199 A MYSTERIOUS AFFAIR OF STYLES-THE NINFATTE 5 POTTERY OF NORTHERN MESOPOTAMIA By Michael roaf and Robert killick ABSTRACT The elusive Ninevite 5 culture of northern Mesopotamia has repeatedly defied attempts by archaeologists to fathom its mysteries. In this article the relationship between the painted and incised styles of Ninevite 5 pottery is investigated with the help of Hercule Poirot who was present when Ninevite 5 w a en its name.' The evidence from earlier excavations and from recent r< $n -vork in northern Iraq is examined and a new verdict on the de\ elo] the Ninevite 5 pottery styles is reached. Archaeologists n compared to detectives, observing and gathering the clues, arrang ilysing the evidence, and finally, we hope, fitting the facts together to giv< . complete and accurate picture of what happened. Such a methodolog) v.- tly expounded by Hercule Poirot in his first appearance in print.*' "One Jail leads to another so we continue. Does the next fit in with that? A merveille! Good! We ran proceed. This next little fact - no! Ah that is curious! There is something missing a link in the chain that is not there. We examine. We search. And that little curious fact, that possibly paltry little detail that will not tally, we put it here!" He made an extravagant gesture with his hand. "It is significant! It is tremendous!" "T-es-" "Ah!" Poirot shook his forefinger so fiercely at me that I quailed before it. "Beware! Peril to the detective who says: Ht is so small it does not matter. It will not agree. I will forget it.' That way lies confusion! Everything matters" L'nfortunatelv archaeological facts are not as well constructed as those in detective stories: many of the dues have been obliterated by time and the principal witnesses are long since deceased. Nevertheless, in this investigation into the mystery ot the styles of Ninevite :> pottery we will try to follow Poirot's methods. First we will set the scene, then examine the evidence (noting the curious facts and paltry little details that do not tally), then we will cross-examine the witnesses, and finally we will suggest what we think is a plausible explanation. THE SCENE OF THE MYSTERY In the late twenties and early thirties the search for the prehistoric remains of t r l~A »Mt nurlr arc all taken from this lxx>k. We OW th« f jiff u are iht\ wr (,hll»ttc I'JHI, M/U. auKR * o,..,,,, |),ttf fJZuhM < fci»iir. Ih, .«r./«.««. Affair at StyU> (Urn- employed on this case to Ma Susan Roat. d'/n l'f/<> '>U Kit <|tii.r.iti.»tM in n.iliii ilitmij(litiiil mh:iiai:l koai and koui m i km i.ick Mesopotamia was all the rai^e.' The excavation nl ihr . ar.M (M|,M,,t ri) historical sites was now supplemented l>\ investigations ol the railirr formative periods of Mesopotamia!! civilization. By 1930 the main cultures of tLiter prehistoric and the early historic periods in tin* south of Iraq had been sue * e\sfu|| delineated and were given the labels 'Ubaitl, I'ruk, Jeindn Nasr, and Karly Dynastic and this nomenclature lias successfully survived the test of time/ With rather less satisfactory results the same terminology has been transferred to regions further north: and this has proved particularly unfortunate for ihe ill-defined Jemdet Nasr and the better known Early Dynastic cultures, since they are nol easily recognisable in the material culture of northern Mesopotamia* To cope with this problem the name Ninevite 5 has been adopted to describe the culture in the north of Iraq and the north-east of Syria at roughly this period. The label "Ninevite 5" <; was coined a little more than fifty years ago and like the names of other cultures has been used to describe not just the culture and the chronological period but also the pottery. But in this case, curiously enough, the term has been used to describe several very different styles of pottery, including painted and incised wares with unrelated motifs and of totally different ware.7 This then is the mystery we are seeking to solve what was the relationship between the different styles of pottery which have been (ailed Ninevite 5 ? **It is certainly curious /' / agreed. "StilL it is unimportant, and need not he taken into account." A groan burst from Pm rot. "What have I always told you ? Everything must hi ! iken into account. If the fact will not jit the theory let the theory go." 0 THE EVIDENCE Ol THE EYEWITNESSES Nineveh 1929 1932 The pottery, which was later to b< called Mirwrvii- was lirsi recognised as distinctive of a particular period in \\v I! 10 ison ol excavations at Tell 3 "In archaeology, pre-hisiory had suddenly become the fashion , , . They examined small, obst urc mounds all over the country, picked up fragments ol painted pottery wherever they went, labelling iheui, tying them up in bai^s, and examining the patterns ii was endlessly interesting. Although it was so old it was navV (Christie 198!, 471). 4 There have been some subdivisions and a few modifications; but attempts to dispose of the Jemdet Nasr culture have not proved successful see r'iukhcincr and Roliig 1986i. Similarly the attempt to abandon Karlv Dynastic II outside the Divala has not yet gamed much support. Sec, however, Porada, Hansen, and Dunham in press. 1 The forcing of an unwilling northern sequence into a southern mould has frequently obscured the situation, particularly as authors have often not been explicit about whether the culture or period was intended. In this article we <*s< the term Late L'ruk in a broad sense u i include northern cultures which strictK speaking may be contemporary with the Jemdet Nasr period in the south. To indicate (he ambiguity ol the term we have often pm Late I'ruk in inverted commas. '■ Ninevite 5 seems to us preferable to Ninevite \ , it onh because n is less easv to misread - 7 Onlv the decorated stvles have attracted much attention: there seems to fie some ambiguity about whether the Ninevite "> plain ware relets only to the line grey ware or includes all types pf undccoraicd Ninevite 5 pottery. In this article we will be almost exclusively rouccincd with the painted unci incised styles. As ive will see below some painted pottery which has been called Ninevite j more properly belongs in a pic-Nine vkc 5 "Late LVuk*1 or transitional phase.) 8 Christie 1920, 119-120, v\ Nivpj r.Kiuwi AFFAIR OF stvi.ks IM^^^^^B 201 Kuvnnjik. the citadel mound of an< irni Nineveh. Two plan- „, a single example ol an incised jar wen- | o i hi ishcr I.'' ' ' P11111™ |"»'"ry «fifJ The following season Thompson and Hamilton published [ mplete painted po.tc.y vessels .„:«• plate of complete in< is|.;, ^ fiv« P^ta of /opiates ol painted and incised sherds.'" I l.miilton «,,,,, ,i, , Zt a fur,,"r « ..vi^nllv mnlrmnnrarv" ili..nirl. I„. „......I___j ' . 11,11 ' two c laMfH Si the both com two are evidently contemporary", though he noted and (henTgNorrľfil "ľ 'W°.fl two types had an "uneven distribution over the site": in squares N O andTP" painted and plam examples urn- found, but few of the incised variety " I "U and V plain and incised sherds were found together h..i'n,i„'(' i " sf!uares >> i • "/"k lv i painted specimen! were rare and in squares Q and \ a large group of all three t vnes was found " " Furthermore he noted that in those areas where all three' types were fou f together "many of these fragments and pots wen- found in and belo* a serie" of rooms which we were subsequently able to date. In means of sherds and bn< ks found in the walls, both to the Romano-Parthian period and to much later " '* In the 1931 32 season at Nineveh Campbell Thompson and Ins wife urn- joined by Max Mailt )\van and his wiir. MallowaiTs main task was in rxcavatc the ere Prehistoric Pit from the top ol the mound down to virgin soil.13 In two months he had dug through 90 feet of stratified layers, most of it prc-Assyrian. and tor thr first time defined a succession of earlier cultures and poiierx styles. í ]<■ labelled these from the bottom upwards, \ine\iie L Ninevitc 2, and so on up (o Xinevite 5. In stratum 5, the painted pottery style and the incised potter) style were found together. Mallowan admitted thai lack i : tohereni building levels in the sounding and the presence of later disturbances m< i I thai the lines of demarcation between each stratum "remain ill-defined'1.M ' • rlheless since painted and incised sherds were found at the same level, he t ísum they w ere contemporary throughout the period. Mallowan also stressed all the incised wares of this period as well as the accompanying painted point • .. e suffii iently homogeneous to enable us to say that they belong to a single period ol culture".15 The sequence in the Prehistoric Pit also allowed a more precise- date to be assigned to N i nevi t e 5. Phe material from Stratum 4had similarities to Late Uruk in the south and an Akkadian st\le sealing was found in the upper part of stratum 5. Mallowan therefore suggested that Xinevite 5 per was "possibly to be dated to the earlier half of the third millennium b.c." ,h iod Tell Billa 1931-2 . v, „ . In 1931 and 1932 at the same time as the work of Hamilton and Mallowan the Joint American Expedition of the University Museum of Pennsylvania and of the American Schools of Oriental Research was excavating Ninevite 5 pottery in the lower levels (strata 6 and 7) at Tell Billa, about 25 kilometres north-east of Mosul The site has yet to be published in full but in the earlier field reports Speiser recorded that grain silos were found in stratum 7 and that "five of them were opened, yielding numerous fragments of plain, painted and incised pottery".'7 In the same report he 9 Thompson & Hutchinson 1931, Pis. XXXIII, XXXIV, XXXV.7. 10 Thompson and Hamilton 1932, Pis. LIII-LX. 11 Thompson and Hamilton 1932, 83-4. 12 Thompson and Hamilton 1932, 84. 13 Thompson and Mallowan 1933. 14 Thompson and Mallowan 1933, 130, 15 Thompson and Mallowan 1933, 174. 16 Thompson and Mallowan 1933, 133, 17 Speiser 1931, 13. 202 MICHAEL ROAK AND ROBERT KILLICK also stated that a tomb in stratum 6 contained "two large painted jars" and "three fine bowls with painted rims'* 18 and in the next year "Billah seven yielded also a large number of painted and incised parts oflarge chalices." 19 It is odd therefore to read in the reports on the pottery written in the same year that "the decoration is either painted or incised, depending on the age of the deposit ... the one occurs in stratum 7 at Billah, the other in Billah 6," 20 or again "the most characteristic ceramic type of this period is the chalice, which is at first painted (7), later on incised (6)." 21 Billa 7 and B were dated by Speiser to the Jemdet Nasr period, slightly earlier than Maliowan's proposed date, on the basis of the introduction of seals, contracted burials, and dubious pottery parallels.*- There may have been a dispute about the relative sequence of Ninevite 5 potterj but there was agreement that the painted and incised styles belonged to the same archaeological culture: "the differences were not so much cultural as chronological" and that Billa 6 and 7 represented "successive stages of one and the same culture".23 Speiscr also followed Frankfort who though! that Ninevite 5 "seems to be a late descendant of the Old Iranian HighJan I j 1 ■ ' and Speiser quoted parallels with pottery from Tejx Hi^;ir, < . n which was also being excavated by the University Museum in support oJ this poth ^>>.2A Tepe Gawra 1931 37 At the same nine that Spriser w :d at !« !! Billa, lie was also excavating at Tepe Gawra about 15 km away, and h u ■ there was confusion about the Ninevite 5 sequence. Only a lew pit < <> ol Nincviu 5 >tterj are illustrated in the final report and those from stratum IV were thought to b out of place.26 More types of Ninevite 5 pottery are discussed in tin text than art illustrated, but it is not always clear which forms are being referred to. Painted chalices, ribbed chalices and other vessel types with horizontal ribbing are all said to appear in stratum VIII-A.27 In stratum V II painted pottery continued while corrugated cups lost their popularity.28 Speiser initially reported that in stratum VI "the pottery is usually incised. The painted ware seems to have been confined to chalices." M But in the final report we read that "painting was absent in stratum VI except for two examples with crude black given by the officials of the museum in Istanbul, arc almost (mainly wrong" Frankfort 1932, 50). His second Jemdet Nasr parallel was a type of "squat pot with sharply marked shoulder and overhanging rim" found in Billa level 6 (Speiscr 1933, PI. LIII.6; but as Mal-lowan pointed oui this type need not be earlier than KD III and eould be later Malta wan 1964, 150 note 35). "Speiser 1933, 265. 34 Frankfort 1932, 50. W Speiscr 1933, 267. M D. Cross in Speiscr 1935, 60. 27 D Cross in Speiscr 1935, 41^*2. Sec below for ribbed CUps and ribbed chalices at Tell Mohammed Arab period 2* D. Cross in Speiscr 1935, 45 and 47. " Speiser 1931, 11. "Speiser 1931, 12. This might well have been an intrusive grave of the earlv second millennium b.c. "Speiser l932aT6. "Speiser 1932b, 6. " Speiser 1933, 267. " Speiser 1932b, 8-9; 1933, 266. The two arguments he uses are the appearance of a pot similar to Ninevite 5 painted ware "in the lowest layer or Fara", and therefore according to Speiscr in dK- Jemdet Nasr period, which was illustrated by Frankfort (1929, Tafel 431, h OPP- P- 86). Anothei Ninevite h put was also illustrated "i ri «r arlk"lr '1,Kl laWlwl ,1S rorni,,« lmrn L"«as}l *«CI ,. sty.ks Fig. 1. Map showing location of sites mentioned in the text. dots".30 Incised Ninevite 5 pottery is not discussed in the text and only one incised and one excised sherd are illustrated.1" Other investigations 1933 1972 Excavations of Ninevite b levels at Nineveh and Billa finished in 1932. Over the next forty years a few sherds and pots of Ninevite 5 styles were found in excavation and survey on a number of sites, Chagar Bazar, Tell Chuera. Mari. Tell Asmar, Ibrahim Bayis, Nu/.i, and in the Rania plain, and the known area of its distribution correspondingly increased.1'1' But on none of these sites were substantial remains excavated which could solve the main problems associated with the Ninevite 5 culture which remained a term used to describe the painted and incised pottery and little else. ,0 Cross in Speiser 1935, 51. Abu ul-Sonf 1%K. 51 Spciscr 1935, PI. LXXV1, 4; Speiscr 1929, Fig. 73. MICH Ml KOAI AND KuilKWI KII IH.K In \9tv\ Mallowiin published a review of ihr slatr nl knowledge about Nfticvite 5,:n 1 Ic and most other scholars who studied the question during this period 14 a^rrrrl with Speiser's chronological di\ ision ol a phase with painird wan- hrir^ followed hy one with incised and with no overlap bf die two st\ les. M alio wan also used thr result* of his excavations at Chagar Bazar to support this theory, In noting that painted and incised pots had not been found in the same graves*55 Furthermore a verbal communication from the excavator of Tell Bardasti in the Kania plain where a thin stratum with painted potters was said to lie below incised Ninevite r) sherds concluded his case.'6 The dale was still disputed/7 as also was the origin of the Ninevite 5 styles.38 A new development was the surest ion that Ninevite f) and I'ruk were closeh connected: in his publication of the results of surveys by the Iraqi Directorate-Genera] of Antiquities of I'ruk. Jemdei Nasr and Ninevite f> sites, Abu al-Soof Stated that "Ninevite V wares, including the painted variety, at both excavated and surveyed sites, are almost always found in association with I"t uk pottery" *r* and this view was accepted by his supervisor Joan Oates.40 Confirmation of this also appeared to come from Abu al-Soof s excavations at Qalinj Alalia when- three sherds identified as painted Ninevite 5 ware were found in Level 11 i and sherds identified as incised Ninevite 5 were (bund in levels II and III which he dated to the I'ruk period.41 Tell Vat Telul elh-Thalalhat 1965 1974 saw the publication of the first modern excavation of a mound undertaken specifically to soke the problem of Ninevite 5, that of Tell V at Telul rth-Thalathat, between Tel Afar and Mosul, carried 0U1 in l%f) bv a team from Tokvo Univers-ity.42 Here for the first time a complete building of the Ninevite 5 period was excavated with its associated finds. This building measuring some 18 by 6 metres was interpreted as a granary; it appeared io belong to a single period of tx vupation and a< terized bj Ninevite 5 incised warr, is contemporary ■ i lati PfHolitcrate in southern Mesopotamia and n nr hi less with Wnuicj CJ.*' mihiritics with tin poner\ oi'Tcpe Hissar levels I, • Pi rind Si.dk ill oil the Iranian plateau, such as •! painnil san at the end of the J. Nasr period, or perhaps in I D incised ware which succeeded it is likelv to hav< ha n introduced in EM II., and although we have to m kon with ihe possibilit> itiai sonw spei in,ens suia i\ c<\ ^ Li, as E+D. Ill, the period of its floruit was probably 1\ I > U; Porada 1965, 159 wrote "tht incised ware of Nineveh 5 has been tentatively taken bv me to be comcniporary with Early Dvnastie II m tin south. Whereas the painted ware belongs to Earl) Dynastw I possibly begiiiiiing at the end of the Jamdat Nasr Peru* , and on the chronolo^u al chart I7h Schwartz "the temporal gap between Stratum 16 and Strata 15 -13 . is the only apparent exception to an otherwise continuous sequence of occupations/ Sir.ua 16 40 comprise the Ninevite 5 levels called Leilan period III. In the sondage 7168 diagnostic sherds were recovered, of which 2651 came from the period III layers. On the basis of this sample, Schwartz stated that "we had achieved our aim - a continuous well-stratified ceramic sequence spanning the millennia of transition from late prehistoric to early historic times,"48 and that "the Leilan sequence demonstrates that fine incised ceramics executed in the same style as the sherds recovered from level V at Nineveh were in use for a very extensive period of time and underwent recognizable changes in shapes and designs during that time span. 'Ninevite 5' style painted ware is shown to have a somewhat shorter period of use, restricted to Period III. The Leilan sequence documents the development of fine incised wares through time in Periods IV-II and the appearance of Ninevite d style painted ceramics in Period III."49 Elsewhere Schwartz concluded that "painted ware did not precede incised ware in an early phase of the Ninevite V period but that the two varieties existed side by side throughout. " 50 48 11 is still not possible, however, even with this excellent publication, to work out to which levels the illustrated pottery belongs. 44 Fukai 1974, 66. 45 Schwartz 1982. "Schwartz 1985, Weiss 1983, 1985a, 1985b. 47 Schwartz 1982, 44. *B Schwartz 1982, 18. * Schwartz 1982, 240. The incised motils in Level I\ arc not distinctively Ninevite 5 (see note 88). In Level I la there are sherds of incised Ninevite 5 style (see Schwartz 1985, 57 and see below note 89). so Schwartz 1982, 223. 206 MlCHAKl. KUAI ANII Rf llif M I Kl! I ICR Jell llrak 1978 Hi One of the continuing goals of the recent excavation! at the nearby %i\r 0f jr|| Brak (like that of Tell Lcilan) has hern the Hucidaiion of the archaeological transition from the Late Uruk to the Akkadian periods and io "place the very distinctive pottery known as Ninevite V in a proper context."'« So far two destruction levels earlier than the administrative huildii^ of" Naram-Sin have been found and a part of the associated pottery assemblage has been published" According to D. Oates this material only included "a very small number" of painted and incised Ninevite 5 sherds which were "clearly out of context", and according to J. Oates "no Ninevite V pottery whatsoever" was recovered from the destruction levels,53 thus demonstrating that in north-east Syria at least the Ninevite 5 culture came to an end before the beginning of the Akkadian period.54 The early date of incised Ninevite 5 pottery as proposed by Abu al-Soof and Schwartz also seemed confirmed by the observations that in two areas at Tell Brak (CH and ST) incised Ninevite 5 pottery was found with Late Uruk pottery including bevelled rim bowls.55 As David Oates wrote there was "unequivocal evidence for the association of bevelled rim bowls and incised Ninevite V."56 Summary Thus there has been considerable disagreement about the relative chronology of the styles of Ninevite 5 pottery. In recent years, in contrast to Mallowan's and Speiser's division of the sequence into an earlier phase with painted pottery and a later phase with incised pottery, the evidence from Telul eth-Thalathat, Qalinj Agha, Tell Leilan and Tell Brak has been interpreted to support the view that the painted and incised wares were contemporary throughout the Ninevite 5 period. Clearly there are problems and contradictious in the published accounts.57 Several hypotheses suggest themselves but how are we to prove which theory is correct? "Oh, my friend, have I not said to you all along that I have no proofs. It is one thing to know that a man is guilty, it is quite another matter to prove him so. And, in this case, there is terribly little evidence. That is the u hole trouble* A Hercule Point, know, but I lack the last link in my chain. And unless Ixanjx the missing link He shook his head gravely.58 THE MISSING LINK Tell Mohammed Arab 1982-1985 This then was the situation when the British Archaeological Expedition to Iraq started excavations at Tell Mohammed Arab in the Saddam Dam Salvage Project 59 51 D. Oates 1982b, 67-8. 52 J. Oates 1982. 5J D. Oates 1982b, 67; J. Oates 1982, 207. 54 Recent work at Tell Brak has suggested that "even the latest examples [of Ninevite 5 pottery] significantly mo'c^10 lhc Cnd of ED ln" 0- Oates 1985a, 175 or Ijoq, 248). "Ficldcn 1981, 157; D. Oaies 1982a, 194 and 197 L hT ° *associal«' wiih Ninevite V pottery md I Oail ,o£ orls"); D °il,cs 19«2b- 67 and 68; %SSL t ' 207" ?1S° °- 0a,cs 1982a> '91 incised "tnevue 5 pouery found in a supposedly immediately post-Jemdct Nasr context. 56 D. Oates 1982b, 68. . 57 Readt- 1982, 72 wrote concerning the pottery M the third millennium "there seemed to be anomalies and complications, in the published evidence, which obstructed attempts at relative dating." "Christie 1920,235. . « Undertaken by the British Archaeological rAJx dition and the Iraqi State Organization for Antique . and Heritage directed by M. Roaf. The excava«on * the Ninevite 5 sounding was the respons.biim rly R. Killick. A MYM i Hint s AH All* c>|. s|v|.k.h The preliminary results have already been publish. .1 -" , ationofthesiteisin progress. Only the- evidene <• of ilx lour, |F"'bli* Along the cliff section and in the Stratigraphic sondage il/rre dSSf"''?!^ occupation were recognised. These periods were separa.ed from one ,,"',7"''V' stratigraphic breaks in the sequence of occupation and each period <\j i !'T, V differences in its material culture. pCnod (,,sPlayrd r|ra, The earliest period (Mohammed Arab period 1) contained pottery related t I which is usually dated to the Late l.t.k period. The assemblage lude, l^td rim bowls, incised four-lugged jars, spouted jars and ring-based carinated bowls*' There was also a painted ware: lour-lugged jars and carinated bowls were painted normally with extremely simple geometric designs .Cross-hatched triangles applied only to the upper half of the pot in red or red/brown paint (Fig. 2). At the end of period 1. there was a period of abandonment which was followed bv a Ninevite 5 occupation Mohammed Arab period 2) with five distinct building levels. These levels contained painted pottery of typical Ninevite 5 style, indistinguishable from that found at Nineveh and at Telul eth-Thalathat. Characteristic vessel shapes include pedestal jars, pedestal bowls, and large hollow-stemmed bowls painted all-over in a variety nfgeomrrri tlMd naturalistic motifs Fig. 3 . Unlike the earlier painted style of period I which was normalh wheel-made in a pink to buff clay with grit temper, this painted Suit ite 5 pottery was often in a hand-made common ware, with a predon; a _ .able temper. In association with the pai N was a series of fine-ware cups and pedestal bowls in grey/green lahri no visible temper. Many of these pots were ribbed or layered on the outsidi and mi of these had notched horizontal bands; and a very few from the latesl bui % levels of period 2 had simple, lightly incised patterns (Fig. 4).62 The run;:- ol p itti rns on these sherds is very limited and compares closely with the Thalathal assemblage. The elaborate incised and excised designs familiar from Nineveh and Biiia are completely absent in this period at Tell Mohammed Arab. At the end of Mohammed Arab 2. the site was again abandoned and. after a period when it was used as a cemetery for people burying incised and excised fine-ware pots in their graves, was resettled by people who used fine-ware incised and excised pottery (Mohammed Arab 3). Painted pottery was not manufactured in this period. The incised potterv of Mohammed Arab 3 displays a much greater range of patterns than the few sherds from the earlier period. The decoration is also often cut as a continuous band of incision around the pot and many sherds are excised, with the clay removed to leave the decoration in relief (Fig. 5). The fine-ware assemblage also differs from that of period 2: the pedestal bowls and cups are not usually ridged or layered, and show a preference for a hemispherical shape. 60 Roaf 1983 and 1984. a" ' arr thc ,at('SI luo 1<"vr,s °' tXTi(Hl " 41 ui 61 Roaf 1983, Fig. 2 and Roaf 1984, Fig. 8. Forfurlher not earlier. As they arc stylistically different from the discussion of thc date and terminology see above note 5 "'late ineised and late excised" ol Mohammed Arab and below notes 101 and 102. period 3 they probably are not the result ol "pottery " 15 sherds of this incised ware have been found in drift" (see below), reliably stratified Mohammed "Arab period 2 levels. In AA A A A A A A A A A A A A © 0 Fig. 4. Early incised Ninevite 5 pottery motifs from Tell Mohammed Arab period 2 This development from an earlier phase with painted Ninevite 5 pottery to a later phase with incised and excised Ninevite 5 potter) is demonstrated by a sherd count of decorated sherds found in the main sounding in 30T-51V Fig. 6 ,63 The results complement those obtained in the excavation of the cliff section,64 In this table, some painted sherds appear in id /? layers: these are residual sherds from the earlier layers. T his is the respe* direction for sherds to move. Occasionally later intrusive sherds are found with earlier material.65 This "These figures are provisional: the final publication of the results of the main sounding is in preparation and it is possible that some changes to the phasing may be made, 64 Compare the counts for trench 51V of the Cliff Section published in Roaf 1984, 151. For a section showing the levels see Roaf 1983, Fig, I (lower) and for the plans of the levels see Roaf 1984, Fig.7. Subsequent excavation in the stratigraphic sounding directly to the north has shown that small modifications should be made to the stratigraphy and phasing in trench 51V but that the main conclusions arc valid and are confirmed. * The reasons are many: a workman may put a sherd m in " - fay, a pit or other intrusive feature m,PV " ;' h"! 1 been recognised, the sherds may have moved thi animal action in the soil, they may have got mixed in washing or, in sorting, the wrong number may haw been written on the sherds. Normally such a mistake is caught before the finds are published and is then suppressed, but occasionally it mav be unwittingly published. The two Ninevite 5 style shcrch in level L may be a case of sucii intrusive pottery drift. The scale of the problem of residual ^hcrd$ is illustrated at Mohammed Arab by the presence of several hundred Ninevite 5 sherds in the Middle Assyrian levels. -rr-r 1 h 5 's > > JO O r Fig, 5. Late incised and excised Ninevite 5 pottery motifs from Tell Mohammed Arab period 3. 212 MICHAKl. KOAF ANIJ ROHI Hl ku.[ Ick ,tu)11, lhe possibility_ of of "pottery drill" in rhhe, Him (..... |n ini,u] when considering the results of any rxravaii.ni. r" kf. "WJWfA U o wry f>r fl very had explanationf" remarked I>otm „ everything, and explains nothing." 66 ' «w*rf The evidence from Tell Mohammed Aral), therefore, supports the contentio painted Ninevite 5 pottery was more common in the earlier part of the Ninev, ^ period and incised more common in the later part. Except for a few residual sha painted pottery died out before the later part of the period and excised and types of incised grey ware were only introduced in the second part of the period"051 But Mohammed Arab does not have tl ie complete Ninevite 5 sequence Th observed stratigraphic gap between periods 2 and 3 and the abrupt change it pottery styles suggests there may be another phase, where perhaps attributes of both the period 2 and the period 3 assemblages are combined. We can also distinguish at Mohammed Arab an "early incised" style, found in period 2 and a "late incised" style found in period 3. Now is the time to gather together all the suspects and confront them, to review their stones and to try to reconcile all the various statements, which up to now have been more than a little mixed up. "The mind is confused? Is it not so? Take time, man ami. You are agitated: you are excited - it is but natural. Presently, when wt are t aimer, we will arrange the facts', neatly, each in his proper place. We will examine a • ' I. Those of importance we will put on e in PERIOD Pha se MA 3 Cemeterv MA 2 MA 1 A B C D E F G H I .1 K L M Excised Incised Ribbed* layered Painted 10:22.7 30(68.2) 3(6.8) 1(2.3; 54(50.9: 33(31.1) 12(11.3) 4(3.8) 21(42.0) 12 24.U 10(20.0) 6(12.0) 41(53.2) 21127.3 6(7.8) 7(9.1) 37(17.1) 109(50.5 39| 18.1) 24(11.1) 2(66.7) 1(33.3) 5(14.3: 1 2.9 7(20.0) 22(62.3) 7(5.0) 12(8.6 33 21.6 87(62.6) 6 1 l.d 37(86.0) 1(3.7' 1(3.7) 25(92.6 3(11.5) 23(88.5) 1 (50.0) 1 (50.0) Total Total number ol'clecorated sherds 44 106 50 77 216 3 35 139 43 27 26 2 0 768 M^hammcdArabT °fdeCOratcd Nbevite 5 potterv from .be 50T-31V sounding atT*J parenthSesafe^r CmmUngPainted and ^faed "Late L ink" pottery). The hguu•» *3UffiBff* St,lnUmbt>r <*«™™ri Ninevite 5 sherds in each level. 4 ntrds ,n Mohammed Aral) period 2 include notched bands. Christie 1920, 156, MA 3 CEMETERY MA 2 B D H A MYSTERIOUS AFPAIR OF s, VI„S EXCISED INCISUJ NUrCHED BANI) K IIHK 0/ I AYI Nf D 213 I K 0 50 100Ä 1 i_l Fig. 6b. Chart of the percentages of decorated Ninevite :") sherds in the 50T-5JV sounding at Tell Mohammed Arab. Level 5 is shown in outline because there were only 3 decorated sherds. one side; (hose of no importance, pouff" - he screwed up his cherub-like face, and pujjed comically enough — "blow them away!" " That's all very well," I objected, "but how are you going to decide what is important, and what isn't? That always seems the difficulty to me."61 " Christie 1920, 55. For onee Hastings seems to have hit upon a major problem, and one which we do not attempt Icj answit ^ MICHAEL ROAF AND KOHF.KT KIl.t.ICK THE RE-EXAMINATION OF THE EVIDENCE OF THE KYK WITNESSES ^ThfpubHcatioii of Nineveh is sketchy bring subjective and selective. Most of ,u illustrated pottery is effectively unstra.if.cd and really only Mallowan s PrrhiSUlri( Pit offers the opportunity for straligraphic analysis and here too d,<- choir,- ,lf wh| . nots were published seems to have been fairly arbitrary. rurlhermore as Mallowan later admitted the Ninevite 5 levels "ai Nineveh were confused"" and there was considerable stratigraphir uncertainty in the excavation so that it would be unwise to rely on the evidence of any particular piece of information. Nevertheless when the published sherds from the Nineveh Prehistoric Pit are tabulated according to their depth below datum (Fig. 7), the pattern that emerges is similar to that from Tell Mohammed Arab given a certain amount of "pottery duff.69 The "uneven distribution" of painted and incised Ninevite "> potiery found by Hamilton should also be remembered and certainly gives support to the sequence found at Tell Mohammed Arab. The material published from Nineveh seems m include examples from throughout the Ninevite 5 period. Some types from Nineveh such as painted ring-based and round-based bowls are not represented in the potters from Telul eth-Thalathat and Tell Mohammed Arab 2 and 3.:" These type in la< t. belong to the "Late Uruk" and "Transitional" phases (see below . Tell Billa In the first reports on Tell Billa it was painted and incised "Chalice ware" or "Billa ware" sherds were found I grain silos belonging to the earliest level (stratum 7 , though no mentii . ;ts made in the later reports on the pottery. Was there an intermediate pin • upation at Tell Billa, and the evidence later suppressed, or did Speiser eorre< I .(curate field observations? It is clear that the published material limited t is does not all belong to a single cultural assemblage: in particular some of the graves and other pottery in level 6 appear to have belonged to the Akkadian period or even later.71 In all only 29 incised sherds said to come from level 6 and 1 3 painted sherds said to come from level 7 have been illustrated. Of the incised sherds 3 have earlv incised motifs. 13 late incised motifs, 9 have excised designs, 1 has a ro« ker pattern probably very late in the sequence) and 3 have uncertain designs. This ini-lit suggest that the sequence at Tell Billa was very similar to that at Tell Mohammed Arab. But whichever version o Speiser's story one believes the Billa strata can be fitted into the Mohammed Aral) scheme. Tepe Gawra rJ*2C r?Uy' the excavation of the lower strata of Gawra seems to have been more reliable than that of the upper strata which clearly contained a mixture of pottery o ^SSSSiT 148 I" Particular, "Thompson * «TS?bS»5 S Sll^r1 UII.3,4,6, IS and PI. L1V .1. 5 and 8. Str bclo* from high,- ! SJtJ*Md mis>alHll.d. or tallcn in 108. wrongy attributed to that tmdspot. A MYSTI-KKurs A I-1- AI J< oi S|YI.,;s + 8 + 7 + 6 + 5 + 4 + 3 + 2 + 1 0 — I -2 -3 -4 -5 -6 -7 -8 -9 -10 -II -12 -13 -14 -15 -16 -17 -18 -19 -20 -21 -22 Extisfd Incised Nolchrd hanil ( rnpt p(ittfr„ ) Ribbed or l^ayertd 2 3 I'awiftf 1 2 3 1 2 1 4 7 2 13 9 2 9 21 I 9 21", •including Painted Nincvitc 5, Transitional, and "Late Uruk". rig. 7. Frequency of published decorated sherds from the Prehistoric Pit at Nineveh at d i Heren t depths in feet. different periods,72 Thus Gawra VII and VI contained painted Nincvitc 5 pottery and also much later Akkadian pottery.™ Because of the close* proximity of the two sites, it would be surprising if the Nincvitc 5 sequence at Gawra was different from that at Billa. The complete Nincvitc 5 sequence is not represented at Gawra so that a gap in occupation at the site during this period has to be suggested. Other discrepancies which we cannot now resolve may be ascribed to less exacting standards of excavation and recording. 7a Although Spciscr would have disagreed with this: "To establish the individual strata with precision is the principal task of the excavator. Now, if this is to be accomplished successfully, the site must be dug level by level, as has indeed been the cast- with the first ten strata" Spciscr 1931, II). " Reade 1968, 245. 216 MICHAEL ROAF ANI> ROBKRT KII.I.ICK 01M Chagar Bazar, levels 1 and 5 of the Area M "Prehistoric Pit" soundiiur^ considered by Mallowan to belong to the Ninevite 3 period - N,arl, *, ; illustrated material from this area was found in graves and thus its use for dating the levels as a whole is limited. The painted pottery from this area seems to be a local style, in imitation of painted Ninevite 5.75 Incised pottery was found in two graves and displays patterns which link it to the early incised style of pcrjod 2 at Mohammed Arab and of Tell V at Telul cth-Thalathat (Fig. 4).7« However, the presence in one of these graves of an excised jar and early incised pottery suggests that this grave at least is later than Tell V at Telul cth-Thalathat. Since the material from Tell Bardasti has not been illustrated one cannot tell if the later level above the one with painted Ninevite ä wares contained early or late incised Ninevite 5 sherds. The association of Uruk and Ninevite 5 wares on survey is illusory.77 Even an the published statistics only 20 sites out of 1 17 link sites and 85 Ninevite 5 sites had both Uruk and Ninevite 5 material on the surface. Furthermore some of these sites had Early or Middle Uruk material or incised or excised Ninevite "> material and were not occupied during the transition from Uruk to Ninevite 5. The sherds identified as painted Ninevite 5 from Qatinj Agha do not form a significant component of the pottery assemblage of that site and it is quite possible that they are examples of "pottery drift" oi been wrongly identified. The incised sherds identified as Ninevite 5 are probabl • i i\ pe of decorated ware found in Gawra XII and XI and in Earlv U ink l< . - ' i! link.78 J Telul eth- Thalathat At Telul cth-Thalathat the excavations pi • ' > i pottery assemblage which appeared to belong to a single period, will ort time-span. It was clearly demonstrated that painted Ninevite 5 pottei and me forms of incised Ninevite 5 pottery were in use at the same time. The pattern n pertoire of the incised material is similar to that of period 2 at Mohammed Arab and the incised patterns typical of Mohammed Arab period 3 are not found. As in Mohammed Arab period 2, excised jars and bowls were not found at Thalathat. Thus l'halathai supports the distinctions we have made at Tell Mohammed Arab between earl\ incised pottery and late incised and late excised pottery. The excavators with commendable thoroughness seem to have published all the decorated sherds with recognisable motifs that they found and it is therefore reasonable to look at the relative proportions of different tvpes and motifs. There appears to be a high percentage of incised to painted sherds at Thalathat when M Mallowan 1936, 11. "Mallowan recognised that they were different from "L T cxamP,cs. jesting that thev represented the degeneration of the older Tell Halaf technique" Malign 1936 39). A hand-made incised pedestal STlTS T^' Carly imiscd Nint,vi,e 5 design «„ Fig 8 8)' 3 l0Ca' Pr°dUCt (Mallowan 1936, Cti7 and V,m. Mallowan 1936. 58 and Figs. 18 and '"See Kitlick I"86. 231; Weiss 19836. 327. 3». °" Abu al-Soofs own Injures this claim is not subst.inw Toblcr 1950, PI. LXXlXa (for reasons forsugg ^ ing that this stamped and incised ware also mi q^ level XII sec Pnrada I9t>5. 145 and I K- '\ 1 1985a, PI. XXXIb ( =J. dates 1986. PI. 7)- A MVS II HUMS All AIU OF STYLE! Excise 0(0.00) hu'lyfd Motched band ( rope pattern) or layered 45(21.1) 12(5.6) 8(3.H) Total number ol dct-orated sherds 213 PctUlttd 140(69.5) Fie 8a. Frequency of decorated wares according to counts of published sherds from Tclul Cth-Thalathat Tell V. The Inures in parentheses arc the percentages. 21 of the „,< ISed sherds include notched hand patients. EXCISI D I NCI SID NO 1CIII I) R1HMI)/ HANI) I AYI Kl I) HAINII I) Thalathat 5 L 'A) 10(1% Fig. 8b . Chart of the percental s old ited sherds from Tclul eth-Thalathat Tell V. ... ■ 9 V:iT a," Furthermore there are some compared with Mohammed Arab | 1 w ■ hi> a[ Tt.lul new'types which are not found ■ • - N »M , eth-Thalathat may therefore be 1 1 man TellLeilan • in , tirnch at Tell Leilan In his analysis of the pottery from th<- Opc.a 11 . ^ ^ sufficit.„t sample Schwartz grouped the strata into 20 phases in oroe _ , is based on rather size from each phase- and then b) means o a comp >^ fivc ^ ds confusing idiosyncratic typological traits - (,u,(1rl was sub-divided by the (Leilan II-VI). Leilan III, containing Nmcviu 1 I |(K nlh phascs 4 b, computer into three periods. IIla (phases / 8, strau ;' 83 fi 9 ^ the strata 21-34), and IIIc (phases 2 3, strata lb 20) the 79 If the illustrated sherds represent some unicprescn-tative subsample of what was actually found at 1 ell \ ai Telul eth-Thalathat, then these statistics have hide value. The similarity in incised motifs with Mohammed Arab period 2 still remains valid. 80 E.g. Fukai 1974, PI. LI, 6; PL XX11L 3. 81 Schwartz's tables list the motifs by phase and u< t h\ stratum, so we have had to follow this in our study. I Ke excavated Level III deposits in Operation I totalled 85.5 cubic metres (the equivalent of one 5 by 5 metre trench excavated to a depth of 3,5 metres). 82 Idiosyncratic because his terminology is dihYrcnt from that of other scholars and because be does not use a typology based on the shapes of whole vessels in (act he claims ihis as a virtue]. Confusing because the types and motifs and their ranges are not defined. Furthermore a RlOtif such as a horizontal f)ainutf band of unknown width is given the same importance in the computer analysis as quite specific patterns such as lined zigzag or again notched bands are considered sigiiiiicant only i found on a sherd without other motifs. [here is little justification for this tripartite division in the published material and since Schwartz's typciogy is an unusual one it is dtfhcull to use his tripartite division. For an attempt to impose the Leilan tripartite sequence on the whole of the Sincwtc 5 period see Schwartz I9i\5, Tables IV on p. 57 and Table VI on p. 59. when »3 'JIH MICHA i I Mi) a v AND KOHI-.R I KIJ.IJiiR PERIOD phase strata Excised Inci Kfd* Uhhed bandf lithhed or (rope pattern) Lajmd nu- 2 16- 18 12(13.2) 62(68.1) 40(44.0) 8(8.8! 3 19-20 22(15.3) 87(60.4) 78(54.2) 5(3.5) 4 21-24 1(1.8) 24(42.1) 17(29.8) 1(1.8) mb 5 25-31 48(23.1) 33(15.9) 6(2.9) 6 32-34 10(28.6) 2(5.7) 3(8.6) 7 35-38 19(30.1) 8(12.7) 5(7.9) Ilia 8 39-40 16(48.5) 4(12.1) IV 9 41 -42 17 0 10 43 44 9 0 5(5.5) 12(8.3) 47.0) II (5.3) 14^40.0) 21(33.3) 15(45.5) 58 49 total number of'dciorated motifs in period III 4(4.4) 18(12.5) 10(17.5) 143(68.8) 8(22.9) 18(28.6) 2(6.1) 7 91 144 40 208 35 63 33 5 614§ ♦Calculated by taking Schwartz's distribution of incised motifs and subtracting excised "panelled" (motif6) and "grooved" (motif 11)), notched bands (motif 3). and ribbing [motif5). The first figure includes incised motifs which may not be Ninevite 5 such as horizontal lines. The figure in italics is calculated taking just Schwartz' motifs 1,2.7.8,9.10.1 Land 12 which are almost certainly Ninevite 5 (sec note 84 for further details). fOnly those notched bands with no other motif are in< ludi 1 ■ {This figure includes all painted sherds not just Nincvitr 3 ones. These arc sherds or pots not the number of pots. §Apparcntly 614 motifson563 (481 incised of all sort an >ainted sherds. Fig. 9a. Frequency of decorated motifs in pei iods ] V of Operation I at Tell Leilan. The figures in brackets are percentages of th< . :< , orated sherds in each phase. fXCISFD [NCI 51 rrjCD RÍBBFD/ I AYI Ri I) PAINTED flic Illb Ilia 8 0 50 100^ I_,_i * incäd <>™«:<,rated motifs in period III of Operation 1 at Tell a moots which may not be of Ninevite 5 stvle are shown in outline only- rm frequency counls lor (he dillovnt categories of figures arc plotted out as a frequency < hail as we havr'do,,,. j'" .Vhrn U,rv \rab and other sites, it is clear that Leilan phases B 3 !„-|-- Mf»tamrnrf| than period 3 at Mohammed Arab. The excised &^7f^SZ^^ "grooved" by Schwartz) found m phase H indicate that this stratum mav be lan' than Thalathat, winch had no excised pottery, and show the cxis.rnr, ,!r . Ju when painted, incised and excised wares were in use at the sameM^m ^ * In the latest phases of Leilan III there is a particular form of cxdsed w by Schwartz "grooved", which combines shallow vertical channels with 2 patterns in the early N.ncv.te o style.87 This material, missing at MohammH \rtb may be an early form of excised Xinevite 5 pottery. ' ' Schwartz has not given convincing evidence to'support his assertion that incised Ninevite 5 pottery preceded painted in Leilan IV.H8 But evidence from Tell Leilan itself contradicts his assertions that the 'painted ware did not precede incised ware' and 'that the two varieties existed side by side throughout', for in levels of Operation 57F02 excised and incised \in< te 5 sherds were found without any painted material.89 Schwartz, however di iol discuss this pottery as a Xinevite 5 style but treated it within a general dis< period 11 [lottery.1*0 It seems more likelv that the pottery from Operation n sponds to the later part of the Xinevite 5 period roughly contemporary Mohammed Arab period 3 when incised and excised pottery were in use 1 ■ • ted style had died out/" 91 •♦From Level III oi <>,„...u->» I sherds (not neeessarilv of Nmevite ' -incised sherds with 532 mnt.ts in. I ■ sherds, whieh many would not . l.-r including 107 examples not cl.su... live ul it». « • ^ vite 5 ware as known from Other sites tnc*c motifs 4 (horizontal lines 14 oceurTcnces I » <*» (fingernail rounded 9 and fingernail other M . . vertical applied bands 9. 16 OiDtifc with dim vrru lines 4), 17 (thumb impressions 3 and 19 38). These motifs are not illustrated by Schwartz j some prohahly an- Nineviie .'). but his d,M"ptfV!ic ,,, notsulficicnt and there is no indication ot die la J ' the pottery. What is meant bv outliers is uncertain these might be sherds that Schwartz himselt mogn _ as being residual or intrusive. Eliminating i >p< . _ I I" '.W ClCC U T " ribbed and notched bands gives a <>» — rences of incised motifs, i 81 Although the percentages of painted sherds in i upper levels arc low and some might be lh<)UghMtM< ^ the presence of complete pots in a grave u* tC ^ Schwartz Pit I) of stratum 19 shows that thev wrr*' * n being used as laie as Leilan 111 phase 3. A < residual. S.l.u,n/ 1985. table [[ on p. 35 show, a similar dcw-l.tFmlrvli. hm the duYcum tvpes of incision and rXt -in- not separated out. Schwartz 1982, Figs, 38,5 and b; 39.23; 40.6 N Imam 1985, Fi.g. 5.23. MThe uuurd sherds found in Leilan IV, which (man/ tailed mused Xinevite 5. were 9 sherds with ln»ii/f»nul tibbing and \2shctd» with hori/ontaJ lines, a motu admitted In Sc hwartz to occur in all periods S< hwartz 1982, 81 . .NVit/icT motif is put in ukirh diagnosis for incised Ninevite 5, thcujtrfi they might be if thev were in the typical Ninevite 5 (jjrey ware: Schwartz did not indicate what tin* fabric of tliese shcTds was. w Schwartz 1982. 223. In >u\iui 4 7 of Operation 57F02, which have been assigned to Leilan IIa by Schwartz 1985. 5b 57 . lanje numbers of sherds with lined zigzag, plain vcTticaJ zigzag Schwarte 1985. Fig. 6 , slashed in< ision and junel incision (Schwartz's term lor a tvpc of excision were found in strata 4- 7. So many that Schwartz wrote that "gaming popularity in late Period III, thev attain their maximum popularity in early Op. 'y7F02 Period IV (Schwartz 1982, 188). -.Schwartz 1982, 186(T. ----1 Kv MICH Ml KM AT AND MOItl'.ia KM. IK. II il r v in n beginning of p Leilan III, therefore, documents very clearly the period of transition from r* to incised with the ratio of painted to incised changing from more painted incised in Leilan phases 6 8 to more incised than painted in I.eilan phases 2 4 Clearly the Leilan Operation 1 sequence, contrary to what the excavators asserte 1 does not span the whole of (he Ninevite 5 period but only the middle part of it with large gaps before and after.93 Tell Brak The problem of "pottery drift" due to residual sherds and levelling fill had been recognised at Tell Brak in a number of areas early in the recent excavations.94 This explanation has now been adopted for the seemingly early occurrences of incised Ninevitc 5 with Late l/ruk pottery." The pottery which provided the "unequivocal evidence for the association of bevelled rim howls and incised Ninevite V" has now itself become "not unequivocally stratified".'*6 Until recently very little Ninevite ;> pot ten from Tell Brak had been published. But now 7 excised sherds, 4 incised shci I I painted sherd have been illustrated in drawings and a further H incised and ci: sherds are illustrated in photographs.97 The painted sherd belongs I -iiional "Late I'ruk" Ninevite 5 style rather than to genuine painted Nil Vmongst the incised and excised sherds there arc some examples of tin ates 1982 s, ; , . rk ,v n- r"-- " ,s not always appreciated *fr> ^cadmgjy ^.raiihYd'V B (hues 1982b, 70: ■ Wl vvithu, the stnn tun contained a nmssofUte ^t«.li S"mr N,*eWt* V sherds, but the surface 'a, 17b or J. Oates 1986, 249). And for area ST: '1r ar(,m'«ectural evidence and the Strati^ti LATE EXCISED NINEVITE 5 EARLY EXCISED NINEVITE 5 EARLY INCISED NINEVITE 5 PAINTED NINEVITE 5 PAINTED TRANSITIONAL "LATE URUK"/ NINEVITE 5 PAINTED "LATE L'Rl'K." INCISED "LATE L'Rl.'K" "variously tailed Laic K.uK Dvn ill ilan II. etc. in the Habur. tig. 11. Decorated potter) I period terminology. Painted "Late Uruk'" and Incised-! Uruk" arc thcwlic?! <-t"the styles under discussion. They precede the introduction ol Ninevite 5 material and were found in period 1 at Mohammed Arab with pottery n lated to that known from Nineveh, Tell Brak, Habuba Kahira, I'ruk itself, and other sites dated to the Late Uruk period. However, the forms are degenerate when compared to the international types of the Late Uruk period.102 9fl ^ J. Oa.es 1986, 251 252. For example Figs. 5.98 and ** It is also possible i«, distinguish «haugc* in the underrated pottery. For example ,(,«• ribbed/layered «'k- ware forms of the "painted and early mcised" P nod are superseded by plain hnc-warc equivalents in 11»! r lHU' indstd*' ^ri"d Or again hl^^semnu-d bowls are found only in the -painted nou-dtnIhnfi,SCf rri°d- Furtht'r ™-"P'« U*££* ^^-'>'-tion of the excavations a, Tell loo •pu earlier thin Th^ CVi^"CC tha' this "8«»ved" ware is examp e of" w^,'"1 'Td " Ldlan ,hw " 1 Polled ( = Late Excised) in phase 4 and 7 in phase 3, while there are nu examples of' ■ rt>,,'r ' - . i_ •} i> flues howoer. v%ai<- in phase 4 and Id in phase j. II , ,* f " it Mohammed Aral, or a, Telul e.h-Thalathat - " " al IHI The distribution outside the Eski Mosul area of these "Late Uruk" and Tran sitional styles ,s not vet clear. The Panned "U,r Iruk" stvle can ulemified on other sites in northern Inu,."- perhaps in south-eastern Turkey «« and on one site in n,°? Vr7"»Syri? 'T'r <,fthe InfiSCd criod III at Leilan which contain painted-wad excised pottery beloi .. In nal pari of the Early Ninevite 5 period or the beginning of the Late \ . . iod. There is no ob\ i< u the Painted Ninevite 5 ware (as there is in the incised). But furth well show both chronological and regional variations in the st\ le. The following stag) ") period, when Incised and Excised pottery were in use but Painted v. , was not, was found in period 3 at Tell Mohammed Arab. The Late Incised and l.at< I xciscd Xinevite 5 styles appear to manifest 103 Falcs et alii, in press; Fujii it alii, in press 104 Nineveh [Thompson ami Hamilton 1932, PI. 1.111 15); Tipe Gawra , Tohle r ITiO, PI. CXLI. no I 19 . 101 Bchm-Blam-kť IOH!. Abb 2l.la-b, 23.6. 106 See Meijer 1985, Fig. 24!> for one Painted Late Uruk sherd from Tell Q_ara$a. I lie published painted sherds from Tell Brak and from grave (JI88 at Chagar Bazar arc more in die transitional style than in the painted Ninevite ") style J. Oates l'JBI). lit; ">.107; Mallowan 1937. Figs. 25.1 and 25.3). ""J. Oatcs 1986, Fig. 3.46. 108 Painted ring-based bowls from Nineveh are similar to examples from "Late I ink" Mohammed Aral) period 1 (Thompson and Hamilton 1932, Pis. I.III.15 & LIV.5). Other bowls appear closer to the |>edestal shape characteristic of Fainted Ninevite T> hut retain the "Late Uruk" geometric patterns, these may he classified as Transitional (ibid. PI. LIV.l and K . Round-bottomed cups with geometrii- designs are also "'Late I ink' I hompson and Hami lum. PL LII 1.3.4 & 6 . This shape in not known in Paintťd Ninevite 5 ware at Mohammed Arab su where it occur* at Nineveh with naiurali>ni patterns Thompson and Hamilton, PILNÍ, 8 and 12 , it may also represent an aspect of the Transitional sfvle. See also note 10b\ '« The distinction between the painted designs from I rlu! eth- I halathat and those from Nineveh which was pointed t.nt 'a i he -Japanese c \cav ators Fukai 1974,6; ami Fig. h on 102 of the Japanese text is not a regional variation but .* chronological development. The Nineveh sample includes examples ol the earlier Transitional stvle, where painting is usually restricted to above the carination see above note 70). I he dlffiClUQ in recognising * development in Painted Nmevite o potterv is perhaps due to the small size of the sample rather than the homogeneity of Painted Nmevite :> pottery from different sites. . MICIIAKI. ROAI- AND Roll Kiel KILLK N regional dillere.iees. The variety and f|ual.ty ol the ,.;,«• |,X( is|lfj an pottery from period S al Mohammed Aral) and siraiurn I, ,,, j,,, ,/.'" -rlleeted in the limited repertoire of the Hahur material. In t m it t|,k i ' '* a strong regional style not louud to ih<- r;tsl no P°t / style in Fig. I I is called Incised Taya IX. This Myi ire dealing primarily with decorated pottery siylrsVi " e Incised Ninevite .r) style and has in the past been m' 'l C'c.arly here are differences in vessel shape and suriac, 'S' cntlf"f) i- * ■ «11 IV- . m,d ,„ i|,r r.lsl no \*>\u-x.; The latest pottery style in Fig. I I is (ailed Incised lava IX. T|,js ttyu i chosen because we are dealing primarily with decorated pottery stvlr, i, :'''S1,J"" derived from the fin as the latter 111 but there -.........-................------l< ^ there is no dear c\ ulcncc to suggest thai < - exerted a definite influence on the other. A IIYITBRIOUI AFFAIR <„. V| vi ».s NORTH IRAQ Mohammed Arab _„ Tell Billa * ;~ —2 Thalathat Tell V 7 18 |» 3 2 1 6 Nineveh MM ?8- V KaranaS u U-18 HABUR Tell Leilan Op. 1 -13._I 5 57F02 -2-3- -4 7- Tell Brak ch 5 8 Chagar Bazar graves g8 ^7 DI Y.ALA skqik.nck Akkadian i vRivnvwa-if ! V i ''' N A S I 1C. PROTOLITERATE ?16 1» ] ineviti and also a plain fine-ware cup from level IVb at Gubba falso Earlv nasiir ! \\ Mari. three incised sherds two Earlv and one unidentifiable : a singh Ex< isc-c 1 ^herd^120 and two "metallic-ware" sherds121 have been found within that pari ol the Trench B sequence dated to the 1,6 Killick 1986, 232 3; Porada l!. MICHAKI. ROAI- AN!) KOHKK I Kill M.k i; irlv nvnastic I period. This would surest lhal both "mrtallfe-ware" aru, , ,.M isrd Ninrviir 5 phase .started as early as Karly I)y„;,su, I. The r,,a ' 1 U, Ninevite 5 fine wares and metallic wares is still ,lf,i (rrtain h(j| "talgal convincing evidence lor any overlap between the styles. The rvidm,,. ,n ''■ correctly dated and if there was no intrusive pottery drill, would imnlv a1** correctly dated and if there was no intrusive pottery drift, would ,m„iv th Ninevite 5 period ended before the end of Early Dynastie I. Bm thm, ■[ * hd' the evidence to date the Trench B sequence and few pottery parallels with otherl,!^ As Lebeau remarked "hormis tin seul lesson de "cut ware", la ceramiquedyn *' archaique I de Mari n est presqnYn rieu comparable a cellc du sud mesotLtam^ de la basse vallee de rEuphrate," 123 This lack of strong connections between M*' and the south and between Mari and the north make it an unreliable gu;tj*. r" determining the relative chronology between north and south. Two Ninevite! sherds have so far been reported from stratified contexts at Tell Chuera,124 but since the chronology of Tell Chucra is disputed the value of these sherds for elucidating Ninevite 5 chronology is dubious.1*'5 8 The duration and absolute dates of the period of Ninevite 5 pottery remain uncertain and the absence of any epigraphic evidence from Ninevite 5 sites126 and the apparent unreliability of Carbon 14 determinations lor this period 127 make it impossible to propose any chronological scheme with confidence. The upper limit is fixed by the clear Late Uruk connections of the pie-Ninevite 5 period 1 occupation at Mohammed Arab.128 Painted Ninevite 5 polu r\ follows on from this, after a period with a transitional style. The start ;h<- Early Ninevite 5 period may therefore be placed towards the end of tli ndel Nasr period, perhaps about 3000 b.c:.129 The lower limit is fixed by th< uiian context of the post-Ninevite , ; i .uul m.i\ be reliable 2673b.c. '■' ; - 1985a, 23 and 25 . The two determinations \'i lioih srem unlikelv. three of the four tti m> iroin phases 6 and 7 seem considerably i and s hwari/'s method of averaging the rtinatkms Ihuu phase b is not theoretically justified Schwara 1982, pp. I77ff. . the two determinations from leiul eth-I halalhal have large standard deviations and so do not give much better information, for further details see Schwartz 1985, 'fable V. "•The Painted "Late Uruk" may be related i» monochrome painted potterv from Jemdet Nasr | field an Man in 1935, PI. XXXV.2 4 6.13: Dates and Oato 1976, pM and from protolitcrate contexts iniW Divala Delouga* 1952. Pb. 28c-d. 29b. 32a-c, !34a-b, l86c.6Q3.253a . though the designs are so simple the similarity may be fortuitous. .mdet '» This is a guess. Some scholars think thai thej Nasr period lasted from 310(1 to 2900 a..:. e.g. ft'^j Hansen, and Dunham in pre.ss while other,.In to 3 100 B.C. e.g. Vcrtcsalji and Kolbus I9BJ.■ Fig. H between pp. 96 and 97). T»W^5£p '"Some of the parallels used for dating the la\«i, ,i< Mari arc from sites which are not well-dated m h layers which arc usually considered laier than K.u k Dynastic I, e.g. Chagar Bazar 3, Tell Bill a Stratum i, Barri Scavo B. A later date lor this part of the I rem h B sequence would explain the absence ol Nincviic 5 pot tery in a similar assemblage of the same dan ,.t Irl Melcbiya (Lebeau et alii 1985, 8: Lebeau c< alii 1986, 10) 123 Lebeau in press. 124 Kühne 1976, 103 and Abb. 404 and 405. '"Moortgat originally proposed an Akkadian dale for Tell Chuera (Moortgat 1965, 14 15 and 47 . IV site was re-dated to the Early Dynastie II period by Kühne (Kühne 1976, 31), but this has been disputed by supports the Akkadian date. Since Late Uruk inscribed tablets have been found at Brak and Nineveh (Kinkel 1986, 187-189; Cotton and Kcade 1983, 33-34), there may be- written documents vet to be discovered. So far no Ninevite 5 tempi, or public administrative building, except possibly al t.hagar Bazar (Mallowan 1937, 115 and Pig. 7 . where tablets might be expected has been excavated. Not only is there increased inaccuracy because of the calibration curve but also there mav be some systematic problem at least with the results from fell Brak (L). Dates 1982a, 197; j. Oaics 1985bj. Four determinations trom stratum 20 (phase 3) at Tell Lcilan which were sent to laboratories in America and japan all cluster in definite wav of telling which is more c<"" formally more radiocarbon determinations, t < J\vh()lt. the duration of Lcilan III irather than ^m Ninevite 5 period) "al Tell Lcilain 1 he t ». ^g dates from period III delimit a duration oi ^ millennium, ea. 3500 2500 B.C. an WW®?* 1 lifespan" Schwartz 1985, 58). A MYSTIKIUI'S AFFAIR <>| s|V. ». Til > deshuction levels at Tell BrikTh | m- begun ... the end of the Karly i)ynas!\, ',,!" ^"1'^ , ] .,M'rli'K! f»" ^lore. havr Early Dynastic III period, perha^25(M) Be ""' '" ^ 'M "f «* I litis we have been able to reromimet ii„.. 'Mysterious Affair of S.vl,' t^Z^\Whivh M "* -nrusion aT;,ut "r'h^""-«»nR .c«i„r, h:: ;v;; ;;:s ™n-> *nd been reconciled so that the relationships between ,he ifh "v lV''n''W styles and their relative sequence have lien firm lal "mZ" d" ' """^ however, remain^unresolved. We tnav be w.ser on thetb ee ' f t ^r.H pottery but the Ninevite 5 culture as a whole still eludes us. It remains a cuIt "out Of focus, represented only by its pottery: a clearer definition of £ pa ru " characteristics is still one of the major goals of Mesopoiamian archaeology Hercule Poirot our infallible companion in this investigation, would not have been so cas.lv satisfied with such a conclusion: for him there was alwavs a solution in which every little fact was neatly explained. . . . At last he heaved a deep sigh. "It is well. The bad moment has passed. ,\'ow all is arranged and classified/'1,1 But here the analogs b< tween 2 . as well as preliminary reports on other excavated sites. Tell Baqaq Tell Abu Dhahir, Tell Fisna, Tell Ger Matbakh. Tell Jambur, Tell Jessari, Telljigan, Kutan, Tell Rijim, Tell Sa'ud, Siyana ITya, in which Ninevite 5 pottery of various styles has been found. Other sites in the Salvage Project from which Ninevite 5 pottery has been reported are Tell Dhuwaij iTell Zummar), Khirbet Hatara, Tell Jaloqa, Tell Karana I and 2, and Tell Scllal. It has not been possible to integrate the results of these excavations into the above discussion and indeed much more detail will be available when the final reports are published, but this new evidence confirms the pottery sequence as outlined above. A recent note in Antiquity M. S. Tite, S. G. E. Bowman, J. C. Ambers, and K.J. Matthews 'Preliminary statement on an error in British Museum radiocarbon dates (BM-1700 to BM-2315)', Antiquity 61/232 July 1987:. 168 confirms that, as suggested in note 127, there is an error in the Tell Brak Carbon 14 determinations which may be approximately 200 300 years too young. Bibliography Abu al-Soof, Bchnam 1964, 'Uruk pottery from the Dokau and Shahrazur Districts . . .'. Su#* 20. •"j.Oatcs 1985b. '" Christie 1920, 122. IUGHABI ROAI-AND Rnni.R l rii.ik.k 228 . i Bchnam 1968, •Distribution of I Iruk. .Jamdal Nasi. ..,„1 Ninev.ie V ,,„Mrry . . / t] 30 74* 86. J Bcliinm I^Wi 'Excavations al Qalinj Agha (Krl.il '. Sumn 2r>. 3 42. Abu a -£oo. ' |072 -Notes on ihr Laic I'rchistorii pottery of MetopotamiaV W, or , . 1978-1980\ htanhuler Mttteilungen 81, 11 93. J«W Ph -stie A IW The Myyterious Affair at StyU* (John UflC, Dir Bodley Head. London rKristic' * 198l\. An autobiography (Fonlana, 1977 Collins . CoHon D and Reade, J. E. 1983, 'Archaic Ninevch\ liaghdadei \Uttalungvn 14, 33 41, Curtis, J (ed,) 1982, /-V/<>' ****** «/ Mesopotamia** Discovery, .British School of Archaeology in I London 1. raq, Delougaz, P. 1952, Pottery from the Diyala Region iOIP 63. Chirac» . Dvson,R. H. 1965, Problems in the Relative Chrormlogy of Iran, 6000 2000b.c;.\ in Ehrich cd ' ' 1965,215-256. Ehrich, R. W. (ed.) 1965, Chronologies in Old World Archaeology Chicago . Fales, F. M., Tusa S., Wilhelm G.. & Zaccagnini C4 in press, 'German-Italian Expedition to Iraq : Preliminary Report on the First (iampai^n nt Excavations within the Saddam Darn Reservoir Archaeological Rescue Project 1984 \ ms. subrnincd tu Sumer. Field, H. and Martin, R. A. 1935. Tainted potters in.n, Jemdci Nasr, Iraq'. American Journal of Archaeology39, 310-320. Fielden, K. 1981, \\ Late Uruk potterv t>n»up from Tell Brak, 1978", Iraq 43, 157 166. Finkbeiner, U. and Rollig, \\\ (cds. 1986, Gamdat /v. i regional ttvle/ TWO Beiheft B62), Tübingen. Finkel, I. L. 1986, inscriptions from Tell Brak 1984', frc I / 201. Frankfort, H. 1929, 'Vase F. Vorderasien5 in M. Eberl Hi n der Vorgeht hu-hte vol. 14 Berlin . 77-93. Frankfort, H. 1932, Archaeology and tin Sumtnan I* Frankfort, H. 1934, Oriental Institute discoveries in Iraq ' fiminäry Report of the Iraq Expedition (OIC 19, Chicago). Fujii, H., ct alii in press, 'Telljikan' ms. submitted r,j \ Fukai, S-, Horiiu hi. k, and Masurani, T. 1974, Tell Excavation of ltd \ ITiivrrsity of Tokyo). Ii, H. and Odani, N. 1981, il. Tell Gubba\ in H. Fuji t-iminarv report oi excavation at Gubba and Son^or\ at-Rajidan 1 .111 16 I Killick, R. (;., 1986, The Eski Mosul Region\ in Finkbein \ Rollig eds. 1986, 229 244. Kühne. H. 1976, Die Keramik vom Tell Chuera und ih re lit liehunsi ■ Funden aus Syrien-Pol äs Una. der Türkei und dem Iraq (Berlin!. Lebeau. M. 1985, "Rapport prcliminairc sur Li scqm-nce ceramique du chantier B de Mari Illeme tnillcnaire Mari, Annales de Recherches Interdtuifdin- i 4, 93 126 Lebcau, M., in press, 'Rapport prcliminairc sur la ceramique des premiers niveaux de Mari i chantier B 1984,', Mari, Annales de Recherches Interdisciplirtaires 5. Lebcau, M., Gubel, E., & Monchambrn. J.-Y. 1985, -Rapport prcliminairc sur la premiere Campagne de fouüles ä Tell Melelnva \loven-khal)<>m printemps 19H4 '. .\kkadica 45. I 31. Lebcau, M. et alii 1986. "Rapport prcliminairc sur la deusieme rampage de touilles a Teil Mclebiya (Moyen-Khabour printemps 1985 \ .IXkadica 4b. 1 4?). .. Mallowan, M. E. L. 1936, 'The excavations at Tall Chagar Bazar and an Archaeological Survey of the .. „ H^ur Region, 1934-35', Iraq 3, 1 86. " . , Mallowan, M. E, L. 1937, 'The excavations at Tall Chagar Bazar and an Archaeological Survey oi the Habur Region. Second Campaign 1936'. Iraq 4.91-177. Mallowan M. E. L. 1964. -.Ninevite 5'. Vorderasiatische Archmlw. Studien und Aufsätze Anton Moorig* •.* (Berlin), 142 154. Meijer, D.J. \\\ 1983, A Survey in Northeastern Syria (Publications of the Netherlands Historical and Mo , Ar.chaTOlo«""»l Institute in Istanbul, Vol. I A' 111, Leidem. . loortgat A. 1965, Tell Chuera in Nordost-Syrien. Bericht über die vierte GrabungskampogU '«and £ Wissenschaftliche Abhandlungen der Arbeitsgemeinschaft für Forschung des Lande. Nordrhcin-Westfalen). A MYSM-.Hioi'S All Alk < >|. S| VI. Oatcs, 0. 1982a. Excavations at I'd! Brak, 1978 81' /,„,, 44 11.7 -,ni Oatcs. I). 1982b. Tell Brak1 in Curtis (c<|.) 1982. 1,2 71 Oatcs, I). 1983, 'Excavations .it Irll Brak. I9H3 84', fau ks , Oatcs, IX, arid Oatcs J. 1976, Tki mi oj nntKatwn (Elsevier-Ph tidon Oatcs, J. 1967, Review of Abdul Jalil Jawad, The Advent of ,»„■ iv, of , i . Mesopotamia, in American Anthiupnlogtst m 401 402 ,wn™ip» in Northrrri Oatcs..]. 1982. 'Some 1Lair Karly Dynastie III pottny Iron, Tell Brak' /w/44. 203 219 OatcsJ. 1985a. Irll Brak: I ruk pottery from the 1984 season' It,,,, 47 \rt u,, Oatrs. J. 1985b, 'Tell Brak and Chronology ihr Third Mill,..,.; ' M#,,«4, 137 144. M'11'..mum. A/«„. ,1.^, rf, Oatcs, j. 1986, Tell Brak: Ihr Cruk Karly Dynastie Scnucn, r. in I'inklxinrr & R„|]ÍR ,ds. ,986> Perkins, A. L Í949, The Comparative Archeology oj Early MesoPotanua Studies in Ancient Oriental Civilization /:>, Chicago . ,,ldl Perkins, A. L. 1954, 4 The relative rlironoltiKy n{ Mesopotamia', in R. W\ Khrirh rd Relali Chronologies in Old World Archaeology Chicago). 42 55. ' ö Porada, E. 1965.'The relative chrtJimlujrx <>f Mesopotamia. Part I, Seals and Trade 6000 lfihi1**>v in Ehrich (cd.) 1965, I :» 200. }' Porada, E. 1985, 'Syrian seals from the late fourth to the tare second millennium', in H. Weiss cd. From Ebla in Damascus Washington . 90 94. ' Porada. E., Hansen, D, P., and Dunham, S, in press, ■Relative: Chronology of Mesopotamia 7000-1600 b.c. \ in R. Ehrich cd. . Chrontdogtes w Old World Archaeolo^ 2nd rew Chicago . Reade.J. E. 1%«, 'Tell 'lava Stmiman Report*, Iraq 30,234 264, Reade.J. E. 1982. 'Tell Taya\ in ( rd 1982.72-78. Roaf. M. D. 1983a, "A rrj..... on 1 ■ ■> B tish Archaeological Expedition in the Eski Mosul Dam SaKayt ProjectIron > i i 1982 to June 1983', Sumer 39,68-82« Roaf, M. D. 1983b. 'Tell Moh Kski Mousl Dam SaUaijr Project British Archaeological Expedition Catalogue ol (objects*, Sume> 3'i. 83 94. Roaf, M. I). I9ÍH. ,i\ai: n-. I '. Mi hammed 'Arabin the Eski Mosul Dam Salvage Project',Iraq 46, 141 156. Schwartz, G. M. 1982, From Pi Habur Plains: The Operation 1 Ceramic Periodization from Tell Leilan l'h I> Di-serf at i n. Dept uf Near Eastern Languages and Literatures, Yale University . Schwartz, O. M. 1985, 'Thi Nincvin \ period and current res« u« h . PaUmient 11 1.53 70. Speiser, E. A. 192"). Trrlimman Km tvati.ms at J'rpr (iawra'. Annual of the American Schools of Oriental Research 9 11929.. I 7 94. Speiser. E. A. 1931. Krttn of Februar) 28' and "letter of March 1*. Bulletin of the American Schooh oj Oriental Research 42 April 193 Í .10 13. Speiser, E. A. 1932a. Reports ... On the Tell Billah a«d repedawra excavations-. Bulletin of the American Schools oj Oriental Research 46 April 1932 . I 9. Speiser, E. A. 1932b. The "< hali.r" ware ol Northern Mesopotamia'. Bulletin of the American Schools of Oriental Research 4Q December 1932}, 5 10. Speiser. K. A. 193:1. 'Tin- potten ol fell Billa', The Museum Journal 23 1932 33 . 249 308. Speiser, E. A. 1935. Excavations at Tepe (.aura. vol. I. Levels 1-171/ Philadelphia . Siirenhařcn, D. 1978. Keramik-produktion in Habuba Kahna-Süd Berlin . Thompson, R. Campbell, and Hamili..». R. W. 1932. The British Musrum Excavations on the Temple of Ishtar at Nineveh 1930-31*. Liverpool Annals of Archaeology and Anthropology \J, Thompson! R. Campbell, and Hutchinson. R. W. 1931. The Site of the Palace of Ashurnasirpal at Nineveh, excavated in 1929 30 on behalf of the British Museum. Liverpool Annals oj Archaeology and Anthropology Iři, 79 112. , v;„..,..h Thompson, R. Campbrll. and Mallowan. M. E. I- 1933. "Ihr British Musrum Excavations at N.neveJl 1931 -32', Liverpool Annals of Archaeology and Anthropology 20, 71 18b. Tobler, A. J. 1950, Excavations at Tepe (,'awra, vol. II. Levels /.V-.V.V (Philadelphia). i a. 2nd rev. rd..