“And in the fourth year Egypt rebelled ...” The Chronology of and Sources for Egypt’s Second Revolt (ca. 487–484 BC)

Scholars continue to give different dates for Egypt’s second revolt against the Persians: Classicists generally date the revolt to 487–485 or 487/ 486–485/484 BC; Egyptologists and historians of the Achaemenid Empire generally date it to 486–485/484; while some scholars date it to 486/485–485/484. Such chronological differences may sound small, but they have important consequences for the way the rebellion is understood. The purpose of the present article is therefore twofold: first, it aims to clarify what we can and cannot know about the rebellion’s exact chronology. After a review of the relevant evidence, it will be argued that the best chronological framework for the rebellion remains the one provided by Herodotus’sHistories, which places the rebellion in ca. 487–484. Second, the article will show how this chronology influences our understanding of the geographical extent and social impact of the rebellion. The adoption of Herodotus’s chronological framework, for example, results in a larger number of Egyptian sources that can be connected to the period of revolt than was previously recognized. These sources, it will be argued, suggest that some people in the country remained loyal to the Persian regime while others were already fighting against it. Moreover, they indicate that the revolt reached Upper Egypt and that it may have affected the important city of Thebes.

the Achaemenid Empire. 1 However, when the rebellion began and ended exactly is still disputed. Classicists generally date the revolt to 487-485 or 487/486-485/ 484 BC, based on several passages in Herodotus's Histories. 2 Most Egyptologists, on the other hand, use a combination of Herodotus and Egyptian texts to date the rebellion to 486-485/484a date which has been adopted in Achaemenid studies more widely. 3 Some yet go a step further and date the revolt to after Darius's death, which would move the revolt to the beginning of Xerxes's reign in 486/ 485. 4 Which dates are we to follow?
Whether the second Egyptian rebellion against the Persians began in 487, 486 or 485 may not sound like a major problem. Most historians of antiquity are used to chronological imprecision; and even though the dates for the rebellion differ with about a year or two, such a margin of error is still relatively small. The exact date, however, has serious consequences for the way the rebellion is understood.
Questions about which Egyptian sources should be connected to the episode, and consequently which regions in Egypt and which classes of society were affected by it, are all caught up in the issue of the event's beginning and end. The purpose of this article is therefore twofold: first, it aims to clarify the chronology of the revolt as given by Herodotus's Historiesa chronology which has sometimes been misunderstood. It will argue that Herodotus's chronology places the rebellion in ca. 487-484; and that this chronology, though vague, remains preferable to a chronology based on Egyptian date-formulae. The dates of the revolt which are current in Egyptology and Achaemenid studies more widely need to be revised accordingly. Second, the article will show how the adoption of Herodotus's chronology influences our understanding of the Egyptian revolt in general. It will be argued, for example, that a larger number of Egyptian texts can be related to the period of revolt than was previously recognized. Though these texts have their limitations, they provide us with important information on the rebellion's geographical extent, and on the division of political loyalties in Egypt at the end of Darius's reign. The Egyptian sources therefore deserve a closer look. Let us start, however, with the issue of chronology.

I The chronology of Herodotus
There are several ancient sources that refer (or may refer) to an Egyptian rebellion in the 480 s BC. According to Aristotle (Rh. 2.20), for example, Xerxes invaded Greece only after he had captured Egypt, which suggests that Egypt had rebelled in or before Xerxes's early reign. Something similar is suggested by a royal inscription from Xerxes's reign, which claims that one of the Empire's satrapies was in rebellion when he became kingafter which Xerxes subdued the unrest (see XPh, the so-called "Daiva Inscription"). 5 The primary source on which all studies of Egypt's second revolt are founded, however, is Herodotus's Histories.
Herodotus mentions the Egyptian revolt at the beginning of Book 7 of his work. The mention follows a long narrative about Darius's failed invasion of Greece (which culminated in a Persian defeat at the battle of Marathon), and it directly precedes the story of Xerxes's campaign against Athens. What Herodotus says about the Egyptian revolt in-between those two events is actually not that much: the only things we are told is that Egypt revolted, and when it didnot why, how, or who specifically. It is only because the other texts that mention the rebellion are even less forthcoming, that Herodotus's brief remarks remain the goto source. The following is his (abbreviated) version of events: § 7.1. After Darius had sent an army against Greece, his Persian forces were defeated at the battle of Marathon. The king was furious when he heard the news. Messages were sent around the empire to raise a brand-new army; and for three years, Asia was in turmoil due to Darius's military demands. In the fourth of those years, Egypt rebelled. § 7.2-4. At roughly the same time, while Darius was planning an attack on both Egypt and Greece, Darius's sons started to quarrel over the issue of royal succession. Artobazanes and Xerxes thought they had the right to succeed their father on the throne. Darius chose Xerxes, however, as his heir apparent; and in the year following this, Darius died. § 7.5-7. Xerxesthe new kingwas reluctant to make war on Greece. Several different parties eventually persuaded him to do it anyway. But Xerxes first sent an expedition against Egypt, in the year after Darius's death. Xerxes crushed the rebellion and reduced the Egyptians to a state of even worse slavery than they had experienced under his father. He installed his brother Achaemenes as satrap of the Two Lands. § 7.20. After Egypt's defeat, Xerxes spent four whole years preparing and equipping his army for the planned invasion of Greece. He and his men set out in the course of the fifth year. And 5 As the Daiva Inscription does not mention the name of the satrapy, it is possible that it referred to a different event, e. g. to the Babylonian rebellions in Xerxes's second regnal year (on which see Waerzeggers 2003Waerzeggers /2004. The Egyptian sources for the revolt will be discussed below.
the army which marched to Europe, Herodotus says, was larger than the world had ever seen before.
The rest of Herodotus's narrative leads up to Xerxes's infamous crossing of the Hellespont, the deadly battle at Thermopylae, the occupation of Athens by the Persian army, and so forth. These events, which cover only two years in all (480-479), take up a third of Herodotus's Histories (Books 7-9). They are the legendary high-point of the hostilities between Greeks and barbarians which he promised to narrate at the start of his book (1.1).
The Julian dates. It is a fortunate coincidence that Herodotus mentions the Egyptian revolt in a larger historical narrativeone which is filled with chronological indicators. Classicists have long used these indicators (in combination with information from other sources) to determine the Julian dates of the events discussed. The battle of Marathon, for example, is connected by Herodotus to the phases of the moon and to a Spartan festival in such a way that it can be dated to ca. September of 490. Xerxes's invasion of Greece, on the other hand, can be dated to the spring and summer of 480. 6 The affairs which Herodotus mentions in-between can be dated in relation to those two events. This creates the following picture. Herodotus says that the battle of Marathon (September 490) was followed by three years of military preparations, and that the Egyptian revolt started in the fourth year. This yearand, with it, the Egyptian revoltmust therefore have begun in 487. Darius died in the year after that, Herodotus says, which must have been a year that began in 486. Xerxes sent an army to Egypt in the second or next year after that, which must have been a year that began in 485. 7 And Egypt's defeat was followed by four years of military preparations, with Xerxes's campaign against Greece happening in the fifth yearwhich must have been a year that began in 481 or 480 (depending on whether the preparations started directly after Egypt's defeat, or only in the year after that). 8 No matter when that fifth year be-6 Some have dated the battle of Marathon to August 490, but that doesn't change the overall chronology that much; see Olson, et al. (2004). For the date of Xerxes's invasion of Greece, see Macan (1908b), 398-401. 7 The word which Herodotus uses in 7.7 to qualify the year in which Xerxes sent an army to Egypt, i. e. δεύτερος, does not necessarily mean "second" (pace the interpretations of Pestman 1984, 147, andRottpeter 2007, 14). Δεύτερος can also mean "next," which is how Herodotus commonly uses it; see Powell (1977), 82. 8 Miller (1959), 40, and Rhodes (2003, 72, for example, place the start of the four years of military preparations in the same year as the defeat of Egypt, i. e. in 485/484; while both Strasburger (1965), 724-725 andHow andWells (1912), 133 think the four years only started with the year following gan and ended, though, the campaign resulted in the occupation of Athens during the summer of 480.
Based on this chronology, we can conclude that the dates of the Egyptian revolt are ca. 487-485 (at least, if we believe the Histories). We could leave the discussion of Herodotus's chronology at that. It is better to go a step further, however, and ask in which exact monthsrather than years -Herodotus placed the Egyptian revolt. Such precision will help us later on, when we try to assess whether several dated Egyptian sources fell before, within or afterand therefore contradict or comply with -Herodotus's chronology for the event.
Herodotus's "year." To know in which months Herodotus placed the Egyptian rebellion, we need to know what a "year" was for the historian. This is less simple than one might initially assume. First of all, when we speak of years, we generally refer to year periods (i. e. random periods of about twelve months) or to years within a specific calendar system (e. g. a year which runs from 1 January to 31 December). We can say, for example, that an issue occurred in November 2016 and that it was solved the next (calendar) year (i. e. at some point after 1 January 2017), or that it was solved one one year (period) later (i. e. in November 2017). The same is true of Herodotus. But the problem is the following: we do not know which calendar system Herodotus used, nor when he referred to calendar years or to year periods in his narrative. The only thing modern scholars can do is read the Histories closely and come up with an interpretation of Herodotus's "year" which would fit his chronology best (as well as the chronological information we have from other sources).
Close reading of the Histories has resulted in roughly four different interpretations of Herodotus's "year:" most Classicists think that Herodotus's "year" was a campaign year, which ran from spring to spring; 9 some think that it was a Persian regnal year, which began in March or April; 10 some think that it was an Athenian archon year, which began in ca. June; 11 while others think that Herodotus used any of these campaign or calendar years and combined them with year periods at Egypt's defeat, i. e. in 484/483. For the different arguments in favor of either chronology, I refer the reader to the relevant publications. 9 E.g. Busolt (1895), 537-538 n. 3; How and Wells (1912), 79, 128, 133;Strasburger (1965), 698 n. 31; Scott (2005), 457 n. 1; Stadter (2012), 44-45. 10 For an elaborate article in favor of Persian regnal years, see Miller (1959). 11 For an elaborate article in favor of Athenian archon years, see Hammond (1955). Rhodes's (2003), 60 interpretation of Herodotus's chronology falls somewhere in between Miller's and Hammond's: he thinks that Herodotus used a Persian series of dates, but that those dates were subsequently worked out "in terms (...) of years beginning in the summer." some points in his narrative. 12 Each interpretation yields different results for the chronology of Herodotus's historyincluding the chronology of the Egyptian revolt. Let me illustrate this in a bit more detail.
When Herodotus writes that Darius began three years of military preparations after the battle of Marathon (ca. September 490), we could interpret this as three year periods. The fourth year, in which Egypt rebelled, would then have started in ca. September 487. But it is also possible that Herodotus referred to three calendar or campaign years instead of three year periods here. If he referred to the Athenian archon calendar, for example, the fourth year after the battle of Marathon would have started in ca. June 487; and if he referred to a campaign year, the fourth year would have started in spring 487 (whereby, in each case, the "first" year would be the year in which the battle of Marathon actually happened). 13 The same problem applies to the end of the Egyptian revolt: cuneiform sources indicate that Darius died at the end of November 486; and Herodotus says that Xerxes sent an army to Egypt in the second or next year after Darius's death. 14 It is possible that we should interpret this as a literal second year period after Darius's death, which would mean that the army went to Egypt at some point after November 485 BC. 15 But it is also possible that Herodotus placed the re-conquest of Egypt in the next calendar or campaign year after Darius's death. This means that, if a year for Herodotus was a Persian regnal year, Xerxes's army could have gone to Egypt as early as 6 April 485again several months earlier than an interpretation of year periods would have us believe.
None of the interpretations of Herodotus's "year" can be proven beyond reasonable doubt, so multiple chronological possibilities remain. It should be emphasized, however, that the possibilities are not endless. One can reasonably say that Herodotus dates the Egyptian rebellion to somewhere between March 487 and June 484, which are the outer-limits of the different chronological interpretations combined (see Table 1). 16 One could go a step further and qualify the outer-12 See e. g. Macan who thinks that Herodotus's years generally reflect campaign years (1908b, 403-404), but that the years of military preparation under Darius and Xerxes are year periods (1908a, 2 and 29). As for the second/next year after Darius's death: Macan considers that both an interpretation of the passage as the next (calendar) year and as the second year (period) are possible (ibid., 8-9). 13 Herodotus counted inclusively, which means that the year (or month or day) in which something happened was the first year (or month or day), and the one following it was the second; see Hammond (1955), 383 n. 2. 14 For the interpretation of δεύτερος as either "second" or "next" see supra, n. 7. 15 As claimed by Pestman (1984), 147. 16 Depending on which "year" one thinks Herodotus was using, the chronological parameters should be shortened. most parameters of the beginning and end of the revolt as well: the beginning fell somewhere between March 487 and September 486; and the end somewhere between March 485 and June 484. We are therefore left with a period of at least seven months (September 486 -March 485) and at most three years and four months (March 487 -June 484), somewhere in which a revolt of unknown length must be placed.  Depuydt 1995, 199 n. 34). Therefore, if one wants to maintain Herodotus's four full years of preparation (understood as either year periods or calendar years), the latest date for the end of the Egyptian rebellion would be June 484. The fifth year could then have started in June 480, which coincides with Xerxes's crossing of the Hellespont (for the latter's date, see Hammond 1955, 383-384).

II The Egyptian termini post and ante quem
As mentioned above, Egyptologists have long had a different interpretation of the rebellion's chronology: the start of the event is often placed in 486 or 486/485, rather than 487 or 487/486. This divergence can be explained by two things: first, while Egyptologists have likewise based their dates on Herodotus's chronology, they have done so in a more approximate manner; and second, the date-formulae of two Egyptian texts have been used as strict termini post and ante quem for the episode. The combination of these two elements has resulted in a more narrowand, in my view, fallibletime-span for the rebellion. Let us start with the Egyptian texts. Egyptian documents are often dated to the regnal year of a specific king. When we're lucky, the documents even mention the month and day of writing. Such date-formulae are valuable evidence: they can show us the number of years that a specific monarch ruled, they can indicate a change of rulership, andat the very leastthey can tells us which king was or was not recognized at a specific moment in time. When we study Egypt's second rebellion, therefore, texts that show us the dated recognition of Darius and Xerxes (or the absence of that recognition) in the period 487-484 are important sources of information. What we find is the following: the last Egyptian text from Darius's reignas far as such texts are excavated and publishedis P. Loeb 1, which dates to 5 October 486 (17 Payni of year thirty-six of Darius), 19 while the first Egyptian text from Xerxes's reign is Posener 25, which dates to 9 January 484 (19 Thoth of year two of Xerxes). 20 The first year of Xerxes, on the other hand, is entirely undocumented in Egypt.
Comparing this documentary situation to Herodotus's narrative, Egyptologists have created the following argument: Herodotus says that Egypt was in revolt by the time that Darius died. The exact date of Darius's death is ca. November 486 (a date based on evidence from cuneiform sources). 21 The last text which recognizes Darius's reign in Egypt is dated to 5 October 486. Therefore, the Egyptian rebellion must have begun somewhere after October but before November 486 BC. Herodotus also says that Xerxes sent an army to Egypt in the second year after Darius's death. As the first Egyptian text which recognizes Xerxes's reign dates to 9 January 484, the rebellion must have ended between November 485 (the start of 19 See Martin (1996), 296-297. 20 See Posener (1936, 120. 21 For Darius's date of death, see Zawadzki (1992).
the second year after Darius's death) and January 484. Based on this chronology, the rebellion must have lasted at least a year and one month (from before November 486 to after November 485), and at most one year and three months (from after 5 October 486 to before 9 January 484)a period of time in which zero Egyptian texts are dated to Persian kings. Roughly speaking, this would date the start of the revolt to the end of 486 and its defeat to the turn of 485/484. 22 The argument that the Egyptian rebellion should be dated to 486-485/484 has been widely adopted in Egyptology and studies of the Achaemenid Empire. 23 The claim that the rebellion began at or even after Darius's death in November 486 must have been the result of this argument as well: after all, the argument has confined the revolt's possible beginnings to a narrow window of time at the very end of Darius's reign (i. e. to October -November 486). Unfortunately, the 486-485/484 date has become so accepted that modern scholars appear to be unaware of its foundation: some studies which mention the revolt seem to attribute the dates directly to Herodotus, rather than to a chronological study. 24 This creates the impression that the dates of the Egyptian sources, which were used to establish the termini post and ante quem of the rebellion in the first place, are perfectly compatible with Herodotus's text. They are not, however. It may be helpful to revisit Herodotus's chronology of the revolt; and to highlight where his chronology contradicts the chronology created by Egyptologists.
Herodotus vs. the Egyptian sources. Let us start with the end of the revolt: Egyptologists have set the terminus ante quem for the revolt's defeat at 9 January 484, based on the date of Posener 25. The terminus post quem for the revolt's defeat (namely November 485) is not based on an Egyptian source, however. The date is only founded on Herodotus 7.7 in which Xerxes is said to have re-invaded Egypt in the "second" year after Darius's death. The Egyptological argument 22 The general argument can be found in Cruz-Uribe (1980), 37-38; but it was most elaborately formulated by Pestman (1984). It is important to note, however, that these scholars used a slightly earlier terminus post quem for the event: while they agreed that the rebellion only started after the date of P. Loeb 1, they dated that papyrus to 7 June rather than 5 October 486. The latter date for P. Loeb 1 is more current now: it was first proposed by Hughes and later followed by e. g. Cook (1983), 99 and Martin (1996), 297 n. 5. The demotic signs for the months involved are very similar, hence the different interpretations; but an October-date is the more likely reading (personal communication with Cary Martin and Joachim Quack, June 2018; I want to thank both scholars for taking the time to answer my questions about the demotic signs). 23 See supra, n. 3 and n. 4. The exceptions to the rule are Ruzicka (2012), 27 and Yoyotte (2013), 257, both of whom date the start of the revolt to 487 instead of 486. 24 E.g. Briant (2002), 161, 525;Kuhrt (2007), 236, no. 59 n. 4, 248, no. 6 n. 2;Rottpeter (2007), 14-17. takes this passage quite literally to mean that, if Darius died in November 486, the second year after his death would have started in November 485. That this is a misreading of the Greek text, and a too simplistic interpretation of Herodotus's chronology, was already shown above: first, the word for "second" can also mean "next" in Herodotus's narrative; second, it is not clear whether Herodotus is talking about year periods or calendar years here. 25 So, in the absence of further Egyptian sources, we are left with Herodotus's vague terminus post quem for the rebellion's defeat: it may have ended anywhere after March 485.
As for the revolt's starting period: it is here that we run into a more significant contradiction between the Egyptological reconstruction and Herodotus's chronology. Herodotus, as was argued above, placed the start of the Egyptian revolt in a year before Darius's death, more precisely between March 487 and September 486. The Egyptian sources, however, seem to indicate that the revolt began near Darius's death: if P. Loeb 1 is our terminus post quem for the revolt's beginning, then it must have started after 5 October 486, and before Darius's death in the following month. This date falls explicitly after Herodotus's outer limit for the start of the rebellion (i. e. September 486). 26 We are therefore faced with two options: either Herodotus's chronology is incorrect, and Egypt still recognized Darius at the end of his reign; or our interpretation of Herodotus's chronology is incorrect, and it should be amended to incorporate October 486. The first option does not require further elaboration. We can look at option two in a bit more detail, though.
It is possible to interpret Herodotus's chronology in such a way that it incorporates October 486 within its starting-period for the revolt. One could argue, for example, that Herodotus placed the start of the three years of military demands after the battle of Marathon in October 490, rather than in September. The fourth year after the battle would then have started in October 487; and the year would have ended in October 486 BC. P. Loeb 1 was written at the start of the latter month, so the revolt could have begun on a later day that month and still have fallen within Herodotus's chronology for the event. This requires us to assume the following: some time would have elapsed (in Herodotus's mind) between the end of the battle at Marathon and the moment that Darius heard about the Persian defeat, and/or some time would have elapsed between the moment that Darius heard the news and the period of military demands. It also requires us to assume that the years which Herodotus referred to were year periods, and not campaign or calendar years. All of this is quite reasonable.
There is one problem, however. If we want to place the start of the Egyptian revolt in the last days of October 486, and consider Herodotus's chronology as correct, we are forced to assume that the quarrel between Xerxes and Artobazanes (which Herodotus 7.2 claims happened at the same time as Darius's plans for an invasion of both Egypt and Athens) happened at the end of October or beginning of November 486. This also requires us to assume that Darius nominated Xerxes as his heir apparent quite quickly after the quarrel began, because Darius died in November of that year. And it requires us to assume that the "year" in which Darius died, which was the year following Egypt's revolt and Xerxes's nomination (as Herodotus 7.4 describes it), was a hypothetical fifth year of military demands. This year could then have started in October/November 486, and be narrowly chronologically compliant with Darius's death in November. Such an interpretation of Herodotus's text is not impossible. It does feel, however, like a stretch of the actual narrative.
If an emendation of Herodotus's chronology does not look like the best solution to explain the date in P. Loeb 1, we are faced with the following dilemma: either we accept Herodotus's chronology for the rebellion, and dismiss the date in P. Loeb 1 as an anomaly, or we take the date in P. Loeb 1 seriously as a terminus post quem, and dismiss Herodotus's chronology as incorrect. Most historians of antiquityor, at least, most Egyptologistswould find this an easy choice: the evidence of contemporary and native sources trumps that of a non-Egyptian narrative which was written decades after the event. Hence, P. Loeb 1 should take primacy over Herodotus, and the terminus post quem of the revolt really is October 486. Such an interpretation of the Egyptian sources, however, has its own particular problems.
The problem of the Egyptian sources. When we interpret the dates of P. Loeb 1 and Posener 25 as termini post and ante quem for the Egyptian rebellion, it is important to emphasize that we are operating under the following assumption: if the writers of P. Loeb 1 and Posener 25 recognized Persian kings at a specific moment in time, so the assumption goes, then the rest of the Egyptian population at that time must have recognized Persian kings as well. This is, in fact, quite a far-reaching generalization.
First of all, we should remind ourselves that the textual corpus of Persian Period Egypt is limited: the history of large parts of the country remains sparsely documented or even entirely undocumented. Second, we should remind ourselves that this limited corpus precludes knowledge of where the revolt began or how it may have progressed. It is possible that the rebellion began in the Delta, for example, though there are no documents from the region which can prove that hypothesis. 27 The rebellion could then have spread to other parts of Egypt in the months after its beginning, but we do not know how much time that would have taken. And the rebellion may eventually have affected the whole of Egypt, but we are not sure whether it actually did. The interpretation of our small handful of date-formulae is therefore seriously hampered: all they can show us is which king was recognized in a specific locality at a particular point in time; but they can never show us which king was recognized in the rest of Egypt. 28 To conclude, I would posit that the date-formulae of P. Loeb 1 and Posener 25 do not prove that the Persians controlled the whole of Egypt in October 486 and January 484. Rebels may already or still have been active in other parts of the country. 29 This conclusion has one important consequence: if P. Loeb 1 and Pos-27 The only tentative indication of the rebellion's affect in northern Egypt is a stopped-up well, excavated at Tell el-Maskhuta; see Holladay (1982), 25-26. The well does not indicate that the rebellion actually began in the north, however. This fact requires some emphasis, as many scholars have nevertheless linked Egypt's second revolt to the Delta and/or Libya; see e. g. Kienitz (1953), 67-68;Rottpeter (2007), 24-25;Ruzicka (2012), 27-28;Yoyotte (2013), 257. The link is partly made in analogy with the third and fourth Egyptian rebellions, which were both connected to those places (see Kienitz 1953, 67-68). It is dangerous to assume that all rebellions would have been the same, however. The link is also made because the name of the rebel-king in ca. 486 was "Psamtik" (known as "Psamtik IV"; see infra for the sources of his reign). One might connect this Psamtik to Inaros II's father of the same name, Inaros being a Libyan man who led most of Egypt in revolt against Artaxerxes I (see Cruz-Uribe 1980, 39 who notes the uncertainty of the connection); or one might use the probably Libyan origin of the name "Psamtik" as evidence for the Libyan ethnicity of its bearer (which is done by Rottpeter 2007, 24-25). Neither arguments are convincing, however. After three kings of the Saite Dynasty had "Psamtik" as their birth-name (see Leprohon 2013, 164-167), many Egyptians gave the name (or a variant thereof) to themselves or their children (see e. g. the name indices in Vittmann 1978, 225;Pressl 1998, 324-325;Chevereau 2001, 377-378; and the study on "beautiful names" by De Meulenaere 1966). It is therefore possible that Psamtik IV was simply an Egyptian man who was born with that name; or that both Psamtik IV and Inaros used the name to stress their connection to Egypt's "glorious" Saite past. Whatever the case, there is not enough evidence to link the origins of the rebellion and/or of Psamtik IV to the Delta and/or Libya; and the possibility that the rebellion stemmed from, for example, Upper Egypt must be reckoned with. 28 The specific social context of P. Loeb 1 and Posener 25 makes them even less suited for generalization; see infra for a discussion of this issue. 29 The observation that P. Loeb 1 and Posener 25 may have coincided with (rather than pre-or postdated) the revolt is not new. Kienitz (1953), 67-78 assumed, for example, that Egypt's second rebellion only affected the Delta; and he used P. Loeb 1 and Posener 25 as evidence for continued Persian control in the south of Egypt. Pestman (1984), 147 likewise considered the possibility that P. Loeb 1 and Posener 25 were contemporary with the event, rather than being its termini post and ante quem (before rejecting it as the less likely option). Scholars have not considered the scenario and the consequences it has in more detail, however. ener 25 cannot be used to chronologically delineate the Egyptian rebellion, then the only chronological framework we're left with is Herodotus's. Subsequently, because Herodotus's chronology is much wider than the chronology created by Egyptologists, the number of Egyptian sources that might be linked to the rebellion widens as well. These sources deserve a closer look.

III The Egyptian sources: a social divide
At the risk of sounding repetitive: Herodotus's chronology allows that the Egyptian revolt began as early as March 487 (in year thirty-five of Darius), and that it lasted until June 484 (in year two of Xerxes). Egyptian texts dated to 487-484 are shown in Figure 1 for easy reference. There are ca. nineteen of them in total. 30 Our current goal is to scrutinize all of these texts, and to see whether any of them (and not just P. Loeb 1 and Posener 25) can inform us about the Egyptian rebellion. Now, when one looks at the texts in Figure 1, four things can be observed immediately: 1) sixteen of the nineteen texts fall within Herodotus's chronology for the revolt (including P. Loeb 1 and Posener 25), which means that all sixteen of them could have coincided with the troubles; 2) all nineteen texts can be divided into six different textual corpora, based on their provenance and contents; 3) five of these corpora stem from Upper Egypt, which means that our understanding of the events in 487-484 remains largely limited to the southern Nile Valley; 31 and 4) three of the six corpora end in or shortly after year thirty-five of Darius, while the other three continue into the reign of Xerxes (and beyond).
All of these observations are important and interrelated. My focus in the following pages will be on element number four, however. I will argue that the three corpora that "survived" the revolt (i. e. those whose documentation continued after the rebellion had been defeated) belonged to a specific social stratum in Egyptian society, while the three corpora that ended near or in the period of revolt belonged to a different one. More importantly, I will argue that these social differ-30 There are some additional Egyptian texts that might be dated to 487-484; however, as their date-formulae are imperfectly preserved, they have been excluded from Figure 1. References to such texts can be found in the footnotes of the following section. Note that one can search for dated Egyptian texts in Trismegistos: Trismegistos has a near 100 % coverage of all demotic sources and a ca. 95 % coverage of all Aramaic sources; see https://www.trismegistos.org/calendar/calendar_ search.php (accessed 18-09-2018) and https://www.trismegistos.org/about_coverage.php (accessed 19-09-2018). 31 The only corpus which cannot be directly linked to Upper Egypt consists of objects of unknown provenance or with a provenance outside of Egypt; see infra.
ences overlapped with differing political loyalties at the end of Darius's reign. How this observation affects our understanding of the revolt's chronology will be revisited at the end of this section. Let us start, however, with a look at the social realities behind the texts. The corpora that survived the revolt. The three textual corpora that "survived" the Egyptian revolt consist of papyri from Elephantine, rock inscriptions from the Wadi Hammamat, and several inscribed vases which were found outside of Egypt. This may sound like a varied group of texts. But they have more in common than one might think. All three corpora, for example, contain texts that are dated to year thirty-six of Darius and year two of Xerxes (though most without the month or day preserved). More importantly, all of the corpora were tied, in one way or another, to the Persian imperial administration of Egypt. This is clear from even the most cursory glance at the contents of the texts.
32 The light grey background indicates the period of time in which the Egyptian revolt may have begun and ended; the dark grey stroke indicates the period of time in which the revolt must have been in progress (both according to Herodotus). Note that the date of P. Hou 8, 7, and 4 is uncertain: they could be placed in year one of Xerxes as well (see infra).
First of all, Posener 24 and Posener 25 were written in the Wadi Hammamat, far away in Egypt's Eastern Desert. Their author was a Persian called Atiyawahi, son of Artames and Qandjou. Atiyawahi inscribed about eight rock inscriptions from year twenty-six of Darius until year thirteen of Xerxes. It is unfortunate that none of them mention the purpose of Atiyawahi's visits: he may have been in the Wadi Hammamat for mining expeditions, patrol duty and/or travel to and from the Red Sea coast. Whatever the purpose of this travels, though, it is clear that Atiyawahi had a relatively high position in the Persian administration of Egypt: his inscriptions clearly state that he was governor of Coptos and an official of Persia at the time. His younger brother, Ariyawrata, later inherited his posts. 33 Secondly, P. Berlin 13582 (Pharmouthi of year thirty-five of Darius), P. Loeb 1 (17 Payni of year thirty-six of Darius) and P. Berlin 23107 (year two of Xerxes) were all found at the island of Elephantine, located at Egypt's most southern border. Elephantine housed a well-known community of foreign mercenaries in the Persian Period, and the island produced a large corpus of mostly Aramaic but also demotic textsa corpus which lasted at least until ca. 400. 34 The three aforementioned texts were all part of this corpus (even though they may have been written by different people). Whereas the Aramaic P. Berlin 23107 is too fragmentary to understand correctly, it is clear that the demotic P. Berlin 13582 and P. Loeb 1 were both related to a man called Parnu. The former concerns a receipt for a payment made to the collection-box of Parnu, while the latter is a letter written to Parnu by an inferior of his. Now, like Atiyawahi, Parnu was a Persian who enjoyed a relatively high position in the administration of Egypt: the Elephantine documents indicate that he was commander of the fortress at Syene (opposite Elephantine) and governor of Tshetres, Egypt's most southern province. 35 Third and finally, BLMJ 1979 (year thirty-six of Darius), Posener 43 and Posener 44 (both year two of Xerxes) all concern brief inscriptions on vases. The in-33 For editions of all of their inscriptions, see Posener (1936), 117-130 andGoyon (1957), 118-120. 34 The corpus of Aramaic texts seems to have ended in ca. 400, i. e. about the time that Egypt became independent from Persian rule. With the end of this corpus, our knowledge of the foreign mercenary community at Elephantine ends as well; see Porten (1968), 301. We do have a sizeable corpus of demotic texts which can be dated to the fourth century BC and later, though. Unfortunately, many of these texts remain unpublished; see Müller (2016), 225, 227-231 for further references. 35 For P. Berlin 13582 and P. Loeb 1, see Martin (1996), 296-297, 374-375; for P. Berlin 23107 see Porten and Yardeni (1999), 56. Note that another Aramaic papyrus from Elephantine might be from Thoth of year thirty-five of Darius (Porten 1996, 257-258). However, as the date of writing is not explicitly preserved, the document has been excluded from Figure 1. scriptions merely record the regnal year, and the name and titles of the Persian king. 36 It is nevertheless obvious that the objects were tied to the Persian imperial administration: the objects belong to a much larger corpus of vases which was found all over the empire; many of these vases were inscribed in both hieroglyphs and cuneiform; 37 all of them allude to the reigns of Darius, Xerxes or Artaxerxes; and a lot of them have been found in the Iranian heartland. Posener 43 and Posener 44, for example, were found in the palace of Susa. Where and by whom these vases were made exactly (e. g. were they made and/or inscribed in Egypt, or by Egyptian artisans in some other part of the empire?) remains obscure. 38 But it has been argued convincingly that the vases were related to Egyptian tribute payments (which would explain their presence in the imperial capitals), and/or to royal gifts (which would explain the presence of a vase in e. g. the famous Mausoleum of Hallicarnassus, or the one found in a burial tumulus in southern Russia). 39 The question that remains is what these texts can tell us about the rebellion in Egypt. The least we can get from them, of course, is that some people in Egypt recognized Darius's reign in his very last regnal year (i. e. between 23 December 487 and 22 December 486); and that the same peopleor, at least, people in the same milieurecognized Xerxes's reign from at least his second regnal year onwards (i. e. between 22 December 485 and 21 December 484). What we cannot get 36 For BLMJ 1979, see Westenholz and Stolper (2002), 1-5; for Posener 43 and 44, see Posener (1936), 141. Note that there might be a third vase which was dated to year two of Xerxes (Qaheri 2012, 325-326). However, as the king's name is not preserved, the vase has been excluded from Figure 1. 37 Both Posener 43 and BLMJ 1979 contain cuneiform. It is unclear whether Posener 44 would have contained it as well, as it is apparently a mere "fragment de vase;" see Posener (1936), 141;Westenholz and Stolper (2002), 1-5. 38 Westenholz and Stolper (2002), 11 note that "[t]here is a consensus that these vessels were made in Egypt, and that the inscriptions were carved in Egypt, the only part of the Achaemenid empire where hieroglyphic inscriptions and Egyptian versions of multilinguals had any meaningful use." This statement does not take into account, however, that most of these vases were found outside of Egypt, that the famous statue of Dariusinscribed in both hieroglyphs and cuneiformwas found at Susa (Yoyotte 2013), and that we find sealings with hieroglyphs on tablets of the Persepolis Fortification archive (Garrison and Ritner 2010). I therefore strongly contest the claim that hieroglyphs had no "meaningful use" outside of Achaemenid Egypt. Moreover, as both the aforementioned archive and the Persepolis Treasury archive actually show the presence of large numbers of Egyptian artisans in the Achaemenid heartland at the end of Darius's reign (see Henkelman 2017), I remain doubtful about the vases' original place of creation and inscription. 39 See Westenholz and Stolper (2002), 5-13 for a general discussion of the Achaemenid inscribed vases.
Egypt's Second Revolt from them, however, is when the rebellion would have begun and ended exactly: based on the texts' contents, it is likely that the people who wrote the texts would have remained loyal to the Persian regime as long as they couldeven though other Egyptians may already have revolted. The Wadi Hammamat inscriptions show, for example, that Atiyawahi maintained his high position in the Persian administration of Egypt after the revolt had been defeated. If Atiyawahi had supported the rebellion, surely he would have been replaced? Another sign of continued loyalty is the date-formula of P. Loeb 1: as we have seen above, P. Loeb 1 was probably written after the rebellion had already begun, and yet the date still referred to Darius's reign. It is therefore safe to say that Parnu and the people who were subordinate to him recognized Darius's reign despite the fact that some contemporaries had already started to fight against him. 40 The corpora that ended near or during the revolt. The three textual corpora that ended near or during the revolt are quite different from those that "survived" the event. All three corpora consist of larger groups of demotic papyri, for examplegroups which I will henceforth call "archives." The archive from Hermopolis is small and contains only six contracts, four of which were written in year thirtyfive of Darius; 41 the archive from Hou is a bit larger and contains thirteen texts, the earliest of which is dated to Darius's twenty-fifth year; and the archive from Thebes is the largest and contains ca. nineteen documents, the earliest of which is dated to Amasis's fifteenth year. 42 More importantly, none of these archives 40 Interestingly, P. Loeb 1 may even refer to the unrest: the letter talks about a shipment of grain which ran the risk of falling into the hands of "rebels." "We are used to seeing the rebels when they are on the mountain on the southern side opposite us," wrote the author of P. Loeb 1. There was a long distance between him and them. But "if [the grain] is brought down, without armed men to guard this grain" then the rebels would come for it by night, and steal the whole lot. The quote is taken from Martin (1996), 297. However, where Martin translates "brigands" I translate "rebels." The difference depends on one's understanding of the phrase rmt.w nty bks: the phrase is traditionally translated as "the men who rebel," but it is often interpreted in P. Loeb 1 as the less politically charged "brigands." This interpretation depends on the idea that "[t]here is nothing in the letter to suggest a civil uprising" (Hughes 1984, 85). See also Spiegelberg (1931), 5;Kienitz (1953), 67-68 n. 8; Briant (1988), 142-143;Martin (1996), 297 n. 8. However, if P. Loeb 1 was indeed contemporary with the revolt, we may have to take the scribe's choice of words more seriously. 41 Only two of these have been included in Figure 1, as the others do not contain the month or day of writing; see Farid (2002), 191. 42 I use "archives" here as a short-hand for groups of documents which show some internal cohesion, and which were either found together or are likely to have been found together. It should be emphasized, however, that 1) the six documents from Hermopolis may have been written on only one or three papyri (Farid 2002, 186); 2) the texts from Hou look like an oddly disparate lot of documents, if it weren't for the scribes and witnesses that held them all together (Vleeming 1991, 6-7); were connected to the Persian imperial administrationin marked contrast with the vases, Elephantine texts and Wadi Hammamat inscriptions discussed above. The archives only show us people with homogenously Egyptian names, people whose lives revolved around geese, mummies, and real estate, and people who primarily interacted (as far as the documents tell us) with other Egyptians.
For the exact contents of all of these texts, I refer to the reader to the relevant publications. The issue that I want to elaborate on here, however, is this: none of the three archives that ended near or during the revolt contain documents dated to year thirty-six of Darius (or to any years of Xerxes). Year thirty-five of Darius is the last Persian regnal year that we find in all of them. We may therefore ask ourselves the following question: are we sure that these people continued to recognize Persian kings in 486-484, like their contemporaries in e. g. Elephantine did? The Egyptian sources, in fact, suggest that they did not.
Psamtik IV and the rebellion in Hou. So far, our discussion of the Egyptian rebellion has been focused on Herodotus's narrative on the one hand, and on several Egyptian texts dated to Darius and Xerxes on the other. One may be excused to think that the Egyptian sources do not provide us with any positive evidence about the rebellion. They do, however: three Egyptian papyri, all of them part of the Hou archive mentioned above, are dated to the second regnal year of a king called Psamtik. While these papyri used to be attributed to Psamtik III, the last short-lived king of the Saite dynasty (527-526), it was argued long ago that the Psamtik in question had to be a rebel-king. Unfortunately, the Hou papyri have received little attention in modern scholarship; so I hope the reader will forgive me for a short digression on their contents. 43 and 3) the majority of Theban texts stem from the (modernly reconstructed) archive of the choachyte Tsenhor (Pestman 1994), but several other choachyte texts belong to the same milieu; it is unclear whether they should be separated from or connected to each other. Among these texts are P. Cattle 13 (Cruz-Uribe 1985, 25-30), and P. Berlin 3079 (Spiegelberg 1902, 5). Note that the latter text might also be dated to year thirty-five of Darius; but as the regnal year is disputed (Thissen 1980, 116 n. 19), it has been excluded from Figure 1. 43 It is unfortunate, for example, that Briant (1988), whichamong other thingsgives a survey of the Egyptian revolts, does not refer to the Hou papyri. The only Egyptian source which is given for Egypt's second revolt is P. Loeb 1 (which is, incidentally, mistakenly dated to 487 in the article). The conclusion is that Egypt's second revolt was not really a "revolt" at all, but mere local trouble (see ibid., 141). One does find a reference to Psamtik IV in Kuhrt's (2007) important sourcebook for the Achaemenid Empire. She refers, however, to an older article by Cruz-Uribe (1980), who first suggested that a Psamtik IV may have existed; and she decides to follow Spalinger's (1982a) doubts about Cruz-Uribe's suggestion. The later study of Pestman (1984), which reinforced the hypothesis of Psamtik IV's existence (see infra), is omitted here as well (Kuhrt 2007, 248, no. 6 n. 2). Finally, The demotic papyri from Hou probably reached the antiquities market of Luxor in the 1890 s. 44 They were bought by Wilhelm Spiegelberg at different moments in time and for different institutions: some of them ended up in the papyrus collection of the University and State Library of Strasbourg, while the others eventually arrived at the Egyptological institute of Munich. Despite this division, it was evident upon publication of the papyri that they were probably found together. 45 Most of them concerned "gooseherds of the Domain of Amun," for example. More importantly, there were several individuals who appeared in multiple documents from both lots of papyri. Pestman was the first who, in 1984, published a full overview of these prosopographical interconnections. His study resulted in the following, crucial observations. The three papyri from Psamtik's second regnal year (later known as P. Hou 8, 7 and 4) were more closely tied to the papyri from Darius's reign than previously thought. One man, for example, acted as a witness in P. Hou 7 and P. Hou 4 (both year two of Psamtik), but also as a witness in P. Hou 6 and P. Hou 5 (both from Darius's reign; the former from an unknown year, the latter from year twentyfive). Another man was the scribe of P. Hou 7 (year two of Psamtik), but also of P. Hou 13 and P. Hou 12 (both from year thirty-five of Darius). And a third man featured as Party A in P. Hou 4 (year two of Psamtik) and also as Party A in P. Hou 3 (year thirty-five of Darius). Pestman observed that if year two of Psamtik referred to Psamtik III, as previously assumed, two of these men would have had an uncommon career-span of ca. forty years (526-487). Moreover, the Hou archive would then have displayed a documentary gap of at least twenty-eight years, as the earliest papyrus from Darius's reign stemmed from his twenty-fifth year (P. Hou 5; see Table 2). Pestman therefore argued that P. Hou 8, 7 and 4 were unlikely to have been written in 526; and he re-dated the papyri to the end of Darius's reign instead. Psamtik was subsequently called "Psamtik IV"a rebel-king in ca. 486. 46 Rottpeter's (2007) study of the Egyptian revolts does include a reference to Psamtik IV and to Pestman (1984). One does not find the reference where one would expect it thoughnamely, in Rottpeter's discussion of the Egyptian sources for Egypt's second revolt, which includes the Wadi Hammamat inscriptions and P. Loeb 1 (ibid., 15-16). Instead, Psamtik IV is merely mentioned in a later footnote, where Rottpeter doubts the rebel-king's significance because Herodotus did not mention his name (ibid., 24-25 n. 37). Nothing else is said about him or about the Hou papyri. 44 That they were found in Hou is indicated by the documents themselves; see Vleeming (1991), 1-2. 45 See e. g. the brief note in Spiegelberg (1931), 70. For a summary of the papyri's acquisition and research history, see Pestman (1984), 145-146. 46 See Pestman (1984. Note that Pestman built his hypothesis on Cruz-Uribe (1980), who first suggested that a document from Hou may have been dated to a "Psamtik IV." Do we know anything else about this "Psamtik IV"? Unfortunately notnot with certainty, anyway. What we do know, however, is that many of the rebelkings of Persian Period Egypt associated themselves with the name "Psamtik." A rebel-king of the 460 s is said to have had a father called "Psamtik," for example (though we do not know whether this Psamtik was a rebel-king himself). 47 It is also said that a "Psamtik," king of Egypt, sent grain to Athens in 445/444. 48 And we know that a "Psamtik," king of Egypt, held power in ca. 400. 49 On top of that, there are three Egyptian objects that bear the cartouches of two different Psamtiks -Psamtiks who cannot be identified with any of their Saite Period namesakes. One is a scarab which refers to a Psamtik Nebkaenra, and the others are a sistrum-handle and a naophorous statue which refer to a Psamtik Amasis (son-of-Neith). 50 The objects may have referred to any of the aforementioned Psamtiks; or to Psamtiks whose existence we are otherwise unaware of.
That there are Psamtiks galore in Saite and Persian Period Egyptknown both from Greek and from Egyptian sourcesrenders Pestman's hypothesis even more plausible: there is no obvious reason why the Psamtik of the Hou papyri has to be identified with Psamtik III rather than with any of the other Psamtiks that existed. The decisive factors in the identification therefore have to be the contents of the papyri themselves, and the coherence of the archive in which they were found. These factors, as shown above, point to a date at the end of Darius's reign. Note that, theoretically, P. Hou 8, 7 and 4 could be dated to a year in-between year twenty-five and year thirty-four of Darius, as those years are undocumented in the Hou archive. However, as the only rebellion known to have existed in Darius's later reign is the one mentioned by Herodotus (and by Aristotle; and possibly by the Daiva inscription), a date for Psamtik IV in ca. 486 remains the most likely hypothesis. 51 47 The claim that the Libyan rebel-king Inaros was the son of a man called Psamtik can be found in Hdt 7.7, Thuc 1.104 and in a Greek inscription from Samos (see Dunst 1972, 153-155). 48 See Philoch. FGrH 128 F 119 and Plut. Per. 37. The claim that this Psamtik was a "Libyan dynast" is not supported by these passages, pace Spalinger (1982b). 49 Diod. Sic. 14.35 used to be the only reference to this Psamtik. We now know, however, that at the end of the fifth century BC a pharaoh called Psamtik was recognized in the demotic ostraca from Ayn Manawir. He is probably the Psamtik that Diodorus knew of; see Chauveau (1996), 44-47. 50 The objects are listed by Jansen-Winkeln (2014), 583-585, who attributes them to Psamtik III. 51 Pestman (1984), 147-148. Note that Pestman dated Psamtik's first regnal year to 486, and his second to 485. These dates were based on the termini post and ante quem of P. Loeb 1 and Posener 25. Once we reject those termini, however, it becomes possible to date Psamtik's first regnal year to 487 as well; see infra. Once we accept the likelihood of Psamtik IV's existence, what else can be said about his reign? First of all, the date-formulae of P. Hou 8, 7 and 4 enable us to reconstruct the rebellion's chronological progression with a bit more precisionan element which I will revisit below. Secondly, it is important to emphasize that P. Hou 8, 7 and 4 give us the only certain indication of the rebellion's geographical spread in Egypt: we may not know where the rebellion started, but we at least know that it came to affect Hou in the southern Nile Valley. This fact alone begs the question about the rebellion's impact on other towns of Upper Egypt. Might the rebellion have reached Hermopolis and Thebes as well?
The rebellion in Hermopolis and Thebes. The archives from Hermopolis and Thebes do not contain any papyri dated to Psamtik IV (or to any other Psamtik, for that matter). Whether their archive-holders eventually recognized Psamtik IV's reign can therefore not be proven. What can also not be proven, however, is that they continued to recognize Darius's and Xerxes's reigns in 486-484: as noted above, neither the archive from Hermopolis nor the one from Thebes contain any documents dated to year thirty-six of Darius or later. This is similar to what we see in the archive from Hou (and unlike what we see in the corpora that survived the revolt). However, whereas the archive from Hou visibly switched to the regnal years of Psamtik IV at some point after September/October 487 (i. e. after Payni of year thirty-five of Darius), the archives from Hermopolis and Thebes simply ended after February/March and June/July 487 respectively (i. e. after Hathyr and Phamenoth of year thirty-five of Darius). What does that mean, if anything?
That an archive ended generally means that its documents were disposed of in antiquity or that it was (suddenly) abandoned by its keepers. The possible reasons behind such acts are numerous: they range from changing environmental conditions, which may have prompted families to migrate and to leave some of their possessions behind, to the simple fact that old documents could have lost their value over time, and were hence discarded on rubbish dumps. Vleeming, the primary editor of the Hou papyri, (tentatively) favored the latter scenario: "One is (...) bothered by the question, why did someone dispose of all these texts in one go (in one's grave? on a rubbish heap?). We can only tentatively suggest that the texts had fulfilled their function: the donkeys had died, the cows had passed away, the land lease had lost its pertinence, etcetera." 54 The same may have been true for the archives from Thebes and Hou. If so, they cannot tell us much about the Egyptian rebellion.
Chronological analysis. A closer look at the texts that fall within Herodotus's chronology for the rebellion has shown a clear divide in the Egyptian corpus. On the one hand, we have people (both Persians and Egyptians) who were involved with the Persian imperial administration. They seem to have recognized Darius's reign until 486, and to have recognized Xerxes's reign from at least 484 onwards. On the other hand, we have Egyptians who recognized Darius's reign until ca. 487. Some of them suddenly disappear from our view in the middle of that year, while others appear to have recognized a rebel-king for a while, until they too disappear from the record. The question that remains is whether we can synthesize these bits of information into a more coherent chronological narrative.
To repeat: all chronologies of Egypt's second rebellion are ultimately based on Herodotus. We have seen that he dated the rebellion's beginning to some point between March 487 and September 486, and the rebellion's end to some point between March 485 and June 484. It is only when we want to establish the details of the rebellion's chronology more clearly that we start to rely on the Egyptian sources. Now, depending on our interpretation of those sources, one could think of roughly two ways in which to sort the Egyptian texts at our disposal. They are as follows.
1). If we want to take the apparent documentary break of the Egyptian sources in year thirty-five of Darius seriously, then we could place the start of the Egyptian revolt in 487 (more precisely between March and December 487). Assuming that Psamtik IV claimed the Egyptian throne from the very beginning of the rebellion, his (undocumented) first regnal year would have started in 487 as well. This means that the papyri dated to his second regnal year stem from 486; and that his reign was at least contemporary with year thirty-five and year thirty-six of Darius. 58 Therefore, P. Tsenhor 17 (Thebes), P. Hou 3, 2, 17 and 12 (Hou), and P. Berlin  Figure 1 based on this scenario; cf. however the scenario presented below. For those who are not familiar with the Egyptian dating system: "year 1" of an Egyptian king began at the moment of that king's accession to the throne. "Year 2" started at the following New Year's. So, even though Psamtik IV enjoyed a second regnal year, this does not mean that he had already ruled one full year; it only means that he must have ascended the throne before the New Year that preceded the Hou papyri. To specify this more clearly: the Egyptian sources show that Psamtik IV must have ruled at least four months (and need not have ruled more) by the time that P. Hou 4 was written. That's because P. Hou 4 is dated to Tybi of his second regnal year. Theoretically, therefore, Psamtik IV could have ascended the throne on 22 December (i. e. one day before the New Year's day of 487 or 486), and P. Hou 4 could have been written on 22 April (i. e. on the first day of Tybi in 486 or the second day of Tybi in 485), with only a four months' difference between those two dates. This is only the bare minimum, of course. It is, however, more exact than the minimum of "six months" proposed by Pestman (1984), 147. 59 In light of this scenario, it is important to note that the regnal year of P. Hou 8 is difficult to read. Vleeming (1991), 128 settled for a reading of "two," because "2 is the highest known year of Psammetichos IV"; but a reading of "1, 2, or 3 (or 4?)" could be not excluded definitively.
2). If, on the other hand, we want to minimize the amount of (visible) political fragmentation in Egypt, then we could place the start of the Egyptian revolt in 486 (more precisely between January and September 486). Psamtik's first regnal year would then have started in 486 as well, which means that the papyri dated to his second regnal year stem from 485, and that his reign was at least contemporary with year thirty-six of Darius and year one of Xerxes. All texts dated to year thirty-five of Darius, therefore, would have predated the Egyptian rebellion; while Posener 24 (Wadi Hammamat) and BLMJ 1979 (vase of unknown provenance) could have predated or coincided with the rebellion. Only P. Loeb 1 (Elephantine) must certainly have coincided with the revolt. As for the rebellion's end: Psamtik's reign may have been thwarted shortly after the Hou papyri were written, i. e. after April/May 485 (year one of Xerxes). 60 It may, however, have lasted until as late as June of 484 (year two of Xerxes). Theoretically, therefore, Psamtik could have enjoyed a third regnal year in Egypt (ca. 484). P. Berlin 23107 (Elephantine), Posener 25 (Wadi Hammamat), Posener 43 and 44 (vases from Susa) may either have coincided with or postdated his reign. In this scenario, the minimum time-span of the revolt would be ca. eight months (September 486 -April/May 485); the maximum time-span ca. two years and six months (January 486 -June 484).
With the evidence we have at present, neither of these scenarios can be proven beyond reasonable doubt. It is therefore best to maintain the vague dates of 487-484 for Egypt's second revolt, with the important caveat that the rebellion may have affected different parts of Egypt at different moments in time with varying degrees of intensity.

V Conclusion
The present article began and ended with the problem of dating the second Egyptian revolt. We have seen, among other things, that all chronologies of the revolt are ultimately based on Herodotus's Historiesa work which dates the Egyptian revolt to some point between March 487 and June 484. We have also seen that Egyptologists have tried to specify this chronology more clearly: for them, the dates of P. Loeb 1 (5 October 486) and Posener 25 (9 January 484) have become the strict termini post and ante quem for the event. This article has argued, on the other hand, that P. Loeb 1 and Posener 25 were written by people who were tied to 60 The last papyrus dated to Psamtik's reign stemmed from April/May of his second regnal year; see supra, n. 58. the Persian imperial administration of Egypt. It is probable that these people remained loyal to the Persian kings for as long as they could; and it is therefore possible that their texts coincided with, rather than pre-or post-dated, the Egyptian revolt. The downside to this argument is that it rids us of specific dates for the unrest: the only chronological framework we're left with is now Herodotus's. What we gain, however, is an expanded corpus of Egyptian sources that might be related to the troubles.
There are several things we can learn from this expanded corpus of sources. First of all, the corpus draws our attention to a documentary divide: whereas some groups of Egyptian texts "survived" the revolt, others ended near or during the rebellion. We have seen above that this documentary divide overlaps with a social divide: the former groups were tied to the Persian imperial administration (and included texts such as P. Loeb 1 and Posener 25), while the latter were tied to Egyptian communities, some of whom supported a rebel-king. Secondly, the expanded corpus of sources draws our attention more closely to the geographical extent of the revolt. We have seen that the rebellion affected Hou, for example, and that it may have influenced the important city of Thebes (and, possibly, Hermopolis). If it did, then it is plausible that other regions of Upper Egypt would have been affected as wellthough in what way exactly remains an unanswered question. 61 In the end, it can only be hoped that future finds and publications will throw more light on the Egyptian troubles of ca. 487-484. Until then, we will have to be satisfied with our meagre trail of breadcrumbs.