PART I THE LETTERS 1. INTRODUCTION

PART I THE LETTERS 1. INTRODUCTION 1.1. Chronological and Archival Context 1.1.1. Précis of Historical Background As outlined by Ristvet and Weiss in their introduction to this volume, recent research on both archaeological and documentary evidence from Tell Leilan and other sites has produced detailed discussions of the identification of the site and its place in the history of Northern Mesopotamia in the late third to early second millennium B.C., so that few remarks on the historical background are needed. 1 First, it may be useful to reiterate that the identification of Tell Leilan with ancient fieón / fiubat-enlil can be considered definitely established. Any possible doubts left by the analyses presented by Charpin (1987a) and Whiting (1990b) are removed by the 1987 evidence. 2 On the other hand, the problems concerning the relationship between Apum/m t Apim and fieón /fiubat-enlil remain unresolved, and the new evidence provides no firm conclusions on this issue. All that can be said is that Apum, in the texts here, refers to areas near the capital fieón /fiubat-enlil. 3 The name fiubat-enlil is sparingly used in the texts published here, but was almost certainly applied to the town by the mighty fiamí -Adad 1 (ca. 1833 1776 B.C.), whose association with the 1. The following brief remarks summarize information and discussion found especially in publications by Weiss (see Bibliography); Whiting 1990a and 1990b (for the Leilan evidence); Charpin 1986 and 1987a; and Charpin, ARMT XXVI/2, pp. 31ff. (for the Mari evidence). 2. This follows not so much from any single piece of evidence, but from the cumulative weight of corroborative data. To mention but two aspects: numerous administrative documents record transactions as taking place in fieón, and in all seven cases in which fiubat-enlil is mentioned in the letters (see index) the logical contextual implication is that the letter was received by someone residing in this town. 3. As correctly pointed out by Charpin (1987a, 137ff.; and Charpin 1990b, 117 with notes 4 5; for a different etymology of the name Apum, see Cohen 1993, 260), extant Old Babylonian references (including those from the new Leilan material) provide a clear distinction between Apum as the area or country around Leilan and fieón /fiubat-enlil as the actual ancient city, but some problems persist. First, the Old Assyrian references to Apum should be to a town (like other localities along the Old Assyrian routes) and, second, divine compounds like B2let-Apim usually involve the name of a town rather than a land. Unfortunately, the new evidence does not solve the difficulties. In [84], in a rather broken and not too clear context, we find a mí-tur uru a- pa? Ÿ-a-yi kiÿ. If the reading is correct and the sign uru not added by mistake this reference seems to prove the existence of a town Apum, but the evidence is slim. In [101] the king of fiun, located somewhere west of Leilan, reports that he has prepared the defense of the town of fiun and the Óalla of the country of Apum, and at the end of the letter he states that the town of fiun and the country of Apum is well (cf. also [102], 26f.). This seems to indicate that fiun, a vassal kingdom of Leilan, was actually within the territory of Apum and that a town Apum perhaps was located west of Leilan. fiun itself is not yet mentioned in any third-millennium or Old Assyrian sources and could be a candidate, especially if we compare the Old Assyrian route through the Habur to the Old Babylonian itinerary and assume that fiun in the latter substitutes Apum in the former (cf. below I.1.2.4). 1

2 THE ROYAL ARCHIVES FROM TELL LEILAN Habur Plains harks back almost to the beginning of his reign. This evolves clearly from the Mari Eponym Chronicle (Birot 1985), in which the annotation for the third regnal year of fiamí -Adad mentions that he was defeated by the Lulleans, i.e., a local hill population, at Lazap t(um), a town located in the region south of Leilan (see I.1.2.4). fiamí -Adad possibly conquered the city from an earlier ruler, but the name of this ruler or, indeed, the name of any other pre-fiamí -Adad ruler of Leilan is still unknown. The archives from Mari and Chagar Bazar (Talon 1997) provide us with the names of a number of officials active in fiubat-enlil, but otherwise give few details about the city. From Leilan itself there is some evidence from this period, primarily from the buildings excavated on the Acropolis in 1985 (Whiting 1990a and 1990b), but also from the Eastern Lower Town Palace, where several rooms have yielded sealing fragments bearing legends of fiamí -Adad, IÍme- Dagan, and some of their officials. 4 Events during the phase immediately following the death of fiamí -Adad and the disappearance of YasmaÓ-Addu from Mari are still poorly known, and a more precise elucidation must await retrieval or publication of additional sources. It is clear, however, that even before the death of the king himself, and while his power seems to have reached its zenith, Leilan and the surrounding area were not under his complete control. Evidence from Mari documents shows how the barbarian Turukkeans, deported from the east Tigris region and settled close to fiubat-enlil (in Amursakkum), rebelled and caused fiamí -Adad and his sons considerable difficulties (Eidem 1993a). To what extent the sons of fiamí -Adad managed to retain control of the Habur Plains after the death of their father is difficult to ascertain, but we know that the old official Samiya stayed in control of fiubat-enlil some five years into the reign of Zimri-Lim. Texts from Mari show that Samiya was in conflict with Turum-natki, who was supported by inhabitants of fiubat-enlil itself. Together they solicited help from Zimri-Lim of Mari to get rid of Samiya, and promised him the treasures of fiamí -Adad, apparently still somewhat intact, in return. Likewise Samiya received a letter from SimaÓ-ilânem of Kurd, who offered to kill Turum-natki and join the country of Apum to fiubat-enlil. Thus Turum-natki is the first documented figure who may have had legitimate, i.e., pre-fiamí -Adad, claims on Apum, but there is as yet no explicit evidence for this. This situation came to an end in the year ZL 3', when Ibal-pî-El II of EÍnunna invaded the Habur Plains supported by Qarni-Lim of Andarig. Samiya disappears from view, and Turum-natki somehow Another solution might be that Apum and fieón (for an etymology Hot (Springs), see Bonechi 1998, 221f.) originally referred to two different aspects of the Leilan settlement, its extensive lower town and the ancient core represented by the citadel. These two realities are, of course, termed respectively adaííum and keróum in Old Babylonian northern texts, and this double nature of major tells, which dates back into the third millennium B.C., could have created different names for parts of the same settlement (cf. the situation at Ebla, where a special designation [sa-za ki x ] was used for the citadel). Since the k rum merchant quarter in Leilan was located in the lower town, the Assyrians would then have used the name Apum rather than fieón. None of this, however, is very convincing on present evidence, and we must conclude that the problem cannot be solved yet. 4. For the 1987 material, cf. the example published here in Appendix 2, no. 7 (fiamí -Adad). Other figures attested in such evidence include Kani-PI, son of atni-addu, servant of IÍme-Dagan [L.87-1281]; Sam ya, servant of fiamí -Adad [L.87-1279]; Lîter-Íarr ssu, servant of fiamí -Adad (cf. Parayre 1991b, 138 no. 14; a bulla with the seal of this official has been found also at Acemhöyük, see Tunca 1989, 483). Samiya and Lîter-Íarr ssu (Whiting 1990a and 1990b; for Lîter-Íarr ssu, see D. Charpin, ARMT XXVI/2, p. 29 c) are well-known figures in this phase of Leilan history, and both are known to have been stationed at fiubat-enlil toward the end of the period. Kaniwe(?) is not attested previously.

THE LETTERS 3 came to grief, since we hear that Qarni-Lim inters him in Apum, mourns him, and then places one of his sons on the throne of Apum (see Eidem 1994). The new king was possibly a certain Z zu, who is known from a few Mari texts. One of these is ARM X 122+ (Durand 1987h = DEPM III, no. 1140), in which Zimri-Lim, campaigning in the north, reports to his queen fiiptu that he has broken the enemy lines and managed to join his allies Z zu of Apum and the troops of EluÓut. 5 Another reference to Z zu, in A.350+ (Charpin 1990b = DEPM I, no. 333), concerns reports of his death, attributed consecutively to an illness, to a serious accident, and finally to natural causes. Following this event, officials of Bunu-IÍtar of Kurd arrive to seal the residence of Z zu and retain a caravan of his, assembled to transport grain from AzamÓul to SapÓum, while Qarni-Lim of Andarig, presumably then residing in fiubat-enlil, proceeds to install ya-abum as king. This and other texts from that time reflect a situation in which the Sinjar city-states of Andarig and Kurd seem to share control of Apum. A new piece of information provided by one of the Leilan treaties (L.T.-1) is that ya-abum was the son of Turum-natki. This leads to a suspicion that the unpublished evidence for ya-abum s accession, referred to by Charpin (ARMT XXVI/2, p. 60), coupled with that of the anonymous son of Turum-natki who replaced him, means that this son was, in fact, ya-abum and not Z zu, who instead may have been installed as a Kurd -sponsored king in AzamÓul. If so, we have evidence for two sons of Turum-natki who may, for a time, have shared the kingship of Apum, residing in fiubat-enlil and AzamÓul respectively. 6 In any event, ya-abum was left as king of Apum, although under tutelage from Andarig during the next years until mid-zl 9', when Elamite troops and their allies invaded the Habur. Events during this year are exceptionally well documented in the published sources from Mari. ya-abum was killed and an Elamite general, Kunnam, resided for a time at fiubat-enlil and controlled a large sector of the Habur Plains, but later the same year the Elamites and their allies were defeated. Qarni-Lim came to a sad end (cf. Heimpel 1996a), and his place was taken by Atamrum, ruler of the Sinjar town AllaÓad, who now controlled both Andarig and fiubat-enlil, where fiupram (also king of the town Sus ) was installed as governor. During the following years Atamrum, and simultaneously and subsequently from ZL 11' his successor, imdiya, were in control of Leilan. The 1987 excavations in the Eastern Lower Town Palace produced only a few inscriptions from this entire period. Neither Turum-natki nor Z zu is attested, but a few sealings have the legend of B2l -em q, servant of ya-abum. Atamrum of Andarig and his governor (Í piˇum) fiupram are not attested, but imdiya, Atamrum s successor, is represented by a few sealing fragments and two tablets. 7 After 1762 B.C., when Hammurabi of Babylon brought an end to the power of Mari, documentation for the history of northern Mesopotamia comes to an end for a very long time, although 5. For this interpretation, which differs from that of Durand, see Eidem 1989b, 365. 6. Such a setup is, of course, reminiscent of that found in the texts published here, for which see below I.1.1.3. 7. The envelope fragments from room 2 (see Appendix 2, no. 1) are especially interesting since they carry the full legend of imdiya s seal. The two tablets are a legal document sealed by a servant of imdiya and dated with the limmu Óa-ab-d[u(?)- ] found in room 22, and a single letter [L.87-887] addressed to imdiya, from room 5. The latter text, which contains interesting new information, will be published separately by F. Ismail.

4 THE ROYAL ARCHIVES FROM TELL LEILAN not completely. Texts from Tell al-rimah in the Sinjar Plain (ancient Qaˇˇar ; see Charpin and Durand 1987; also Eidem 1989a) provide some information about a few years within the decade 1760 50 B.C., and show that Hammurabi of Babylon now was in control of northeastern Mesopotamia, 8 while texts from Tell Asharah (ancient Terqa) on the middle Euphrates document developments under the so-called ana dynasty (Rouault 1984 and 1992). When the first rooms of the Eastern Lower Town Palace at Leilan were excavated in 1985 a handful of tablets and sealings appeared bearing names of kings, evidently dating to a period later than that documented at Mari. Several of these tablets were administrative texts dated with the otherwise unknown limmu IÍme-El and warki IÍme-El, and sealed with the seal of Yak n-aíar, son of Dari-EpuÓ and king of Apum. This royal name provided a welcome chronological link with southern Mesopotamia, where the year-formula for the 23rd regnal year of the Babylonian king Samsu-iluna records the destruction of fiahn (= fieón ), the capital of the country of Apum, and in a variant version mentions a certain ia-ku-u[n- ]. Thus the year Samsu-iluna 22, which equals 1728 B.C., could be considered a terminal date for the king Yak n-aíar. The Babylonian yearformula is the latest extant reference to Leilan, where major occupation also seems to have come to an end at about this time. 1.1.2. Evidence from the Eastern Lower Town Palace The material excavated in the Lower Town in 1985 has been dealt with in some detail by Whiting (1990a and 1990b) and need only be briefly summarized here. It was found mainly in the two partially excavated rooms numbered 2 and 5, and it included sixteen administrative texts dated with the following limmu eponyms: Adad-bani (1 text), AÍÍur-takl ku (1), IÍme-El (7), warki IÍme-El (1), and Niwer-Kubi (6). All these texts came from room 2 except the one dated to Adad-bani, which was found in room 5. In addition, half a dozen undated or fragmentary administrative texts and a single letter addressed to Samiya were found. The sealings had inscriptions relating to the kings fiamí -Adad, imdiya, Till-Abnû, Mutiya, and Yak n-aíar. The seal of Yak n-aíar himself was also found impressed on all the administrative tablets dated with the limmu IÍme-El (including the one dated warki IÍme-El). From this admittedly limited material Whiting was able to define the main premises for the chronological situation. As mentioned above, Yak n-aíar can be connected to the 22nd regnal year of Samsu-iluna, which provides a terminus ante quem for the material, whereas the lower floors of the building obviously dated back to the time of fiamí -Adad. Whiting could define the intervening period as follows: a) any time that Dari-EpuÓ, Yakun-AÍar s father, may have ruled; b) any time that other unknown predecessors of Yakun-AÍar (such as possibly Mutiya, may have ruled; c) the length of Yakun-AÍar s reign prior to the limmus attested in the archive (Whiting 1990b, 572 n. 106). During the 1987 season rooms 2 and 5, which produced the majority of the tablets found in 1985, were completely exposed and, in addition, a large new area of the palace was excavated. The epigraphic finds from 1987 were recorded with 1068 field numbers, but since subsequent study has separated fragments recorded as one field number and joined others recorded separately, this figure is only a rough guide to the number of separate items found. Leaving aside the ca. 250 sealing and 8. For some remarks on the order and extent of this control, see below I.1.2.1.

THE LETTERS 5 envelope fragments, for which absolute figures are less meaningful, the material can be summarized as follows: 219 letters or letter fragments published in this volume. Although a few theoretical joins among the sometimes very small fragments may have passed unnoticed, this material can safely be said to represent at least 200 different texts. Some 80 fragments from at least six larger tablets containing the texts of political treaties and a version of the Sumerian King List. 328 administrative texts or fragments with preserved (or reconstructible) limmu-date representing this exact number of individual tablets. Some 125 administrative texts or fragments without preserved date. Some 140 smaller fragments, of either administrative texts or insignificant pieces. As discussed in detail by Ristvet and Weiss in their introduction, most of these tablets and fragments were found in two main groups, in room 2 and rooms 17/22/23 respectively. The first group consists of administrative texts, many dated to the limmu-year IÍme-El, firmly associated with the ruler Yak n-aíar. The texts predominantly concern wine, and seem to be part of a smaller, specialized archive, and are the latest texts found in the palace. The much larger second group, scattered in three different rooms, consists of administrative texts, letters, and fragments from political treaties. Most of the texts presented in this volume belong to this group. 9 With very few exceptions the material belongs to the latest phase of Leilan history, i.e., the period ca. 1755 28 B.C. Apart from imdiya, who may have continued to be in control of Leilan some time after the Mari archives came to an end in 1761 B.C., three different kings, Mutiya, Till-Abnû, and Yak n-aíar, are attested for this phase, their names occurring in seal legends, in letters, and as contracting partners in treaties. The chronological framework for this period has to be developed from the Leilan material itself, and some of the basic premises for such a construction, from several levels of analysis, must be discussed here. 1.1.3. The Leilan Kings The sequence of the three latest Leilan kings presents no immediate problem. The basis for a reconstruction is provided by two synchronisms with material from Babylonia. The first is the Samsuiluna year formula already referred to, which shows that Yak n-aíar was king in 1728 B.C., a year that may well have been the last of his reign. The second synchronism is provided by the limmu abil-k2nu, firmly associated with the reign of Mutiya, and which is also found in tablets from Sippar. From the Sippar material the limmu can be dated to a year either very late in the reign of Hammurabi or very early in the reign of his successor Samsu-iluna, i.e., approximately 1750 B.C. (see Veenhof 1989; cf. Charpin 1990d). Finally, the Leilan material itself provides ample evidence that Till-Abnû directly succeeded Mutiya, who is referred to retrospectively and in very specific terms in several letters sent to Till-Abnû. 9. Exceptions are (1) two letters from room 12: a fragmentary letter exchanged between two palace officials [168], and a small fragment from a letter sent to Mutiya from an unidentified writer [21]; (2) a letter fragment found in room 2 [160].

6 THE ROYAL ARCHIVES FROM TELL LEILAN Mutiya The name Mutiya is a hypocoristicon of Mutu-AbiÓ, 10 but the short form is exclusively used in the letters. The complete form is attested in the seal legend of Mutiya found impressed on sixteen administrative texts dated in the limmu abil-k2nu: mu-tu-a-bi-ió, dumu Óa-lu-un-pí-mu, na-ra-am d im, ù d nin-a-pí-im Mutu-AbiÓ, son of alun-pî-(yu)mu, 11 beloved of Adad and B2let-Apim In this inscription Mutiya is not given any title, but this is found in L.T.-2, where he is repeatedly referred to as lugal m t Apim king of the land Apum. Till-Abnû Although Till-Abnû 12 is the current form of the name, the hypocoristicon Till ya is found once, in the address of [23] sent from Hammurabi of alab (also possibly in the address of [27] sent from Bin-Dammu, the alab general). A number of administrative tablets are sealed with seals of servants of Till-Abnû or with his own seal. The two first lines in the legend invariably read: Till-Abnû, dumu Dari-EpuÓ Till- Abnû, son of Dari-EpuÓ, but the third line is often not visible, since it has been erased by repeated impression of the seal legend onto the tablets. On a total of seven tablets, however, the third line appears as: ìr Ía d [ ] servant of the god [ ]. Six of these texts are dated in the limmu Amer-IÍtar (months iii, iv, v, x, xi) and one in a limmu [ ]-IÍtar (month iii), which almost certainly is identical to Amer-IÍtar. On one tablet (dated Amer-IÍtar, vii 15) the third line appears as: lugal ma-a-at [a-pí-im (ki) ] king of the land Apum, and this is the title found in several of the treaties, e.g., L.T.-3, which is dated to the limmu Amer-IÍtar. Yakun-AÍar Again the full version is the most current form, but a hypocoristicon, Yak ya (ia-ku-ia), is found in several seal legends belonging to servants of the king. The legend of his own seal is found on numerous tablets from room 2 (all dated in limmu IÍme-El and in two cases warki IÍme-El): ia-ku-un-a-íar, dumu da-ri-e-pu-uó, lugal ma-a-at a-pí-im Yak n-aíar, son of Dari-EpuÓ, king of the land Apum (see Parayre in Weiss et al. 1990, 564, fig. 34) 10. For this name meaning Man of EbiÓ (Jebel Hamrin), see Durand 1991c, 85. It could be speculated, of course, that this name has a bearing on the history of the family to which Mutiya belonged. 11. This follows writings like Óa-lu-(un/m-)pí-PI-mu in texts from Mari (ARMT XVI/1, 97) and here L.T.-2 iii 5' Óa-lu-pí-ú-mu. 12. For this name and its possible etymologies, cf. Durand 1987g. The current spelling is ti-la-ab-nu-ú, but we also find til-la-ab-nu-[ú] ([52] and [55] from alu-rabi), ti-i[l-la-a]b-nu-ú ([89] from fiukrum-teííup), tilab! - nu-úÿ ([89], 29), til-na 4 ([105] from MeÓilum and [128] from BaÓdi-Lim), t[i-l]i-a[b]-nu ([113] from Sumu-Ditana), ti-la-ab-ni ([118] from AÓuÍina, and probably [122] from [ ], also in two letters from Kuzuzzu [137] and [139]). For variations in the spelling of -Abnû, ab-nu/ab-nu-ú/ab-ni in letters from Ayaabu of fiun, see below ad [93]. These variations fit Durand s observations and justify the normalization Till- Abnû used in this volume. Again, it could be speculated whether the name if formed from the homonym GN of a town west of the Habur Plains has a bearing on the history of the family.

THE LETTERS 7 THE DYNASTY Although a sequence Mutiya Till-Abnû Yak n-aíar seems secure, there remains a number of problems concerning these kings: their origins, their relationship to each other, and their periods of rule. It is clear that Till-Abnû and Yak n-aíar, both sons of Dari-EpuÓ, were brothers, and maybe other members of the family are attested. 13 Since Mutiya and Till-Abnû and Yak n-aíar had different fathers, it might seem that two different dynasties were involved, but apparently the situation is not that simple. A diachronic view of our material shows that men named Till-Abnû and Yak n-aíar were actively supporting Mutiya during his reign, and the evidence suggests a geo-political construction in which Mutiya, as king of Apum and based in fieón, was supported by two sub- or junior kings on the borders of Apum. Till-Abnû, in contrast to Yak n-aíar, is not attested as a correspondent of Mutiya, but the letter [110] (Ewri to his lord Till-Abnû) can with confidence be dated to the reign of Mutiya. Further, Till-Abnû is mentioned in two letters from the official Kuzuzzu, [137] and [139], almost certainly sent to Mutiya (for the historical context of all three letters see I.1.3.2). The implication of this evidence is that Till-Abnû supported Mutiya and must have been an ally, vassal, or governor of Apum. This observation may be connected with information in administrative texts dated to the limmu abil-k2nu. The first is [L.87-625] (19 v), which lists a delivery of wine from [ti]-la-ab-nu-ú lú fiurnat. An individual named Till-Abnû is mentioned also in [L.87-665] (11 ix) and [L.87-1411] (25 ix), but in these instances in an uncertain context. At the same time, we have evidence for another lú fiurnat, a certain Kuzzuri: [L.87-646] (25 viii abil-k2nu) lists a garment given to Kuzzuri lú fiurnat; [L.87-539] (7 viiib abil-k2nu; same entry in [L.87-1412] with same date) is a note concerning four jars of wine brought by Kuzzuri lú fiurnat when he came to meet with the king (in ma itti lugal ana nanmurim illikam). Although the possibility of homonymy precludes definite proof, it is a reasonable theory that Till- Abnû, for a time at least, was based at fiurnat prior to his accession, and that possibly he was assisted or succeeded by Kuzzuri. 14 The town fiurnat, apart from references in the Leilan texts, is mentioned in a number of Mari texts that provide some evidence for its location, and it has recently been suggested that it should be sought northeast of Jebel Sinjar (see ARMT XXVI/2, p. 83 sub e). ARMT XXVI/2, 422 provides 13. Two sealing fragments are relevant in this context. The first is [L.87-151], where the fragmentary three-line legend reads: [x x]- d [ ], [dumu d]a-ri-[ ], [x t]i-la-[ ]. If Dari-EpuÓ and Till-Abnû are involved here, we must have a figure related to, but presumably subordinated totill-abnû. The second is [L.87-152], again with only part of the legend preserved: [x x] xÿ- d Ÿ[ ], [dumu ia-k]uun-a-[íar], [ìr? ti-la-a]b-nu- úÿ. In this case it could be speculated that a nephew of Till-Abnû is involved and, hence, that Till-Abnû was fairly aged or considerably younger than his brother Yak n-aíar. 14. The Kuzzuri mentioned in [L.87-656+717] (27 x abil-k2nu), a section leader, could be a homonym. A man, Kuzzuri, is mentioned twice in letters, but no good links with a lú fiurnat can be established: as sender of [17] (reading, however, not completely certain) to his father Mutiya, hence, this figure was presumably a king. The letter is just a fragment, but the town Amaz far from fiurnat is mentioned. The other reference in [179] (address not preserved) is to an individual facing trial.

8 THE ROYAL ARCHIVES FROM TELL LEILAN interesting information showing that fiurnat must have been a fairly large walled town: (Hammurabi of Kurda) sent 2000 soldiers and they attacked the town fiurnat, which belongs to Zû- atni, and captured as much of its salóum, its cows and sheep, and people, as they could get hold of, but the townspeople mounted to the citadel, and saved themselves in the citadel (ll. 25 31). Charpin (1990b, 118f.) has shown that fiurnat probably belonged to Apum, and it can tentatively be suggested that fiurnat should be identified with Tell Qal at al H d on the wadi Rumeilan southeast of Leilan (see I.1.2.6, s.v. Ewri; and analysis of the historical events in I.1.3.2). 15 In the extant material Yak n-aíar appears as king of Apum on sealing fragments and in the sealed tablets dated to the limmu IÍme-El and its warki year, found exclusively in room 2. A man named Yak n-aíar occurs also, however, in the letters found in rooms 17/22: Sender of one letter to Mutiya, his father ; Sender of three letters to Till-Abnû, his brother ; Receiver of one letter from alu-rabi; Mentioned as brother of addressee in a letter sent to b lum (here probably Till-Abnû); Also mentioned in letters sent from Sangara of Till to b lum. Finally, a single-entry administrative text, dated to the otherwise unattested limmu AÍÍur-kaÍid, mentions an issue to Kazikuk, lú-tur Yak n-aíar, lú Il n- ur [L.87-1461] (dated 18 xii, from room 22). Again, granted the possibility of homonymy, we can assume that Yak n-aíar, prior to his accession to the Apum throne, was based in Il n- ur. Since Yak n-aíar may have retained a position as viceroy through the reign of Till-Abnû, the evidence for his activities and sphere of action is more extensive and provides some support for a location in Il n- ur. This important town cannot yet be located very accurately, but it was placed west of Leilan (see I.1.2.4). What emerges then is the theory that Mutiya had placed Till-Abnû and Yak n-aíar as viceroys in fiurnat and Il n- ur respectively. It is important to note that these two towns marked major border points for the country of Apum. In the time of Zimri-Lim, Il n- ur appears to have been an outpost for the territory controlled by Mari and areas in the eastern part of the Habur; in ARMT XXVI/2, 301 this is explicitly stated by the Mari envoy Yam ûm, who refers to the town as the l p ˇi Ía b l ya. Similarly, fiurnat s position may have been useful for protecting the southeastern border of Apum. Whether this geo-political reconstruction is strictly correct or not, we have so many corroborative pieces of evidence in the texts that the basic situation of the brothers Till-Abnû and Yak n- AÍar supporting Mutiya hardly can be doubted. This, of course, raises the question of the exact relationship among the three men. imdiya s reign and control of Leilan beyond the last year of the Mari archives, 1761 B.C., are unknown, but presumably of short duration, and the year of the limmu abil-k2nu, which may be the last regnal year of Mutiya, can be dated to ca. 1750 B.C. Consequently we need to fill a ca. ten-year gap in the history of Leilan. 15. fiurnat is not attested outside the Old Babylonian sources referred to here. If identical to Qal at al H d, the town may have changed its name, since there is evidence for later occupation on the site (see Meijer 1986, 19).

THE LETTERS 9 No kings other than Mutiya are attested directly, but it seems possible that Dari-EpuÓ, the father of Till-Abnû and Yak n-aíar, once ruled Leilan. Some evidence to this effect can be found in [28], where Ea-malik states that Till-Abnû has ascended his father s throne, and in [149], where Tak2 writes to his lord: This is the advice that your father gave me: If you write to Till- Abnû, he shall come to you(r aid) like one man, and if he calls on you(r help), go to him at once! Assuming that the father in question really was Dari-EpuÓ and that he once ruled Leilan, it might be thought that Mutiya was an outsider who had usurped the throne after the death of Dari-EpuÓ. In [28] Ea-malik says that Mutiya ascended his throne and Till-Abnû the throne of his father, but this difference is hardly of any consequence. It seems likely that Mutiya died a natural death (cf. [128]), but the circumstances of Till-Abnû s accession are not revealed and it cannot be excluded that it involved some crisis. In [127] Abbutt num, referring perhaps to the time of Till-Abnû s accession, writes: When the elders of [Apu]m went to KaÓat [to] my lord [ ], a statement that could be interpreted to mean that Till- Abnû, having being deprived of his rights by Mutiya, had sought refuge in KaÓat. At the same time, we have a hint that Till-Abnû s accession may have been disputed in certain quarters, since Hammurabi of alab in [24] finds it necessary to affirm his kingship. However, the best theory that can be offered at present is perhaps that Till-Abnû and Yak n- AÍar were nephews of Mutiya. It could also be suggested that some of the problems concerning the accession of Till-Abnû relate to a competition between Till-Abnû and Yak n-aíar for the succession, rather than any crisis between Till-Abnû and Mutiya. Although again there is no direct evidence for enmity between the two brothers (cf. [48]), the likely brevity of Till-Abnû s reign makes it entirely possible that he could have been ousted by a malcontent brother. Finally, some questions concerning these three kings can be raised. Where did they come from? Did they have any connection to older rulers of Leilan? How did they come to power? We simply do not know. Dari-EpuÓ, Mutiya, and his father alun-pî-(yu)mu are not known from other sources, and our texts reveal nothing about their origins (excluding the doubtful evidence from the names Mutu-AbiÓ and Till-Abnû). The same names occur in texts from Mari, but in contexts that render it unlikely that the same individuals are involved. The various low-status people carrying these names must, of course, be left out of consideration, but vaguely suggestive is the case of a certain Mutu-AbiÓ mentioned in ARM V, 2 (see Durand 1987c, 212 15; = DEPM II, no. 533). Here, YasmaÓ-Addu reports to IÍme-Dagan that he has defeated enemies who were preparing to attack Mari, and among the prominent individuals killed is a certain I ur-dagan, brother of Mutu- AbiÓ. The letter cannot be dated precisely, but the events described suggest a connection to texts that concern YasmaÓ-Addu s trouble with the Yaminu tribes, such as ARM I, 5, which can be dated to the year of the limmu Aw liya (= fiamí -Adad 30, ca. 1783 B.C.), for which the annotation in the Mari Eponym Chronicle mentions a victorious YasmaÓ-Addu campaign against the Yaminu tribes (Birot 1985, 232). ARM I, 5 should, in any case, date to one of the last years of fiamí -Adad s reign and is, therefore, some thirty years earlier than our material. This time gap renders an identification with the Leilan king rather unlikely, but since there was a tendency to use the same names within noble or royal families, the two might still have been related. Although the evidence from ARM V, 2, is, most likely, irrelevant in this context, the possible connection between our kings and old enemies of fiamí -Adad at least has some historical probability. Both the family of Turum-natki (see I.1.1.1) and(?) that of Mutiya and Dari-EpuÓ could have been related to a dynasty that ruled around Leilan shortly before fiamí -Adad conquered Apum. For the Sinjar kingdom Andarig there is evidence for a brief pre-fiamí -Adad dynasty preserving local position, and subsequently regaining the ancestral throne during the reign of Zimri-Lim (cf. I.1.2.4,

10 THE ROYAL ARCHIVES FROM TELL LEILAN s.v. Buriya), while from Kurd we have a pre-fiamí -Adad king, AÍtamar-Adad, matching the homonymous Kurd king in the texts published here (cf. I.1.2.5., s.v.). This shows clearly how fragmentary and ambiguous the present evidence is, and it can only be hoped that future discoveries will serve to fill in some of the serious documentary gaps and reveal a clearer pattern. Meanwhile we have presented the outlines for an operational understanding of the texts published here: three members possibly of the same family ruling Apum, with Mutiya as king assisted by his juniors, Till-Abnû and Yak n-aíar. Subsequently the triad dwindled when Mutiya died and was succeeded by Till-Abnû, who, in turn, was followed on the throne by Yak n-aíar. 1.1.4. The Leilan Limmus In order to date the reigns of the last Leilan kings more accurately we turn next to the twelve limmus identified in the dated administrative texts. 16 A number of recent studies of the limmu eponyms from this period results in a sequence for the whole period, based primarily on the evidence from the Kültepe Eponym List (Veenhof 2003) and the Mari Eponym Chronicle (Birot 1985). Using these key sources, which list almost all limmus from ca. 1975 B.C. down probably to the death of fiamí -Adad, it is possible to isolate approximately forty eponyms later than this event, and belonging to the last part of the period related to level 1b at Kültepe/KaniÍ (ca. 1800 1725 B.C.; see Veenhof 1985, 1998, and 2003). A number of these late eponyms is still missing in extant sources, and it is not yet possible to establish a continuous sequence. The fact that only two of the eponyms related to sizeable portions of texts from the Lower Town Palace, viz., abil-k2nu, IÍme- El, Amer-IÍtar, Nimer(or Niwer)-kubi, and Ipiq-IÍtar, are attested elsewhere in the North is hardly surprising, since the Leilan material probably is later than any other text group so far discovered (cf. Charpin 1988). Helpful for our material are the texts found at Leilan in 1991, which have provided four limmus belonging to the time when Qarni-Lim of Andarig controlled Leilan, i.e., the period ca. ZL 4' 9'. Three of these limmus, AÍÍur-takl ku, Zabzabu, and AÓu-waqar, can be shown to have followed each other directly, whereas the fourth, Adad-bani, cannot be placed (see Van De Mieroop 1994). The table at the end of this section provides an overview of the limmu-dated texts found in 1987. 17 abd[u(?)- ] is associated with the seal of a servant of imdiya. It should be noted that abdu-iítar (limmu earlier reported from the Leilan Acropolis) has proved not to be an eponym (see Whiting 1990b, 573); abil-k2nu texts are associated with seals of Mutiya or his servants; Amer-IÍtar texts are associated with seals of Till-Abnû or his servants. A single text from month iv, however, is sealed with the seal of a Mutiya servant, making it likely that this was the first regnal year of Till-Abnû; A single Ipiq-IÍtar text is sealed with the seal of a servant of Till-Abnû; IÍme-El/warki IÍme-El texts are associated with the seal of Yak n-aíar. 16. For this evidence, see also the editions by Vincente 1991 and Ismail 1991. 17. For the names and the sequence of months (so-called fiamí -Adad calendar ), see Charpin 1985; and Cohen 1993, 255ff..

THE LETTERS 11 Most likely abil-k2nu, Amer-IÍtar, and Ipiq-IÍtar provide a consecutive series beginning ca.1750 B.C., followed by a gap of unknown duration before the years IÍme-EL/warki IÍme-El, which must belong before 1728 B.C. The remaining limmus, not explicitly associated with particular kings, are best considered in direct conjunction with their texts, but some brief remarks can be offered. AÍÍur-takl ku and Adad-bani are most reasonably identified with their namesakes in the 1991 texts from Leilan, and can, therefore, be dated to the reign of ya-abum. It must be noted, however, that both of these names are attested with at least two different eponyms. For AÍÍur-takl ku (cf. Whiting 1990a, 216) we have (A) son of Ennam-AÍÍur from AÍÍur itself and from the Leilan Acropolis, and (B) son of Enlil-nada, attested in Kültepe Ib (in a group of texts that also have PilaÓ-Sîn and b- ill-aííur; see Veenhof 1998, the limmu is no. 13 in his list). Eponyms AÍÍur-takl ku without patronymics are found in the Mari Eponym Chronicle (D4) and in the Temple Stair texts from Rimah. Adad-bani presents the most difficult case, since two or three different limmus Adad-bani are attested elsewhere, and all are dated to the reign of fiamí -Adad. Adad-bani also has a single occurrence in room 5 (found in 1985), and Whiting (1990b, 572) opted to identify him with probably the latest eponym, Adad-bani son of Puzur-il, which can be dated to the third year before fiamí -Adad s death. AÍÍur-kaÍid may be identical to the eponym attested in Kültepe, son of ZI-lá-mu (see Veenhof 1998, no. 8). The single text from Leilan is the one that mentions Yak n-aíar lú Il n- ur (cf. I.1.1.3). If this is the same individual as the later king of Leilan, the text should date before his accession to the throne in Apum. Azzubiya is probably identical to the limmu of this name attested in the Iltani archive from Rimah, and can be dated to the years ca. 1760 50, i.e., late in the reign of Hammurabi of Babylon. PilaÓ-Sîn, for which the single text provides no conclusive internal evidence, but its occurrence elsewhere supports a late date. Veenhof (1998) suggests a date ca. 1770 B.C., but this seems too early (the limmu is no. 34 in his list). Nimer-kubi is attested also in the tablet found at Qal at al H d southeast of Leilan, and dated to warki Nimar-kubi (see below I.1.2.6, s.v. Ewri). Since it is best represented in the texts from room 2, it may be placed close in time to IÍme-El. A limmu fiu-b2l is known from the Mari Eponym Chronicle (B7) for the year before fiamí - Adad s accession, but this definitely seems too early for our limmu. The relevant text mentions the official Bayy nu, who can be firmly associated with the reign of Till-Abnû. Although his activity may cover a long period, it can hardly be stretched back that far. It seems likely then that all these limmus belong within the time ca. 1760 28 B.C. Azzubiya clearly comes before abil-k2nu, and the same may be true for AÍÍur-kaÍid, PilaÓ-Sîn, and fiu-b2l. If correct, we may, adding some eponyms from the Iltani archive at Rimah, have virtually all the limmus from the reign of Mutiya and his predecessor(s), imdiya (and possibly Dari-EpuÓ), after the fall of Mari. Then, after Ipiq-IÍtar (which may be the last regnal year of Till-Abnû), there is a considerable gap, since, dating Ipiq-IÍtar to ca. 1748 45 B.C., we seem to have only the limmus from room 2 representing a total of four years to cover the period until 1728 B.C. Although it is likely that the texts from room 2 belong toward the end of this period and that warki IÍme-El may actually equal 1728 B.C., this cannot be proved.

12 THE ROYAL ARCHIVES FROM TELL LEILAN DISTRIBUTION OF LIMMU-DATED TEXTS limmu/month i ii iii iv v vi vii viii viiib ix x xi xii abd[u? - ] 1 1 abil-k2nu 179 16 27 24 32 17 20 14 13 1 Amer-IÍtar 60 2 11 9 7 7 5 3 4 5 2 1 Ipiq-IÍtar 17 1 3 2 3 5 1 1 1 IÍme-El 51 1 2 1 1 11 8 3 8 9 2 warki IÍme-El 1 1 Adad-bani 2 1 1 AÍÍur-kaÍid 1 1 AÍÍur-takl ku 2 1 Azzubiya 1 1 Nimer-kubi 11 1 3 6 PilaÓ-Sin 1 1 fiu-b2l 1 1 The names of limmus are followed by the total number of texts, including tablets where the evidence for a month is lost. Month viiib, which is attested only for abil-k2nu, is the intercalary Addarum. Figures in boldface include sealed tablet(s) with evidence of the royal name. All texts dated with AÍÍur-takl ku, IÍme-El/warki IÍme-El come from room 2, where also ten of the tablets dated with Nimer-kubi were found. With the exception of a few texts from isolated contexts, all the other dated texts were found in rooms 17, 22, and 23. From the tablets found in 1985 can be added attestations for months ii, iv, and vi (two texts) in the year Nimer-Kubi, but not for additional months in other years. 1.1.5. Archival Context of the Tablets Having presented an overview of the chronological situation, we may briefly focus attention specifically on the archival context of the texts published here, which were virtually all found in room 22. The tablets from this room, together with the much smaller groups from the two adjacent rooms 17 and 23, form one group, as is evident from the fact that joins can be made with fragments between rooms, like the letter [9] and several of the treaties. It is apparent that this archive consists of several different sub-archives. On a synchronic level there is a mixture of letters and administrative texts similar to other smaller palace archives, such as the Iltani archive at Rimah and the Kuwari archive at Shemsh ra. This can be explained by the kind of administrative texts involved at all three sites,

THE LETTERS 13 i.e., documents pertaining to expensive items such as metals, garments, expensive food products (or, at Rimah, documents pertaining to the household of the archive owner), 18 and requiring closer control by the archive owner or the top-level administrators, whose seals at Leilan were applied to many of the texts. Other administrative documents concerning agriculture and the circulation of agricultural products were kept elsewhere as at Shemsh ra (cf. Eidem 1992, 33ff.). Also, accounts for the circulation of wine and beer were kept completely separate, perhaps close to the actual cellars. Apart from such texts excavated in the Rimah palace, the small archive from Leilan found in room 2 of the palace is a good example (see Whiting 1990a; also Ismail 1991). Finally, the inclusion of treaty texts in royal archives is not surprising. On a diachronic level the archive represents several separate archives and we can, for present purposes, distinguish two main groups: A. Texts relating to Mutiya: 22 letters addressed to Mutiya (+ X letters sent to him as b lum) 179 administrative documents dated with limmu abil-k2nu (= last regnal year) 1 political treaty B. Texts relating to Till-Abnû: 99 letters addressed to Till-Abnû (+ X letters sent to him as b lum) 60 administrative documents dated with limmu Amer-IÍtar (=1st regnal year) 17 administrative documents dated with limmu Ipiq-IÍtar (= 2nd regnal year) 3 political treaties These groups constitute the core of the archive and should provide a key to its composition. Turning again to a comparison with the archives from Rimah (the Iltani archive and related texts) and Shemsh ra, which, in contrast to our archive, cover only single administrations, we note that both have a roughly similar composition: letters covering a few years and administrative documents heavily concentrated in a single year. This distribution can be explained as a tendency to keep letters while periodically selecting older administrative texts that were summarized on larger tablets to be recycled or simply discarded. Charpin has recently discussed such procedures specifically for the Mari archives and introduced a significant distinction between living and dead archives, the latter type being exemplified by the groups of small administrative notes found as fill in benches in the Mari palace (Charpin 1985, 253ff.). Adding to Charpin s typology, we might introduce an intermediate category, namely, that of an inactive archive not discarded, but no longer a current working body of material. 19 A good example of this category would be the letters found at Mari from the time of YasmaÓ-Addu no longer part of an active archive, but still kept for reference. These observations have obvious relevance for the Leilan material. Knowing that Till-Abnû succeeded Mutiya, we can assume that the tablets in group A were regarded as an inactive archive during the reign of Till-Abnû. Since we also know that Yak n-aíar succeeded Till-Abnû and is hardly in evidence in the tablets from rooms 17/22/23, we may further assume that group B prior to 1728 B.C. (when presumably the reign of Yak n-aíar ended) also came to be regarded as an 18. The complete analysis of 189 administrative texts undertaken by C. Vincente shows three dominant groups: silver/metal: 36.5%; foodstuffs: 30.1%; garments etc.: 18.5% (Vincente 1991). 19. I owe this idea to MacGuire Gibson.

14 THE ROYAL ARCHIVES FROM TELL LEILAN inactive archive. This means that the main active archive of Yak n-aíar, if preserved, is located elsewhere in the palace. Dealing with a composite inactive archive, however, leads to a complex situation, since perhaps several levels of selection must be reconstructed in order to clarify the composition of the extant material. The complexity involved is illustrated by the fact that, in contrast to the archives at Rimah and Shemsh ra, where the administrative documents cluster in a single, presumably final year, we are here faced with the exact reverse, namely, a marked diachronic decrease in the number of administrative texts. Rather than review the whole range of possible reorganizations for the archival composition, we shall instead search for the most reasonable solution, suggesting two main stages in the formation of the archive. STAGE 1 On his accession Till-Abnû selected from the archives of Mutiya series of letters and administrative texts to be kept. The small number, the limited range of correspondents, and the narrowly circumscribed vista of subject matter in the letters sent to Mutiya make it unlikely that the corpus is complete. As will be shown below (I.1.3.2), the letters deal mainly with events that occurred shortly before Mutiya s death, and the texts may have been kept as still relevant. The very compact and perhaps near complete series of administrative texts from months v xii of the year abil-k2nu clearly constitutes the result of a deliberate selection. Till-Abnû on his accession, which occurred at the end of this year or shortly into the next (Amer-IÍtar), may have wanted to keep a fairly complete set of accounts dating some months back for easy reference and checking. All these texts were presumably kept with the main archive of Till-Abnû through his reign, which may have lasted only the ca. 2 years for which we have explicit evidence. STAGE 2 After Till-Abnû s disappearance, Yak n-aíar inherited the archive of his predecessor and decided to deselect most of it while keeping selected letters and some administrative documents from the time of Till-Abnû in increasing numbers relative to a diachronic scheme. Placing these stages in a wider framework, we can posit three different groups of texts from the period of the last three Leilan kings: A. Dead archives (if still preserved, likely to be found in secondary deposits): Mutiya texts deselected by Mutiya and by Till-Abnû Till-Abnû texts deselected by Till-Abnû (primarily administrative texts); B. Inactive archives (texts found in rooms 17/22/23): Mutiya texts selected by Till-Abnû Till-Abnû texts deselected by Yak n-aíar; C. Active archives (texts yet to be found, if preserved): Till-Abnû texts selected by Yak n-aíar Yak n-aíar texts.

THE LETTERS 15 It might be assumed that traces of the original arrangement of the many tablets found in rooms 17/ 22/23 are revealed by their spatial distribution. In theory, analysis of this problem is possible, since the debris containing the tablets was divided into a number of excavation units (lots), and all objects further given sequential numbers as excavation proceeded. Obviously the units defined archaeologically would hardly correspond exactly to possible archival units, but, despite overlaps, some clusters might still be visible. In particular, one could look for possible divisions according to genre, date, subject matter, or, in the case of the letters, according to receivers or senders. In order to illustrate the possibilities of such analysis, a few examples can be given. Area 8, lot 37, which contained 89 epigraphic objects, equals debris from a well-defined space in the northeastern corner of room 22. The tablets found include 30 letters, of which 9 are addressed to Till-Abnû, 5 to Mutiya, and 2 to b lum. The rest include 10 fragments, some specimens with partly broken address, and the single letter to Yak n-aíar from the archive. Among the 20 limmu-dated texts, 11 are dated to abil-k2nu, 7 to Amer-IÍtar, and 2 to Ipiq-IÍtar. This example is not encouraging, since the material statistically seems to be a virtual microcosm of the entire archive. The letters addressed to Mutiya (by name) were found in five different excavation lots and in no apparent cluster. However, all four letters sent from Hammurabi of alab came from a single excavation lot, namely no. 37. Turning to the series of letters sent to Till-Abnû from different correspondents, we find that the letters from Aya-abu were found in four different lots, those of Yam i- atnû in seven lots, and those from Buriya in three lots. In provisional terms it seems unlikely that such analysis will significantly alter the conclusions drawn here concerning the archival composition of the tablets. Considering the inactive status of the archive, the collapse of the building, and subsequent disturbances, it is not surprising that the tablets have become so mixed that only smaller segments of an original archival arrangement have survived. The observations on the composition of the archive have implications for the analysis of the texts. We return to some of these issues below (cf. I.1.3.1). 1.2. Synchronic Survey 1.2.1. alab and Babylon The state of YamÓad, centered in alab (modern Aleppo) in northwest Syria, 20 appears to have been the decisive political power in the northern Jezira at the time of the Leilan archives. Given the brevity of the period documented, the establishment of alab s position so far east cannot be reconstructed in detail, but can be surmised as a consequence of the power vacuum left in the region after the collapse of the Mari state and subsequently the diminished influence of Babylon at the end of Hammurabi s reign. 20. For a recent survey of the evidence outside Leilan, see Klengel 1992, 44ff. Particularly relevant for our material is the letter from Samsu-iluna to Abban, AbB 7, 1, and the letter AbB 4, 24, which also concern relations between alab and Babylon. The texts from AlalaÓ (level VII) generally cover a slightly later period than the Leilan material (see Zeeb 1991).