodorus Siculus, xi.6g.1-6; xi.71.I (Loeb ed., IV, 304-307,
308-309). He places the final settlement in the archonship of Tlepo-
lemus and the consulship of Titus Quinctius and Quintus Servilius
Structus.
14 Justinus Frontinus, History of the World, Extracted from Trogus
Pompeius, xiii.1, in John Selby Watson, trans., Justin, Cornelius
Nepos, and Eutropius (London, 1876), pp. 37, 38.
 ACCESSION OF ARTAXERXES I 67

(1st century B.C. to 1st century A.D.) merely alludes to Xerxes'

murder by Artabanus, "one of his satraps," and Plutarch
(2d century A.D.) says that when the exiled Greek general
Themistocles came to the Persian court he obtained an
audience with the king by applying to Artabanus, the
Chiliarch, or commander of a thousand men. l5 Neither of
these writers dates the event, but both accept the view of
Thucydides (5th century B.c.), who says that Themistocles
came to Persia when Artaxerxes "had lately come to the
throne," as against the views of others that it was in the
reign of Xerxes. l6 Nepos points out that Thucydides was the
nearest in time to Themistocles and was from the same city. l7
One other ancient writer mentions the death of Xerxes:
Aelian (3d century A.D. or earlier) says merely that he was
"murdered a t night in bed by his son." ls In attributing
the murder of Xerxes to his son, he agrees with none
of the other historians extant. That could be merely the error
of a later writer, but it could be possible, though unlikely,
that it reflects a variant tradition stemming from the partisans
of Artabanus.

Ancient Chronological Works

In addition to the historical narratives, there are several

chronological works of the early Christian period that are
relevant to the question of Xerxes and Artaxerxes.

16 Cornelius Nepos, Lives, xxi ("Of kings").^ (in Watson, op. cit.,
p. 413) ; Plutarch, Themistocles, 27.1-5 (Loeb ed., 11, 72-75).
16 Thucydides, i.137.3 (Loeb ed., I , 232, 233). Plutarch (loc. cit.)
says that Ephorus, Dinon, Clitarchus, Heracleides, and others hold
that it was Xerxes, but he prefers the view of Thucydides and Charon
of Lampsacus (the latter contemporary with Themistocles) that it
was Artaxerxes because the chronological data agree better with this
view. Diodorus holds t h a t it was in the reign of Xerxes (xi.56.5 to
58.3 [Loeb ed., IV, 270-2771).
17 Nepos, ii ("Themistocles").g (in Watson, op. cit., p. 321).
18 Claudius Aelianus, Va~iaHistoria, xiii.3 (Leipzig, 1819, p. 194).
68 JULIA NEUFFER

Ptolemy, noted Greek-Egyptian astronomer (zd century

A.D.),in his Canon of the Kings, already mentioned, gives a
scale of Egyptian years (of 365 days, with no leap years)
beginning with the year I of Nabonassar of Babylon on
February 26, 747 B.C. He assigns 21 years to the reign of
Xerxes and 41 years to Artaxerxes immediately following. l9
This does not indicate whether he, like the compiler of the
EPitome of Manetho, regarded the period of Artabanus as a
separate reign, for Ptolemy's Canon omits al! kings who
ruled less than a year. But since the Canon is dated beyond
doubt by nineteen eclipses and other astronomical synchro-
nisms, it is certain that in the official Egyptian reckoning
Xerxes' year 21 (the year 283 in Ptolemy's Nabonassar Era)
began on Thoth I, December 18, 466 B.c., and that Arta-
xerxes' year I was the Egyptian calendar year beginning with
Thoth I, December 17, 465, and ending with December 16,
464-
Among the Christian chronographers, Julius Africanus
(3d century A.D.)and Eusebius (4th century) used Manetho's
chronology. They both included Artabanus with a seven-
month reign between Xerxes and Artaxerxes, i.e., in the 4th
year of the 78th Olympiad (46514). They also dated Arta-
xerxes' year 20 in the 4th year of the 83d Olympiad (which
makes his year I fall in 46413). 20

19 For Ptolemy's Canon, see Claudius Ptolemaeus, The Almagest,

R. Catesby Taliaferro, trans. (Great Books of the Western World, vol. 16:
Ptolemy, Copernicus, Kepler) ,Appendix A, p. 466 ; for the Greek Text,
see Claudius Ptolemaeus, Mathematike Syntaxis [Almagest], [Nicholas]
Halma, trans., I (Paris, 1813), lxx, Ixxi. I t is also printed in Thiele,
op. cit., p. 216. Any year in the Nabonassar Era can be computed
from the starting point by years of 365 days only, beginning a day
earlier every four years, because of the difference a t each leap year.
20 Julius Africanus, Cbronograpby, Fragments in ANF, VI,
135, 137; also table in Eduard Meyer, Forschungen zur alten
Geschichte, I1 (Halle, 1899), 487; Eusebius, Chronici Canones,
Jerome's Latin version, J , K. Fotheringham, ed. (London, 1923),
p. 192; cf. Armenian version, J. B. Aucher, ed. (Venice, 1818), pp. 208,
209.
 ACCESSION O F ARTAXERXES I

Evaluation of Ancient Accounts

In evaluating the ancient historical accounts it is clear that
the situation portrayed is fairly consistent in the various
narratives. Probably Ctesias' story is the nearest we can get
to the original-at least as told from Artaxerxes' point of
view, which naturally became the official version. The extant
summary of Ctesias says nothing of how long Artabanus was
in power or how he met his death, though additional details
in his original account, now lost, may have been the source
for later narratives of Diodorus and Trogus.
Diodorus seems to imply, though he does not say, that the
whole upheaval was over immediately; yet he goes on to
say that it was two years later that Artaxerxes settled the
kingdom. The stories of Ctesias and Trogus, even in their
present abridged state, definitely require some interval to
allow for the first coup to "prosper" and for the development
of the threat of a struggle among the nobles before the
inception of the second plot, to put Artaxerxes out of the way.
In the nature of the case, the fact that Artabanus did not kill
Artaxerxes a t first but allowed him to occupy the throne a t
least in name, and only afterward plotted against him, would
indicate that some time must have passed before he felt
strong enough to make the attempt to seize the kingship for
himself.
Then even after Artaxerxes killed Artabanus he had to
fight his way to control. There was a t least one battle against
the latter's adherents, and there was a revolt in Bactria,
possibly representing the claim of his brother Hystaspes. In
all, the events could account for much more than seven
months.
Yet an actual seven-month reign of Artabanus preceding
Artaxerxes' accession does not fit the picture drawn by the
historical sources (which, of course, represent mostly the
official story from the side of Artaxerxes). Not one of the
extant accounts calls Artabanus "king." He is referred to as
7O JULIA NEUFFER

"powerful" (Ctesias), "captain of the royal bodyguardJ'

(Diodorus), "chief officer" (Trogus), "satrap" (Nepos),
"commander of a thousand men" (Plutarch), while Arta-
xerxes is called king.
There are no known documents dated in the reign of
Artabanus in either Egypt or Babylonia. The king list based
on Manetho seems to be the only source for such a reign.
I t is possible that he could have been recognized in Egypt
only, or the attribution could have been an error rising from
the fact that Artabanus for a time-and possibly for about
seven months-was the real power while the young Arta-
xerxes was the puppet king.
I t may be that the confusion as to whether Themistocles
came to the court of Xerxes or of Artaxerxes could be account-
ed for by supposing that he came during the period of Arta-
banus' ascendancy, while Artaxerxes was king but not yet
ruling (note Plutarch, as cited above). And this situation may
find an echo in the artificial extension of Xerxes' regnal
numbering after his death as attested by papyrus A P 6,
as will be discussed below.

Use of the Ancient Sowces

Before modern archeology furnished contemporary dated
documents from ancient times, and when the only authority
for chronology was Ptolemy 's Canon and the ancient histo-
rians, many writers on Biblical interpretation and chronology
in the last three hundred years discussed the chronology
of Artaxerxes because of the Biblical mention of his 7th and
20th years. They included Johann Funck (1564), Archbishop
Ussher (1650), William Whiston (1702), and Isaac Newton
(1728, 1733), as well as numerous 19th-century writers.
Several, including Ussher, accepted Thucydides' identification
of Artaxerxes as the king to whom Themistocles went, but
accepted a dating of Themistocles that put his visit, and
therefore the accession of Artaxerxes, nine or ten years
 ACCESSION OF ARTAXERXES 1 71

earlier. 21 Newton reckoned Artaxerxes' year I as beginning

in August or September, 464-a regnal year based on several
erroneous assumptions : (I) that the Canon, always antedating,
placed the death of Xerxes after Thoth I (December 17).
465 ; (2)that Artabanus ruled seven months after that before
Artaxerxes' accession; (3) that Artaxerxes came to the throne
two or three months after the summer solstice and counted -
his regnal years in the same manner as the British kings-
as beginning always on the date of his accession.
Modern historians tell the story by piecing together bits
of the various ancient accounts. W. W. Tarn, in the Cambridge
Ancient Histov ( ~ g q )says, that Artabanus reigned seven
months and was recognized in Egypt (based apparently on
Manetho) and that he defeated Artaxerxes' brother Hystaspes
(a recombination of elements from Ctesias and Diodorus?)
before Artaxerxes killed him. 23 A. T. Olmstead presents
Artaxerxes as eighteen years old (a guess from Trogus);
Megabyzus as involved in the original conspiracy; and
Hystaspes, Xerxes' other son, as heading the Bactrian revolt
and being defeated by Artaxerxes after Artaxerxes killed
Artabanus (Diodorus ?). 24
Most historians disregard Artabanus, largely because the
absence of tablets dated to his reign would indicate that he
was not recognized in Babylonia. Indeed, when it was believed
that the nearest contemporary documents (papyrus A P 6 and
the Ur tablet UET IV, 193) meant that Xerxes was living
until near the end of 465, there could be no room for Artabanus

a1 James Ussher, Annales Veteris Testamenti (London, 1650), on

Anno Mundi 3531;in the English version, Annals of the World (London,
165% pp. 131, 132.
32 Isaac Newton, The Chronology of Ancient Kingdoms Amended
(London, 1728), pp. 353-355; Observations Upon the Prophecies
(London, 1733)~pp. 130, 131, 142, 143.
8s W. W. Tam, in Cambridge Ancient History, VI (New York,

1927)~2 .
a4 A. T. Olmstead, History of the Persian Empire (Chicago, 1948),
pp. 2 8 9 , 2 9 0 .
72 JULIA NEUFFER

as a factor in the chronology. The ancient writers are against

his recognition in Persia, though he could have been recognized
in Egypt. Yet A P 6, written in Egypt-possibly during the
period when he was in de facto control-ignores him. (How-
ever, its dating formula does imply that the transfer of power
to Artaxerxes was not immediate and normal, and implies
the sort of confused situation pictured in the other ancient
sources.)

Contemporary Papyrus A P 6 Examined

Though historical sources furnish an interesting and prob-
ably relevant background for understanding the contents
of the tablets or papyri, actually the conditions implicit in
a contemporary document outweigh those in late copies of
worked-over historical narratives. As primary evidence, then,
the only known contemporary document, papyrus A P 6,
must now be examined, and with it must be considered the
Hellenistic tablet (LBART No. *141g), which places Xerxes'
death in Abu 14-18 (August 4-8), 465. Though the evaluation
of this tablet must await its publication, it can meanwhile be
accepted tentatively as possibly correct and be considered in
the light of the contemporary Aramaic papyrus A P 6.

Double Date in Two Reigns

The dateline of A P 6 reads: "On the 18th of Kislev, that
is the [17th] day of Thoth, in year 21, the beginning of the
reign when King Artaxerxes sat on his throne."
Like many other papyri from this Jewish colony in Egypt,
it is double-dated in two reckonings, the Egyptian solar
25 Cowley, op. cit., p. 16. Cowley reads the broken day number
conjecturally as "7th day of Thoth." But he did not do any calendar
computation; that was done later by others. The lunar-solac calendar
synchronism is possible only if the Thoth date is read "17th," which
is equally possible paleographically; see S. H. Horn and L. H. Wood,
"The Fifth-Century Jewish Calendar at Elephantine," JNES, XI11
(1954)~8, 9, PI. I.
 ACCESSION OF ARTAXERXES I 73

calendar and the Semitic (either Persian or Jewish) lunar

calendar.
I t has already been explained that the first part of this
dateline, with its synchronism between a solar and a lunar
month date, leaves no uncertainty that this represents
January 2/3,464. The remainder of this article will examine
the last part of the dateline-the regnal year formula : the
year 21 (of Xerxes, obviously), and the accession year of
Artaxerxes. Does this double dating of the year represent
the difference between the Egyptian and Jewish reckonings ?
There are two other papyri from Elephantine that furnish
examples of such a dating in two regnal years: AP 25 and
AP 28. For example, AP 25 equates "Kislev 3, year 8" with
"Thoth 12, year g" in the reign of Darius 11. That is, by the
3d of Kislev, the ninth month of the Semitic lunar year, the
Egyptian New Year had passed, and this was the 12th of
the first month in the new regnal year g by Egyptian count. 26
But A P 6 not only has two regnal year numbers; the two
are irt two different reigns. I t does not represent a coregency
of Artaxerxes with his father. The historical accounts of
Xerxes' death show that Artaxerxes was not even the crown
prince, and did not become king until after the death of his
father and his older brother.
There are three possibilities in explaining this unusual
dating in two reigns a t once: (I) It was a scribal error. (2) It
represents, like the two month dates, the difference in reckon-
ing between two calendars, Egyptian and Semitic. (3) I t is a
double year designation in one calendar.

Was the Double-Year Formula an Error?

Some have thought that this unusual double-reign dating
formula was an absent-minded error of the scribe who wrote
it. This was plausible when it was believed that Xerxes had
only recently died, in late December, for the scribe could

Horn and Wood, op. cit., p. 17.

74 JULIA NEUFFER

have begun with "year 21" as he had been doing for some
time, and on remembering that Artaxerxes was now king,
merely added the accession-year formula without correcting
the initial error. 27 But this was an official document written
by a professional scribe; he would be expected to begin over
rather than merely to add the correct dating to the erroneous
phrase, especially since "in the year 21" stood in the first line
of the document. And forgetfulness is not an easy explanation
if, as the Hellenistic tablet (LBART No. *141g) indicates,
the change of kings had not been recent but some five months
earlier.
Other Examples of Dating in Two Reigns. But it is not
necessary to suppose a mistake, since there are other examples
of this unusual type of year formula. In the case of the next
regnal transition, after the death of Artaxerxes I, there are
three tablets double-dated in two reigns. That was also a
period of murders, plots and counterplots, and competing
claimants, with the resultant uncertainty of the status quo.
This is not the place to go into the problem of exact dates and
intervals, but suffice it to say that a t the death of Artaxerxes I
his son Xerxes I1 occupied the throne briefly (45 days), then
was killed by a half brother Secydianus, or Sogdianus, who
was himself killed (after about seven months) by another
half brother who reigned as Darius 11. 28 There are no known
tablets recognizing Xerxes I1 or Sogdianus. Perhaps the
length of time assigned to them by the Greek historians was
exaggerated.
There are tablets dated to Artaxerxes as late as the 9th
month of his year 41 (December, 424)) possibly also in the
11th month (February, 423); and there are two dated un-
equivocally to Darius' accession year in the 11th month. Yet
there are two other tablets in the 12th month and one (yet
See PDBC (1g42),p. 16, for this interpretation in a similar case.
27
Manetho, loc. cit.; Ctesias, op. cit., 45-48, (Brussels ed., pp. 44-46) ;
88
cf. Diodorus Siculus, xii.64.1, 71.1 (V, 60, 61, 78, 79) ; cf. Thucy-
dides, iv.50.3 (11, 298, 299).
 ACCESSION OF ARTAXERXES I 75
unpublished) supposed to be some months earlier-dl three
double-dated in the last year (of Artaxerxes) and in the
accession year of Darius. They appear to reflect an unwilling-
ness to abandon reckoning by Artaxerxes' reign, as if it were
still uncertain as to whether the reign of Darius was perma-
nent. I t is significant that these tablets and the papyrus
AP 6, which seem to have the only such double datelines
known, come in both cases from periods when the uncertain
political situation would provide a reason for such an unusual
extension of a king's regnal numbering even beyond the
beginning of another reign.
Reign Artificially Extended Into Another Year. There are
several other tablets, from an earlier period, that similarly
show an abnormal prolongation of regnal dating, and in this
case using a ruler's name, not only after his death, but even
into a new year, with a new regnal number. This was in
another period of upheaval, when Assyria's rule over Babylon
ended.
In 627 the last known Babylonian tablet dated in the
reign of Kandalanu (who ruled Babylonia under Assyria)
was written on the 13th of the 2d month of year 21. Then
there were two later ones obviously after his death: one in
Marcheswan, or Arahsamnu (the 8th month), dated year 21,
not "of Kandalanu," but "after Kandalanu"; and the other
a year later, Marcheswan 2, in year 22 "after Kandalanu."
The intervening year was afterward reckoned an interregnum,
after Kandalanu was gone but before Nabopolassar succeeded
in fighting his way to independence for Babylonia and in
winning the throne; but during that time the old regnal
reckoning in Kandalanu's name was continued, even into a
new and fictitious "year 22." And a chronicle tablet calls
this year "after Kandalanu, in the accession year of Nabo-
polassar ." 29

a@ D. J. Wiseman, Chronicles of Chaldaean Kings (626-556B.C.) in

the British Museum (London, 1961), pp. 89-90. PDBC (1956), p. 11,
76 JULIA NEUFFER

Nabopolassar had already won recognition as king in a t

least a part of Babylonia, as attested by tablets dated to
his accession year in the 2d and 6th months; but while the
fighting and uncertainty lasted, the old reign was carried
on artificially until the 8th month, within 24 days of the time
when he occupied the throne. For the Babylonian chronicle
tablet says that on the 26th of the 8th month (approximately
November 23,626) "Nabopolassar sat upon the throne in
Babylon. (This was) the 'beginning of reign' of Nabopolas-
sar." 30
Except for the distinction made by the term "after Kanda-
lanu," this reckoning of a year 22, although he had died in
year 21, furnishes an exact parallel to the other examples of
dating in the name of a king after his death, and after a new
king was recognized as ruling.
Since the extension of one king's regnal reckoning beyond
his lifetime, into the reign of another king, is attested both
before and after the time of papyrus AP 6, then its double
dateline in the reigns of Xerxes and Artaxerxes is not neces-
sarily a scribal error. Nor is it necessarily a double dating
in two calendars, for the tablets just discussed involve only
the Babylonian-Persian calendar. Yet, in order to test all
the possibilities, the AP 6 dateline will be investigated in
both alternatives-whether the two year datings are expressed
in two calendars, Egyptian and Semitic, or whether both
are in one calendar (and if so, which one).

I s the Year Formula Expressed in Two Calendars ?

First, suppose that the double year formula of AP 6 repre-
sents the two calendars in which the month dates (Kislev
and Thoth) are expressed. Then obviously either "year 21"

presents this interpretation of a reign extended artificially, citing

Wiseman's first (1956) printing.
30 Wiseman, op. cit., p. 51 ; cf. PDBC (1956), p. I I. (This exact date
for the accession, not known before, shows that Ptolemy's canon
postdated Nabopolassar's reign.)
 ACCESSION OF ARTAXERXES I 77

or "accession year" must be in the Egyptian calendar. Yet

a glance a t the Egyptian calendar as represented in the
horizontal band labeled "Egyptian" on Fig. I, will show that
the heavy arrow representing the papyrus date does not fall
in either the year 2 1 of Xerxes or the accession year of Arta-
xerxes in the Egyptian calendar. What is wrong?

Figure I. Artaxerxes in the Egyptian Calendar

The last regnal year (2 I ) of Xerxes and the early years of Artaxerxes
are shown here as reckoned in the Egyptian calendar, compared
with the B.C. scale. The Egyptian years, beginning in December in
this period, run a little earlier than the B.C.years (shown extended
by the broken lines). The N. E. (Nabonassar Era) numbering, derived
from Ptolemy's Canon, is indicated for the years 465 and 464. The
arrow shows the date of the papyrus A P 6 (January 213, 464 B.c.).
On the one hand, Xerxes' Egyptian year 21 undoubtedly
began on Thoth I (December 18), 466, according to the
astronomically fixed canon of Ptolemy and a double dated
papyrus (AP 5) of his year 15; then on the next Thoth I
(December 17, 465) the year number would have changed
to year 22. Yet sixteen days later, on January 213, 464, AP 6
was still dated in "year 21"!
On the other hand, the Egyptian year 465/4, in which A P 6
was written, was officially numbered Artaxerxes' year I,
not his accession year. (This is attested not only by Ptolemy's
Canon but also by several double-dated papyri written during
Artaxerxes' reign, all of which require year I to begin in
December, 465 B.c.) If he came to the throne before Thoth I
(December 17)) 465-perhaps the preceding August 4-8, as
the Hellenistic tablet (LBART No. *141g) indicates-his
"beginning of reign" could cover only the rest of that calendar
year, and his year I would begin on Thoth I. 31
31 The common custom of the Egyptians was to "antedateD'-to
begin dating in "year I" immediately after the accession and change
78 JULIA NEUFFER

If AP 6, written as late as Thoth 17, was dated in the

accession year in the Egyptian calendar, this would indicate
that Artaxerxes was recognized as king in Egypt only after
Thoth I (else this would have been his year I) 32 and before
Thoth 17 (else his name would not have been on the dateline
at all).
Further, if this was the accession year, then year I would
not have begun until the following Thoth I, December 17,
464. That would conflict with the official year numbering,
also with the tablet that places the death of Xerxes in August
465, unless there was a delay in the recognition of Artaxerxes
until after Thoth I.
Could such a delay be accounted for by supposing it to
be during the seven-month reign assigned by Manetho to
Artabanus ? Yet the interval between the August death date
and the January date of A P 6 is less than seven months.
And an intervening reign of Artabanus would still require

to year 2 a t the first New Year's Day. Yet there is some reason to
think that they sometimes applied the Persian postdating method to
their Persian kings. See Parker, "Persian and Egyptian Chronology,"
AJSL,LVIII (1g41), 285-301.
82 The present writer formerly, in the above-mentioned thesis
(see note 3), accepted Cowley's designation of AP 6 as dated "year 2 I "
in the Semitic calendar and "year I" in the Egyptian calendar
because the date (January 213, 464) arrived a t by the synchronism
was in the Egyptian year I. But it seems necessary to abandon
"year I" in favor of "accession year" for the following reasons: (I)
The phrase r 'S mlwkt ' (sic.), "beginning of reign," in A P 6 is the exact
Aramaic equivalent of the Akkadian accession-year formula rEf
Sarrfiti (literally "beginning of reign"), defined as the accession year,
the time of reign before the beginning of the first full regnal year;
see Riekele Borger, Babylonisch-Assyrische Lesestucke, Heft I (Rome,
1963), Glossar, p. lxxvi; (2) a completely different phrase is used
for "year I" in Aramaic, "Snt I (with the king's name) ," which is
also the exact equivalent of the Akkadian date formula used in
Babylonian tablets; and (3) the explanatory but redundant clause
translated by Cowley "when King Artaxerxes sat on his throne"
can also be translated "when King Artaxerxes seated himself" or
"was seated" on his throne, that is, "when he became king" (Horn,
Letter to the author, Feb. 15, 1967).
 ACCESSION OF ARTAXERXES I 79

a change in Artaxerxes' year numbering afterward to continue

with the later attested numbering. Such a change is unattested
by any evidence and seems to be unknown in any other case.
Since the Egyptian year 21 is impossible for AP 6, and the
accession year is incompatible with known Egyptian data
and so unlikely as to be negligible, the logical result is to
rule out both as possible Egyptian datings; and therefore
to abandon the first alternative-a double (Egyptian-Semitic)
year formula-and proceed to the second:

A Double Year Formula in One Calendar

Not Egyptian. If both "year 21" and "accession year of
Artaxerxes" in the dateline of AP 6 constitute a double year
formula in one calendar, then it means that both are desig-
nations of the same year-the one that begins as year 21 of
Xerxes and ends as the accession year of Artaxerxes. This
cannot be an Egyptian-calendar year, since the Egyptian
year 21 ended seventeen days before this papyrus was written.
Then it must be a Semitic lunar-calendar date-in either
the Persian year (beginning in the spring with the month
of Nisanu) or the Jewish civil and regnal year (beginning
in the fall with the 7th month, Tishri).
Most Probably Jewish. Papyrus AP 6 (an agreement over
a disputed piece of land) was written in the name of a Persian
for the benefit of his neighbor, designated as a Jew; and the
scribe was a Jew, as well as most of the witnesses. 33 Although
the lunar calendar synchronism in A P 6 could be valid in
either the Persian or the Jewish reckoning, it seems logical
to conclude that it was a Jewish dating as used in a Jewish
community.
That this calendar was Jewish would be expected for
several reasons. These Jewish colonists of Elephantine had
been there some time before the Persians took over Egypt; 34
53 Cowley, 09. cit., pp. 16, 17.
"1 Ibid., p. xvi.
80 JULIA NEUFFER

hence they would have no reason to adopt the Persian

calendar, since they obviously had not adopted the Egyptian
calendar outright, or they would not have needed double
dating. That their Jewish calendar would have been the
same as the regnal reckoning of the Kingdom of Judah, from
which they had originated, and of the returned Jews of the
contemporary period of Ezra and Nehemiah, seems most
likely.
Some writers hold that these Jewish colonists, like the
Babylonians and Persians, used a spring-beginning year,
while others hold that they employed the Jewish autumn-
beginning year. The evidence for the Jewish reckoning by
years beginning with Tishri, in the autumn-used in the
early Hebrew kingdom, in the Kingdom of Judah, in the
restored Jewish community in the time of Ezra and Nehemiah,
and in this same Jewish colony in Egypt in a later reign 35-
makes it seem a reasonable conclusion that AP 6 was dated
in the Jewish fall-to-fall year.
However, since opinions differ, the dating of this papyrus
will be examined in both Persian and Jewish reckonings.

The a ~ s Year
t of Xerxes and the 1st Year of Artaxerxes
I t has been explained already that throughout the reigns
involved here the regnal-year numbering in the Egyptian
calendar is known from the astronomically fixed reckoning
of Ptolemy's canon and the synchronisms of several double-
dated papyri. I t is also known in the Persian calendar from
the saros list, based on the 18-year saros cycle. 36

Horn and Wood, op. cit., pp. 14-16, 20; Thiele, 09.cit., pp. 28-31.
85
The saros list is extant on two clay tablets containing a series
of regnal years a t eighteen-year intervals based on a Babylonian
eclipse cycle (published by J. N. Strassmaier in reports in Z A , V I I
[~Sgz],zoo, 201; VIII [1893], 106). Beginning with the 7th year of
Nabonidus, this list includes the year g of Xerxee and the years
6 and 24 of Artaxerxes.
 ACCESSION OF ARTAXERXES I

JAN. L
.4 A

ACC
I I I

Figure 2. Xerxes and Artaxerxes in Three Calendars

The regnal years of Xerxes (shortened in this drawing by an 18-year
gap) and the early years of Artaxerxes are shown as reckoned ( I ) in
the Egyptian calendar (with years beginning in December), (2) in
the Persian calendar (with years beginning in the spring), and (3)
in the Jewish calendar (with years beginning in the autumn), all
three aligned against the background of the B.C.years (extended by
broken lines). The three vertical arrows represent, from left to right,
(I) the accession of Xerxes (some time in November, 486 B.c., (2) the
death of Xerxes as indicated by the tablet LBA R T * 1 4 1 9 (August 4-8,
465 B.c.), and (3) the date of the papyrus A P 6 (January 213, 464 B.c.).
(I) In the Egyptian calendar (see Fig. 2, first band), the
year 2 1 of Xerxes was 46615 and the year I of Artaxerxes
was 46514, beginning in December.
(2) In the Persian calendar (Fig. 2, second band)-with
years beginning with Nisanu I, in the spring-year 21 of
Xerxes was 46514 (beginning approximately March 25, 465),
and the year I of Artaxerxes was 46413 (beginning approxi-
mately April 13,464), several months later than the beginning
of the corresponding Egyptian years.
(3) Then in the Jewish calendar the zrst year of Xerxes as
reckoned according to the fall-to-fall year can be determined,
with equal accuracy, as 46514; it began with the Jewish
New Year, the 1st of Tishri, the 7th month (approximately
October 18, 465),3' half a year later than the Persian New
Year. (Discussion of the year I of Artaxerxes according to
this Jewish reckoning will be deferred until after the expla-

37 The equivalents of the Jewish dates are taken from the recon-
structed calendar tables of Horn and Wood, but they are approxi-
mately the same as those in PDBC (1956).
82 JULIA NEUFFER

nation of why the ~ 1 s year

t of Xerxes runs later than the
Persian year 2 I .)
Alignment Deeends on Accession Date. The alignment of
the Persian and Jewish years of Xerxes tor of any postdated
reign) depends on whether the Persian or Jewish New Year
came first after the accession. This can be explained best
with the aid of Fig. 2. Since Xerxes' year 21 in the Persian
calendar was 46514, his year I was 48514, from spring to
spring; he must have come to the throne some time before
that, since his "accession yearJJ was the part of his reign
that preceded his first full calendar year. The date of his
accession can be determined as some time in the preceding
November, 486, because the latest known tablet dated in
his father's last regnal year was in the 7th month, and the
first dated in Xerxes' reign was in the 8th month (approxi-
mately December I). 38
After his accession in November, 486, the first New Year's
Day to arrive would be the Egyptian 1st of Thoth, in Decem-
ber (Fig. 2, band I) ; 39 the Egyptian year I began then, but
those of his subjects who used the Babylonian-Persian
calendar (band 2) would not begin to date by his year I until
the next Nisanu I, in the following spring; and those who
used the Jewish fall-to-fall calendar (band 3) would continue
to date in the accession year until their New Year's Day
came-the next Tishri I, the 7th month-almost a year
after his accession and half a year after the Persian year I
had begun. (That is why, in some of these papyri, the dateline
in two calendars can have two regnal year numbers.)
Thus, throughout his reign, any specific year of Xerxes-
from year I through year 21-began earliest in the Egyptian
calendar, then in the Persian calendar, and last in the Jewish
calendar. Then it is demonstrated that the Jewish year 21,
reckoned from Tishri I, must be 46514. And since year 21

s* PDBC (1956),p. 17.

SQ For this Egyptian postdating of Xerxes, see note 31.
 ACCESSION OF ARTAXERXES I 83

is equated with the accession year of Artaxerxes, it is obvious

that the accession year according to this fall-to-fall reckoning
would end in 464 (and consequently the Jewish year I would
be 46413)
Since January 213, 464, the date of AP 6, falls in year 21
and the accession year in both the Persian and the Jewish
reckoning (see heavy arrow in Fig. z), this papyrus date could
have been either Persian or Jewish.
But there remains the question : Why would there have
been a "year 21''for Xerxes in the Jewish calendar (see Fig. 2,
band 3), beginning in October, 465, if Xerxes had already been
murdered in the preceding August, in the Jewish year zo?
The question of the artificial extension of Xerxes' reign must
be answered regardless of whether the date is in the Jewish
or the Persian calendar. However, such a practice of extending
a regnal year after a king's death-even of beginning a new
year number-has been shown to be a normal, if exceptional,
practice under certain circumstances, as demonstrated in
the Kandalanu-Nabopolassar transition and the Artaxerxes
I-Darius I1 transition.

The Implications of AP 6
The next step, then, will be to examine AP 6 to see whether
its unusual dating, in either the Jewish or the Persian calendar,
can likewise be considered an unusual but normal dating
formula in relation to the political situation.
As a help in visualizing the following possibilities in both
the Persian and the Jewish calendar, the year in which AP 6
was written is marked on Fig. z in heavy lines in the second
and third bands. In each of these calendars it is the year
that began as Xerxes' year 21 and ended as the accession
year of Artaxerxes.
Here is what the AP 6 dateline itself tells us, as can be
seen on Fig. z: (I) A change of reign was recognized a t some
time before January 213, 464, when this papyrus was written
84 JULIA NEUFFER

(heavy arrow), but not earlier than Nisan I in the preceding

spring (otherwise the papyrus dateline, if in the Persian
calendar, would have had "year I," not "accession year"
of Artaxerxes), and scribes in Elephantine began to date their
documents in the name of the new king, Artaxerxes; yet (2)
they retained the regnal numbering of Xerxes, a t least as
late as January, as if he were still alive and still reigning;
(3) if the death of Xerxes occurred in August, 465, they
extended his last year for a t least five months longer, (4)
continuing year 21 if they were using the Persian calendar
or (5) year 20 if they were using the Jewish fall-to-fall calendar,
and if the latter, they were so unwilling to drop Xerxes' regnal
numbering that on Tishri I, two months after his death, they
even began a fictitious, additional year 21 rather than change
to the accession year of Artaxerxes; in that case (6) they
did not recognize Artaxerxes' reign until after Tishri I, 465
(otherwise they would have dated A P 6 in year I, not accession
year) ; (7) if Xerxes' death occurred, not in August, but
after Tishri I, in the autumn, the Jewish year 21 would have
begun normally, in his lifetime; (8) Artaxerxes may have
been recognized immediately, but with reservations, since,
(9) in either calendar, year 21 would have been artificially
extended after Artaxerxes' accession.

Relation to Historical Situation

What could have been the reasons back of this reluctance

to relinquish the old year numbering of Xerxes? Was it
unwillingness to recognize the young Artaxerxes as king or
uncertainty whether someone else might prevail in the end?
(I) If the unpublished tablet (LBART No. "1419) is in
error-if Xerxes did not die in August-he could have, as
was formerly supposed, lived until December, not long before
papyrus AP 6 was written. In that case Artaxerxes would
have been given immediate recognition, and there would
 ACCESSION OF ARTAXERXES I 85
have been no time for an intervening reign of Artabanus. *O
Why, then, was the recognition of Artaxerxes qualified by
the retention of Xerxes' "year 21" ? This would indicate an
initial, though possibly brief, uncertainty as to his hold on
the throne. Was it the presence of an older brother in Bactria ?
Or the immediate control by Artabanus?
(2) If the interpretation is not to be built on the supposition
of errors in these source documents-then from the combi-
nation of these two documents, the tablet and the papyrus,
it should be possible to derive an interpretation that is not
incompatible with any of the data. The persistence of the
regnal dating of a long-dead Xerxes indicates a prolonged
period (at least five months) of uncertainty or unwillingness
to give unequivocal recognition to the reign of Artaxerxes.
If the new king was still so shaky on his throne after five
months, there must have been a powerful rival or rivals
who threatened his authority.
I t is not clear whether his older brother Hystaspes, absent
in Bactria, was a menace, but certainly the most powerful
man in the kingdom was Artabanus. This was the commander
of the royal guard, who, according to the ancient historians,
was the most influential of the courtiers, the real power
behind the throne, the man who had murdered Xerxes and
to whom the young Artaxerxes owed his somewhat precarious
occupancy of the throne. Possibly, in Egypt a t least, the
de facto power of Artabanus overshadowed the de jure authori-
ty of Artaxerxes.
If the dating of AP 6 is Persian, the double dating would
indicate a prolonged period of uncertainty as to the situation
of Artaxerxes. If it was in the Jewish calendar, it would
40 Hence the present writer formerly, in the above-mentioned
thesis (see note 3), ignored Artabanus as having any place in the
chronology. But if August 4-8 is correct for Xerxes' death, then
Artabanus' initial control of Artaxerxes may be taken into account
as an explanation of the extension of Xerxes' regnal numbering in
this papyrus dating; also Ptolemy's Canon, in agreement with tablet
and papyrus dating, postdates Artaxerxes' reign.
86 JULIA NEUFFER

appear to indicate even more-a gap between the death of

Xerxes and the recognition of Artaxerxes. For a scribe using
this Jewish calendar could not have begun dating in the
accession year of Artaxerxes until after Tishri I, in October,
or even later (otherwise AP 6 would have been dated "year
21, year 7 of Artaxerxes").
Did the Elephantine colonists, or all of Egypt, recognize
someone else in the interval? Perhaps Artabanus, who was
assigned a seven-month reign in Egypt by the Manetho
Epitome? If so, they must have abandoned him to recognize
Artaxerxes sometime between Tishri I and January. Even
then they did not feel free to abandon the old Xerxes dating,
as if the outcome were still not settled.
A reign of Artabanus in Egypt seems a doubtful explanation
because it does not fit the historical narratives, because the
interval between August and January is less than seven
months, because the retention of Xerxes' year 21 would
seem unlikely if another king had been recognized in the
interval, and because there is no evidence of Artabanus'
recognition in the Babylonian tablets (though neither is
there any known tablet for Xerxes' last year or Artaxerxes'
accession year). However, since AP 6 seems not to be dated
in an Egyptian year, it would not be expected to furnish any
indication of what the Egyptian regnal formula would have
been.
Nor is it necessary to suppose that Artabanus was actually
a king. If during seven months, or less, he was dominant as
the real ruling power behind the throne, that would account
for a situation in which f u l l recognition of the young Arta-
xerxes was delayed. This divided and delayed recognition
finds parallels, as has been shown in the dating formulas of
tablets written in other periods of upheaval and confusion,
when it was not clear which of the contenders would prevail.
In the present case the historical sources corroborate the
papyrus in picturing just such a situation of dynastic struggle.
Then we may take AP 6 as reflecting such an interim
 ACCESSION OF ARTAXERXES I 87

situation, and there is no disagreement between the earlier

death date for Xerxes and the dating of this papyrus.

Summary
In summary, then, the evidence of this contemporary
papyrus, combined with that of the Hellenistic tablet and
viewed against the background of the earliest historical
narrative and compared with later accounts, leads to the
following conclusions :
(I) There is not necessarily any basic discrepancy between
these sources.
(2) A period of uncertainty between Xerxes' death and
Artaxerxes' full recognition implied in the papyrus is com-
patible with the tablet, and the reflection of such a situation
in the dating formula is paralleled by other examples in
similar periods.
(3) Such an interval of instability agrees with the historical
accounts concerning Xerxes, Artabanus, and Artaxerxes.
(4) The use of "year 21" and "accession year," in either
Persian or Jewish dating, agrees with the fact that no known
Babylonian tablets recognize any other king between Xerxes
and Artaxerxes, though a Jewish dating implies a gap before
the beginning of Artaxerxes' accession year in Egypt.
(5) The alignment of the regnal years of the same number
in the Egyptian and the two Semitic calendars (attested by
the synchronisms of Ptolemy's canon, the saros list, double-
dated papyri, and dated tablets) follows this order: Egyptian
(December), Persian (spring), Jewish (fall), in the reign of
Xerxes; likewise in the reign of Artaxerxes the order is:
Egyptian, followed by Persian, followed (if A P 6 has a Jewish
date) by the Jewish.
(6) This alignment makes it clear that the year formula in
A P 6 does not fit the Egyptian calendar, but is an exceptional
but normal double formula in either the Persian or the Jewish
calendar.