The Burial of the Dead'
Dianna Bolt and Anthony Green
l. Introduction
Apart from the ceramic assemblages by which "Ninevite 5" was originally defined, it is still, perhaps, only in the field of burial customs that the "culture" is as yet at all distinctive. Burials of the "period" have now been excavated at a number of sites, yielding a total insufficient, to be sure, for valid statistical purposes, but nevertheless suggesting the possibility of certain genera! outlines. Figs. 3-31 summarise the Ninevite 5 graves for which we have detailed information, and it is primarily upon these data that our analysis of general and particular practices will be based.
Discussion of Ninevite 5 burial practices and mortuary data rests very largely upon the wider question of what is meant or implied by the term "Ninevite 5". The name has been used to refer to a chronological period, a culture or cultural assemblage and, as Roaf and Killick (1987: 200) have stressed, to a number of entirely different styles of pottery — painted, incised and plain. It is unnecessary in this paper to become embroiled in the until recently controversial questions concerning the relationship and relative chronologies of these pottery styles (cf. below, 4). Suffice it to say that in determining the inclusion of a burial in our Ninevite 5 inventory, we have taken as the primary indicator the presence in the grave of pottery of one or more of these "Ninevite 5" styles. Admittedly this excludes by definition burials which lack ceramic grave-goods, and thereby distorts somewhat the picture of the relative wealth or impoverishment of the graves themselves. We have, however, felt justified in including graves that did not contain vessels when they have been assigned to Ninevite 5 levels in recent, small scale and comparatively well controlled excavations. Unlike Schwartz (1986), we have not, on the other hand, felt it appropriate to include burials without Ninevite 5 pottery which have been assigned to "Ninevite 5" levels by earlier excavators working on large scale operations at the multi-period sites of Chagar Bazar and Tell Billa —- especially in view of the fact that we know some of their assumptions were wrong (cf. Schwartz 1986: 47-50, on Speiser's reassignment of Tell Billa Grave 61 from Ninevite 5 level 6 to Khabur level 4; sec below, 3.3). Burials from Tell Mohammed Arab, Tell Karrana 3 and Tell Jessary which may belong either to the Late Uruk qr Ninevite 5 periods have also been omitted from the sample.
While these principles have, hopefully, ensured that the primary data arc reasonably reliable, they have, admittedly, also ensured that the already pitifully small sample-size is reduced still further. Nevertheless, it will be seen that with the removal, as unreliable, of 12 of the burials in Schwartz's sample of 31(1986: 48-49), new data on recent excavations bring the sample back up to a total of 3 1 (and excluding those for which plans and/or details of grave-goods are not presently available).
Admittedly, the exclusion of the burials at Chagar Bazar and Tell Billa that were without good datable pottery biases the sample significantly by ignoring at these sites the lowest strata of society; the sample from Tell Mohammed Arab indicates clearly [hat even adults were buried without pottery at times (see grave 62X:08, Fig. 23). This problem should be kepi in mind in any future consideration of the social stratification of the period based upon our sample or an increased sample in the light of new excavations.
Acknowledgements: For inviting our participation at the Yale Ninevite 5 Conference, and tor a number of helpful references, we thank Professor Harvey Weiss. For their very kind provision of much unpublished detailed information on burials and grave goods from their excavations, wc should like lo thank Drs. Michael Roaf and Robert Killick. Professor Carlo Zaecagnini and Dr. Elena Rova, Professor Hideo Fujii and Messrs. Hiroyuki li, 1 lirotoshi Numoto and Hiromichi Oguchi, Dr. Piotr Biclinski and Dr. Marc Lebeau. For allowing access to the object registration cards for Tell Billa housed in the University Museum, Philadelphia, we thank Professor Robert Dyson. We are grateful to Dr. Gienn Schwartz for helpful comments on an earlier draft of this paper and for drawing our attention to certain unpublished information on the Tell Billa burials. We have also benefitted from Conference discussion of the original draft ■of this paper. Dr. Piotr Bielinski kindly arranged for the final inking of the illustrations. For a copy of his unpublished manuscript, we must thank Mr. Stuart Campbell. Much of this paper was written and revised during one of the author's A.G.! tenure of the 1988-89 J. Paul Getty Postdoctoral Fellowship in the Department of the History of Art at the L r.i*.ersity of Pennsylvania, and he would like to thank both organizations, the Getty Grant Programme and the L    ersiiv of Pennsylvania, as sponsor and host, for the support which has made this and other projects possible.
519
2. Mortuary evidence
Schwartz's (1986) study of Ninevite 5 burials was concerned solely with the question of social stratification as reflected in the data of the graves (cf. now also in this volume), an aspect which has increasingly become one of the main preoccupations of mortuary analysis in general. In the present case the paucity of the data makes deductions about social status extremely difficult and problematic, as Schwartz discovered. Nevertheless, similarities and dissimilarities in burial practices, even when they cannot always be related directly to social stratification, can be of interest in themselves, and will be considered here without prejudice to the major questions of social ranking, which will be discussed separately (cf. below, 15).
Although there are severe problems in mortuary studies in the inconsistencies of recording and reporting of data, at least the presence of burials is one of the more usually reported items in archaeological preliminary reports, perhaps because of the interest of the objects that they can yield. As Schwartz (1986) found — and this contribution must echo —^- the main problem with funerary analysis for the Ninevite 5 "culture" is certainly the small size of the sample, for even in ideal circumstances it has been calculated that for divisions within burial groups to be statistically valid it requires a very large sample of graves (Orton and Hodson 1981: 113-114). With a very small sample there is a tendency to highlight certain rather random and possibly superficial similarities and necessarily ignore differences, or, conversely, to imagine, for other aspects, that there is no pattern at all, when it might simply be the absence of sufficient variables that prevent one seeing it. These are inescapable trap-falis. Nevertheless, what information we do have presents some intriguing, if presently unprovable, possibilities.
X Mortuary evidence from Ninevite 5 sites 3.1 Nineveh
Thompson and Mallowan (1933: 134) refer to a burial at the top of Level 4 "probably of the late Jemdet-Nasr period." The skeleton was in a flexed position, but no further information about it is given. Rova (19KB) has suggested that this burial might in fact have belonged to Level 5. It was, however, disturbed.
Another feature at this site which has received comment in respect of Ninevite 5 burial practices and evidence for social complexity is the presence of what have been interpreted as vaulted tombs of the period. The 1929-30 and 1930-31 excavations uncovered (in trenches BB and N) massive mud-brick vaulted structures, some thirty courses or fifteen and a half feel in height, immediately to the east of the Ishtar Temple platform, interpreted by the excavators as a scries of vaulted tombs of the Uruk period (Thompson and Hutchinson 1931: Si; Thompson and Hamilton 1932: 78-79; Thompson 1934: 96). Weiss (1985: 330) states that these structures, which he accepts as tombs, must have been built in the Ninevite 5 period, but AI gaze, who, against the protestations of later reviewers, has defended the excavators' original dating to the Protoliterate (1986: 126-127) -— mainly on the grounds that bevellcd-rim bowls are stated to have been found in large numbers in and around rhe area of the structures and, more to the point, "even well within the tombs themselves, at floor level" (Thompson and Hutchinson 1931; 81), has interpreted the construction not as individual tombs but as a single building roofed by vaulted units arranged around a central open court (cf. his reconstruction. Algaze 1986: 128, fig, I), and perhaps representing "some sort of a storehouse" {ibid.: 129).
Although no evidence of burials as such was retrieved, so that if the interpretation as tombs is correct they must have been robbed in antiquity (so Weiss 1985: 330: cf. Schwartz 1986; 50 [more dubious]), the matter is of some importance because Weiss (loc. cit.) has argued that the tombs "in themselves provide an indication of Ninevite social stratification perhaps similar in scale to that of the contemporary south." Schwartz (1986: 50), though citing parallels of vaulted tombs from the Early Dynastic south, is at first more cautious, with more tentative comments on the possible dating and admitting, on comparison of ED I fortified structures at Tell Razuk and Tell Gubba in the Hamrin, that the Nineveh structures "may not necessarily be tombs", but he too concludes (ibid.: 56) that "The most compelling piece of evidence in favor of the existence of highly stratified society is, in fact, provided by the vaulted tombs of Nineveh, a likely Ninevite V urban center."
For the present, with the limited information available, it is hardly possible to solve the question conclusively, so that as evidence either of "the importance of the Protoliterate period at Nineveh" (Algnze 1986: 129) or of a high ranking social stratum in the Ninevite 5 period in the north (with Weiss and Schwartz), these enigmatic structures should be viewed with some caution. However, the interpretation of these structures as tombs does have considerable difficulties, for even had they been robbed of grave furnishings, the skeletons would still be present, if disturbed. The modern excavators may have overlooked some bone, but it stretches belief to suppose that they missed it all, especially as they themselves regarded the structures as tombs and would presumably have been looking for skeletal remains. It seems very improbable, therefore, that these structures could have been tombs.
320
In any event, whether as tombs or as some other structures, it would seem likely they had been dug in to a considerable depth, and could well be later than the Ninevite 5 levels in origin (cf. the comments of Roaf, this volume). If, on the other hand, they represent some standing structure, such as Algaze's storehouse, they would probably have to be Uruk period in date, and again would not in any case be relevant to our present discussion of funerary evidence.
3J. Tepe Gawra
In the excavations at Tepe Gawra there were a multitude of burials in the Late Uruk levels (X-VIII), some very rich in grave goods (Tobler 1950: 51-97). No graves, however, could be attributed with any confidence to the Ninevite 5 levels VIIIA and VII (Speiser 1935: 140-143).
Parenthetically, it may also be worthy of note that a vaulted structure in Level VIIIA at Tepe Gawra (Tobler 1950: pis. 13, 24a) has been cited by Algaze (1986: 129, n. 19) as a close parallel to the vaulted structures at Nineveh: "In its construction, this isolated vaulted hall at Gawra resembles closely any one of the interconnected vaulted halls that comprise the Nineveh structure." Algaze would presumably regard this structure in the light of his interpretation of the Nineveh parallel, as perhaps a storage-hall. A single tomb might also be possible, though again, in the absence of skeletal remains, perhaps unlikely. In any case, as with the Nineveh structures, it would then post-date the level in which it was unearthed, and thus in this case certainly date to sometime after the Ninevite 5 period.
3.3 Tell Billa (Fig. 1)
The cemeteries at Tell Billa assigned to the Ninevite 5 levels 7 and 6, located outside of the settlement area (assuming there was one), remain unpublished. Schwartz (1986: 47, 49 Table lb), however, was able to reconstruct some grave groups by examination of the objects register in association with published pottery reports. His method, however, presupposes the correct attribution of level to these graves by the excavators, an assumption which, in the light of the complexities of the stratigraphy and the known case of their changing their minds about such an attribution (see below), seems unjustified. If we consider which of these burials actually contained definite Ninevite 5 vessels, we can, we believe, eliminate at least half of them from our enquiry.
Schwartz (1986: 47-50) notes that the objects register makes it clear that the large level 6 tomb referred to by Speiser (1931: 12) as "a square chamber of sun-baked bricks upon a foundation of stone, covered with a " tent-roof' of mud-brick", and which contained an apparently well-preserved human skeleton, a silver hair-pin and finger-ring, a copper or bronze bowl, two painted pottery bowls and two painted pottery jars, is Grave 61, subsequently reassigned to level 4, on the grounds of the painted pottery which is Khabur in style (Speiser 1933: 257. pi. LIXM. 3. 4; cf. also Roaf and Killick 1987; 202, n. IS: "This might well have been an intrusive grave of the early second millennium B.C.").
The three graves 70-72. Schwartz (1986: 47) writes, "can be associated with stratum 6 with certainty, since ceramic vessels from each of them were published in the Billa pottery report with a stratum 6 provenience." None of the three published vessels, however, is of itself a distinctive Ninevite 5 type (Fig. 1). Grave 72, which contained the vessel jar of Fig. la (Speiser 1933: pi. LIII: 4, p. 278) probably can be assigned to the Ninevite 5 period because according to the objects register for (he excavations the grave also contained a small plain cup with beaded rim and pointed base, a "classic" Ninevite 5 form. On the "squat pot" from Grave 70 (Fig. lb = Speiser 1933: pi. Lllf: 6, p. 278), however, Mallowan (1964: 150, n. 35) has remarked that it could be ED III or even later. (NB. The attribution of this vessel to level 6 was one of the bases for Speiser's dating of Ninevite 5 to the Jemdet Nasr period: cf. Roaf and Killick 1987, 202, n. 22). Finally, the lugged bowl from Grave 7 I (Fig. Ic = Speiser 1933, pi. LIII: 7, p. 278) is undistinctive cooking ware.
Schwartz (1986: 47) tentatively attributes three further Tell Billa burials, Graves 73-75, to level 6, on the grounds that "they appear on field photoghraphs in a group with Tomb 72, and two of them contain potter) characteristic of the stratum." For Grave 73 (he objects register again records the presence of a small plain cup with beaded rim and pointed base. It is not clear to us which other example of a vessel "characteristic of the stratum" is meant, but Grave 74 may also, possibly, be a Ninevite 5 burial because of the presence of an "ivory" cylinder seal with a row of ibexes (Schwartz 1986: 47) — perhaps rather an example of a "glazed steatite" seal (cf. Pittman 1988). There is no real evidence to assign Grave 75, however, ro the Ninevite 5.
It is unnecessary to repeat or elaborate upon the comments of Schwartz (1986: 47) about the apparent inaccuracies and contradictions in the published accounts referring to these burials. Suffice it to say that -he inaccuracies stand as further testimony of the unreliability of these accounts, and should be taken into jjr„rideration in using the Tell Billa data to supplement the primary sample, and in any general comments _"rvj: ihe mortuary evidence.
521
.14 Tell Chagar Bazar (Figs. 2-12)
At Chagar Bazar, levels 5 and 4 are attributed to the Ninevite 5 period. Level 5 of the sondage (Pit M) was a cemetery site, of which seven graves were excavated, G67-68, G70-72, G74 and one without grave goods, so unnumbered in publication (Mallowan 1936: 16-18, 54, 58-59). Individual graves in level 5 of the A.C. area are G188 and G190 (Mallowan 1937: 116, 124; plan of positions: fig. 7:4). Seven graves are attributed to level 4 of the sondage, G48, G50, G60, G65 and three unnumbered (Mallowan 1936: 16-18, 57-58) and two graves were assigned to this level on the A.C. Site, G177 and G178 (Mallowan 1937: 124; plan of positions: fig. 7: 4). This makes a total of 18 graves in all (cf. Schwartz 1986: 48-49, Table 1).
Very largely the attribution of these burials to the Ninevite 5 period appears reliable, as many of them contained Ninevite 5 pottery of one type or another. On the other hand, we know that on the A.C. Site, at least, intrusive graves of later periods were present (Mallowan 1937: 116).
The following graves with distinctive Ninevite 5 pottery are assigned by Mallowan to level 5:
Grave No.	Style qfNin. 5 pottery	Fig.
G67	incised/ plain (grey ware?)	4
G68	incised/ plain grey ware	5
G70	painted/ plain (cooking ware?)	6
G71	painted	7
G74	plain grey ware	
G188	painted/ plain (grey ware?)	il
G190	painted/ plain (grey ware?)	12
and these are assigned to level 4:
G60 painted/ incised/ plain grey ware 3
G177 painted 9
On the other hand. Grave G72, assigned to level 5, contained only a tripod bowl (Fig. 2a), while four graves assigned to level 4 also contained only undiagnosttc plain wares, G48, G50, G65 (Fig. 2b-d), and G178 (Fig. 10), although the last of these may fairly be regarded as contemporary with G177 because of their clear and related stratigraphy (Mallowan 1937: 116). The burials without grave goods, one assigned to level 5 and three to level 4, have been excluded from consideration.
The Ninevite 5 graves from Chagar Bazar can be divided into two main groups, the level 5 cemetery (G67, G68, G70, G71, G74) and individual graves of levels 5 and 4 (G188 and G190 assigned to level 5, G60. GI77 and GI78 assigned to level 4). Graves G177 and GI78, assigned lo level 4, were intramural, as possibly also was G188, assigned to level 5 (Mallowan 1937: 116).
.1.5 Tell Mohammed Arab (Figs. 13-24)
The largest sample of definitely Ninevite 5 burials, and the best recorded, now comes from the excavations at Tell Mohammed Arab in the Saddam Dam Salvage Project. Twelve burials can be attributed with certainty to Ninevite 5 levels. Again the burials fall locationally into two categories: the cemetery of between periods 2 and 3 in the sondage, assigned to the earlier pan of the Late (incised) Ninevite 5 period (cf. Roaf and Killick 1987: 207) and individual graves, without any clear architectural associations, from the excavation of the cliff section. Unpublished information on the Tell Mohammed Arab burials is by courtesy of the excavators.
hi summary, the Ninevite 5 burials from the site were as follows:
Grave
Sex
Ag.e
49V: IH Female 9-11
49V-.20
SOT: 172
.01
.02
female
Female female
4>49
23-27 K-Ul
Both'
N-S
N-S
N-S F.-W
Head Alignment.
N W
W
up up (Ni
Goods
bowl 2 jars
pin beads 2 jars
pin
jar
pin (pebble)
13
14
1.5
501:173	unkn.	3 months	E-W	E	s	beads	16
SOU: 180	male	13-15	N-S	S	E	worked bone frag.	17
50V:28						jar	18
.01	male	8-10	N-S	S	E	"gravestone"	
.06	male	33-37	N-S	s	up	(2 sealings)	
50V: 123	female	23-27	N-S	s	E	pin	19
51U:176	female	60-70	N-S	s	E	bowl 2 jars food (animal bones)	20
54V:23	female	19-21	N-S	S	w	6 piece crockery set 3 piece crockery set 2 separate (?) jars stone jar 3 cylinder seals 3 pins silver bead stone bead	21
61W:17	male	19-21	N-S	S	E	(pebble) (broken pedestal bowl in fill)	22
62X:08	female	38-42	E-W	E	s	No goods	23
64X:44	unkn.	1-2	N-S	s	E	2 miniature jars heads	24
3.6 Tell Leilan (Figs. 25-27)
Three burials attributed to Ninevite 5 (Leilan 111) were encountered in the excavation of the "Operation 1" step-trench at Tell Leilan (Schwartz 1982: 30-31, 37-38; cl". now Mayo and Weiss, this volume; Calderone and Weiss, this volume). Burial 1 (Fig. 25) was a relatively rich and elaborate grave. Burial 2 (Fig. 27) was an incompletely excavated adult burial, with a single vessel. Burial 3 was an infant pot-burial: such burials from other sites (Mohammed Arab. Tell Jessary) have been excluded from the sample here because of doubt as to whether they should be assigned to the Ninevite 5 or Late Uruk periods, but in this case, although it was low in the Ninevite 5 deposits (tentatively assigned to level 39, and 0.4 m below its surface) and "Its stratigraphic position was unclear because of the intrusion of a large modern pit in the area" (Schwartz 1982: 30), it can hardly be earlier than the Ninevite 5 period, even if its precise level and subperiod attribution may be in doubt.
The pit filled with pottery from level 19 (Fig. 26). called by Schwartz either "Pit I" (1982: 301-307. figs. 36-39, 1986: 50) or "Pit 2" (1982: 37. 65-66. figs. 22-23). and believed to be associated with Burial 1, may be a separate grave, in which the body was never deposited, had decayed, or had been robbed (for expensive personal adornments?), but is perhaps more likely part of the access shaft to Burial I (see 6.4).
3.7 Tell Rijim (Figs. 28-29)
An area of a Ninevite 5 period cemetery. largely destroyed by second millennium B.C. foundation trenches and pits, but with two intact burials, has been uncovered in Polish excavations at Tell Rijim in the Saddam Dam area (see especially Bielinski. this volume: earlier Bielinski 1987; Ball and Black 1987: 244). Of the two intact graves, one (in trench F) was only partially excavated, and produced the skull of a child with, between the top of the skull and the edge of the grave-cut. an incised and excised jar with a small plain cup within its iim (Fig. 28). The other (in sondage S-2 within trench A-l) is a very rich grave of an adult, provided with reed matting and 31 pottery vessels of Ninevite 5 incised, painted and plain wares, including "sets" of bowls of different sizes placed one inside another, and for the most part placed within two very large pedestal bowls, one plain, the other painted (Fig. 29).
is Tell Thuwaij (Fig. 30)
An adult burial with two vessels was found by the Japanese Expedition working at this site in the Saddam Dam Salvage Project (Fujii el al. 1989-90; cl. Numoto. this volume).
3.yTell Karrana 3 (Fig. 31)
The graves from the Late Uruk and Ninevite 5 levels at Karrana 3 are presently unpublished. They are largely the intramural burials of infants and babies.
Since the attribution of most of these burials either to the Ninevite 5 or to the Late Uruk levels is in doubt (and most of them are probably Late Uruk), they have mainly not been included in the sample here. One grave, however, Burial 10, the burial of a six to eight year old child, is certainly of Ninevite 5 date, as it contained a very early ("Transitional") painted Ninevite 5 vessel. Information on this burial is by courtesy of the excavators.
3.10 Tell Jjgan
Two graves of Ninevite 5 date have been identified in the eastAvest bulldozer trench cut by the State Organization of Antiquities and Heritage across the northern part of the mound, in the north-east. According to Killick and Black (1985: 231), they "contained some cylinder seals and a gold ear-ring."
The three rock-sealed graves (cf. 6.5) and six pit-burials excavated at this site by the Japanese Expedition and tentatively assigned by them to a late Ninevite 5 to "Akkadian" horizon (Ii, this volume), are not considered here, as the illustrated vessels appear to be post-Ninevite 5 in date (cf. also Ii and Kawamata 1984-85; 189-196, with figs. 23-26).
3.11 Tell 'Atij
An unspecified number of Ninevite 5 burials have been excavated in the uppermost level of this site, yielding a number of pottery vessels as well as ornaments worn on the person, jewellery and metal bracelets "too decomposed to be described" and some items of fine workmanship, including a pendant shell engraved with the form of a stylised bull (Fortin and Schwartz, this volume). When more information on these burials is available, it will add to our picture of the range of funerary goods deposited in the Ninevite 5 period.
3.12 Tell Raqa'i
An unspecified number of burials excavated in levels 2 and 3 have been attributed to the very end of the Ninevite 5 period or to a post-Ninevite 5 "transitional" phase (Fortin and Schwartz, this volume; Schwartz 1989). No details are yet available, but they may provide some insight into the deveiopmenf of Ninevite 5 burial practices on a village site in the latest phase.
3. ij Tell Kutan
Again these arc largely intramural graves of children and babies, in one case of an adult (Forest 1987). Insufficient detail is available to allow the inclusion of these burials in the present survey.
3.14 Tell Melebiya
At this site a number of apparently associated infant pot-burials was uncovered, apparently belonging to the Ninevite 5 period. They arc presently unpublished (Personal communication of Fvl. Lebeau).
.us Tell Jessary
An infant pot-burial in level 3 may belong to the Ninevite 5, but could equally be of Late Uruk date (Fujii etal. 1989-90).
4. Chronology
We have not in this paper made any real attempt to consider changes in burial practices within the Ninevite 5 period through time and space. Once graves have been classified for one reason or another as "Ninevite 5", the group has been considered as a whole. This is mainly because of the already small size of (he sample, so that subdivision would create groups too small for viable analysis. Additionally, we have not wished to get involved in the question of Ninevite 5 ceramic phasing. Bven though, as evidenced by many of the contributions to the present volume, there is now some broad agreement on the relative chronology of the Ninevite 5 pottery styles; (which only a year or so ago was a matter of high controversy: see Roaf and Killick 1987), since the now generally accepted periodisation of these styles (painted — mixed — incised) is, in good part, based upon the material from the burials (especially at Tell Mohammed Arab), we are in danger of creating some circular argument.
Nevertheless, we can say that the earliest burial included in the sample is probably the child's Burial 10 from Tell Karrana 3, with its vessel of "Late Uruk/ Ninevite 5 Transitional" type (for which see Rova, this volume). Its occurrence marks a significant break from the Uruk period, with its apparent total absence of burials other than infant pot-burials (and so suggesting some other means of disposal of the dead). The remaining graves in our sample which contained pottery could be roughly set in sequence on the basis of the pottery styles, while other associated burials (e.g. graves from the same cemeteries) might be assumed to be roughly contemporary.
524
There is one odd case- the five painted vessels in the pit ("Pit 1:') associated with Burial 1 at Tell Leilan (Kg- 26) Strati«raphically the Burial is now recognised as lying well within strata of the Incised Ninevite 5 period Calderone and Weiss (paper given at the Yale Conference) have suggested as an explanation that these vessels whose paint is extremely fugitive and unlike that of painted Ninevite 5 proper, were "archaizing funerary'vessels which were used within high, status burial rites during the late Ninevite 5/ Leilan HI period." If this interpretation is correct, we should in the future, as our sample of buiiais grows, expect to find other cases of this practice. In the meantime, however, whether as part of such an archaising funerary practice, or for some reason peculiar to this particular burial, it is difficult to escape the conclusion that these painted vessel were in some way "out of style" for their time.
5. The siting of burials
We now propose to consider whether there are features of internal consistency, or, foT that matter, inconsistency, in the burial practices within the Ninevite 5 group of burials. Our reference will be to the details of the graves in our sample (Figs. 3-29), with secondary reference to comparative evidence as seems appropriate. Although the different aspects of a burial — the age and sex of the deceased, the nature of the construction, the grave goods, etc. — are not independent variables, it is convenient and gives a certain clarity to consider certain individual traits, similarities and dissimilarities, in isolation.
Apart from the levels 7 and 6 "cemeteries" at Tell Billa, the attribution of whose graves to the Ninevite 5 period we have in a number of cases reason to doubt (above, 3.3), three distinct cemeteries of the period, sited outside of the settlements, are known, namely the level 5 cemetery at Tell Chagar Bazar, the post-period 2 cemetery at Tell Mohammed Arab and the largely disturbed cemetery at Tell Rijim. In the Mohammed Arab cemetery the graves of women predominate, but that is probably an accident of excavation; a male is represented in grave 50V:28. Since the youngest child included was four years of age, however, we can say that infants are not represented. Interestingly, in the case of both of the dual burials of an adult and a child (a young woman and a girl in 50T: 172, a man and a boy in 50T:28) there is some pathological evidence that the deceased were related (below, 14). The other most striking element in the cemetery is the almost invariable orientation, the body lying north-south, with the head to the south, facing east (and so lying on the right side); this, however, is the preferred orientation for burials at the site generally (3.5; 8). For the Chagar Bazar cemetery burials we have no information on the sex of the occupants, but we do know that, though mainly the burials of adults, they did include at least one of a child (G71) (There are graves from other periods at this site where the report mentions "infants", so we conclude that this really was the grave of a child, not an infant). Again the almost invariable orientation, in this case east-west with head to the east or west, was particular at the site in general (Mallowan 1936: 17) (below, 8). Of the two remaining intact burials in the cemetery of Tell Rijim, one was of an adult, the other a child. The orientations of both skeletons were the same, north-south with the head to the north (below, 8); in this case, however, we have no burials from the site in non-cemetery contexts to compare.
Other burials of adults and children are isolated individual graves not in cemeteries. Usually there is no clear or recorded relationship to architectural features. As they can be widely distributed spatially, they were probably often sited at any vacant locations in and around the settlements. Graves G177 and 178 assigned to level 4 at Chagar Bazar, however, were intramural, that is to say set beneath the floor of an inhabited private house; possibly grave G188, assigned to level 5. was also (Mallowan 1937; 116). The orientation here, though usual for the site, was determined by the orientation of the rooms {ibid.: 118).
The evidence at present can therefore be summarised as:
m Adults, men and women, and children, but not infants, were both buried in cemeteries, without segregation, sometimes with an adult and child in the same grave (although it is impossible to know whether the evidence at Tell Mohammed Arab of possible family relationships was typical or a special case).
(2j In the known cases (at Tell Mohammed Arab) where an adult and a child share a grave, they are of the same sex and there is osteological evidence for a possible family relationship between the two.
(3) Intramural burial beneath the floor of a private house is attested, but without information on whether these were restricted to infant, child or adult burials.
(4) Infants were apparently never buried in cemeteries, and were usually (but not always) placed in pots (see 6.6).
6. Grave construction
The majority of burials are simple inhumations ((here is no evidence of the practice of cremation or the use of fire in the graves). Our detailed information on the construction of graves of the period comes
mainly from Tell Leilan and Tel] Mohammed Arab, as no plans or detailed records were published for Tell Chagar Bazar.
«.r Simple pit burial
It is convenient, in discussing the construction of the graves, to distinguish between the grave-chamber where the burial was placed and the grave-shaft which gave access to it. In the most simple type of burial we find a relatively shallow shaft, directly upon the floor of which the body and grave goods were placed, prior to back-filling. An example of this type is Burial 2 at Tell Leilan (Fig. 27); it was incompletely excavated, but as far as is known contained only a single vessel (incised jar) and the deceased displayed no personal adornments such as beads.
This is comparable, perhaps, to Mohammed Arab 62X:08 (Fig. 23), a simple pit burial without any grave goods. An infant burial, 64X:44 (Fig. 24), was also a simple pit, but contained some personal jewellery (stone and shell beads) and two miniature vessels. A further example is Burial 10 at Karrana 3, which also contained a single vessel, although the deceased wore a pin and cylinder seal, both of copper or bronze, and a worked bone tool was also present (Fig. 31). In its square vertical shaft, Mohammed Arab 61W:17 (Fig. 22) resembles more the grave type discussed in 6.2, but the relatively shallow depth of the shaft and the fact that the body was placed directly upon the floor, without elaboration and, save for a possibly accidental pebble and a broken pedestal bowl high in the fill of the shaft (cf. 10.1), without grave goods, relate the burial more closely to the simple pit graves of this broad type. The adult burial at Tell Rijim (Fig. 29) similarly can be compared in its rectangular shaft to the burials of the type outlined under 6.2, but the relatively shallow depth and the absence of any real distinction between "access shaft" and "burial chamber" relate the grave more to the present group. In this case, however, the deceased was well provided for in the matter of grave furnishings (see Bielinski, this volume).
A comparison of the basic dimensions (in metres) of these simple burials shows them to be remarkably consistent:-
Grave	L.	W.	Depth
Leilan 2	1.4	0.47+	0.96
MA 62X:08	1.2	0.7	►T
MA 64X:44	e. 0.7+	c. 0.5	)
MA6IW;I7	1.2	II	0.7+
Tell Rijim, grave inA-1	c. 1.9	e. 1.55	c. 0.6
f>.2 Deeper pit. with subterranean belling-out on one side for burial
Mohammed Arab 49V:20 (Fig. 14) is similar to the graves mentioned in 6.1, but is deeper (1.8 m), with a distinct bciling-out of the lower shaft (to max. diam. of 1.7 m) on the west, thus creating a sort of separate "chamber"; it was on this side that the body and grave goods were placed. This widening of the pit at its lowest level was presumably the inevitable consequence of the deeper shaft, since the body would have to be lowered from a greater height and it would be convenient for it to be placed in position by someone standing in the shaft. The added space for manoeuvre also allowed for a further elaboration in the construction, a mud-brick vault built over the body itself (cf. 6.7).
63 Vertical rectangular shaft with burial chamber dug-out on one side near its floor
This is a further elaboration of the simple pit burial, but with a now more defined and distinguished "burial chamber" dug-out on one side to contain the body and grave goods; again the increased case of working in a less confined space is the probable explanation.
There are two clear examples of this type of construction, Mohammed Arab 50U:I80 (Fig, 17) and 50V: 123 (Fig. 19), both from the post-period 2 cemetery. Both have the burial chamber dug-out on the west side. The dimensions are again comparable:
Grave			Shaft			Chamhei
	L.	W.	Depth	L.	W.	Depth
50U:IXO		1.5	1.9	1.1	0.55	'f
50 V: I2.i	1.3	1.3	1.6	1.3	0.7	0.3
The chamber of 50U: 123 differed from that of 50V: I tiO in that it was dug not only into the west wall of I he shaft but below the level of the shaft itself (cf. 54V23 under 6.9).
The fill of the burial chamber of 50U:180 contained a section of collapsed wall, interpreted by Killick (in prep.) as possibly part of an original partition wall blocking off the chamber from the shaft (cf. 6.4). It is also possible that it was the remnants of some kind of mud-brick structure built over the body (cf. 6.7).
6.4 Vertical rectangular shaft with burial chamber or chambers dug-out on one or more sides near the floor and bricked-off by partition wall(s)
This further elaboration served definitely to distinguish as something separate the burial chamber, or chambers from the access shaft. The burials that fall within this general type were in each case rather unusual in their other features and are therefore difficult of comparison, but there is no doubt that the partition wall blocking-off the burial from the shaft was a definite distinguishing feature of a group of burials which also show some other similarities.
The clearest examples of this feature are in the Tell Mohammed Arab cemetery, graves SOT: 172 (Fig. 15) and 173 (Fig. 16). The former is a burial of an adult woman and a nine year old child, possibly, on osteoloaical grounds, mother and child (cf. 14). The adult was placed in a side chamber dug-out on the western side (as with the burials discussed in 6.3); the child was in a second chamber to the south. Both burial chambers were bricked-off from the main shaft. Similarly, in 50T:173 (Fig. 16), a baby of some three months was placed in a burial chamber to the norlh of the shaft, again bricked-off with a partitioning wall.
A variation on this is represented by 50V:28 (Fig. 18), a very unusual "two-tier" kind of pot-burial, but also, in a sense, with the shaft bricked-off from a chamber dug-out on the western side. In the lower burial a row of five very large bricks in effect filled the entire shaft east of the burial, and acted as a "platform" for the funerary goods; a second course of three smaller bricks effectively blocked-off the upper burial from the rest of the shaft, and from the grave goods. The grave contained the bodies of a man and a nine-year old male child, once again, on osteological evidence, possibly related (cf. 14).
These three burials have a number of features in common. Firstly, they all involve the burial of young children, in one case a baby, in the other two cases quite possibly parent and child. Grave 50U: 180, which, as we have mentioned, may also have originally contained a partition wall, may be broadly comparable in this respect, in that it contained the skeleton of an adolescent, of some fifteen years. However, although this boy was still a minor by our standards, some 5000 years ago he might have been considered a fully fledged adult member of society, and could even have been married (?).
Apart from personal elements of dress and ornamentation (a copper or bronze pin on the woman in 50T:172, a bone pin on the youth of 50U: 180, a stone bead worn by the baby of 50T:173), none of these burials contained any grave goods in the burial chamber itself. 50T: i 72 contained an incised jar sealed by a pebble, but this was in one corner of the shaft, placed on a layer of silt. In 50V:28 also, in a sense, the grave goods were not contained within the "chamber" but within the eastern, separated, part of the shaft, upon their specially constructed brick shelf.
Grave 54V:23 at Tell Mohammed Arab (Fig. 21). the wealthiest Nineviie 5 grave from the site, can also broadly be included in this category, although it defies precise classification along with any of the other burials. Into the floor of the roughly rectangular access shaft a small dug-out "chamber" on the norlh side contained the body, originally covered with a brick structure which had since collapsed into the main shaft, on top of some of the grave goods, for these in the main lay outside of the dug-out and of the brick structure (including a couple of copper or bronze pins which one might have expected to be a part of the personal dress of the deceased; but cf. below, 11),
Another grave showing certain unique aspects of construction is Burial 1 from Tell Leilan and the pit containing a mass of pottery which the excavators believed was associated with the burial (Figs. 25-26) (cf. above, 3.6). Schwartz (1986: 50) has remarked that this and grave 54V:23 at Mohammed Arab bear "a compelling resemblance ... in the brick structure, the relatively rich assortment of grave goods, and the large deposit of pottery outside of the grave itself." The items in 54V:23 were not exactly outside of the grave, however, but upon the floor of the access shaft, outside of the burial area. One wonders, therefore, whether in the Tell Leilan burial, the mass of pottery accumulated in the "pit" to the west was not, in fact, also lying within the shaft of the grave, with the burial chamber dug out on the eastern side. As the brick structure in this burial covered the body, it is likely to have been similar, perhaps, to the structures in graves 54V:23 and (?) 49V;20 at Mohammed Arab (6.7). but it could also be interpreted as a (collapsed?) partition wall, similar to those in 50T: 172-173, or as a combination of covering structure and wall. Be this as it may, a re interpretation of the area to the west of the ••chamber" of Tell Leilan Burial 1, including the dugout "pottery pit", as the access shaft of the grave (Fig. 25) seems a possibility which would bring this enigmatic burial at least partly in line with what is now known from Tell Mohammed Arab.
6o Rock-sealed tombs
Two graves from Tell Mohammed Arab were distinctive pit-burials sealed by very large capping stones, hewn from the bedrock, and wedged in position by smaller stones. A similar construction can be seen in the slightly post-Ninevite 5 (?) graves from Tell Jigan (li and Kawamata 1984-85: figs. 23, 26; cf. above, 3.10). The method vaguely recalls, perhaps, the sealing of the hewn-stone tombs within the later "Barbar period" burial mounds of Bahrain (cf. e.g. Ibrahim 1982; see Lowe 1986 for the dating). The explanation, here and in Bahrain, may partly be that the burials happened to be sited at places where the relatively shallow accumulation of soil necessitated excavation into bedrock, which made it appealing to use the larger rocks to seal the grave,
Admittedly, 49V: 1 8 (Fig. 13) was not itself dug into bedrock, however, and if the rope impressions preserved in the mud between the stones are an indication of the expenditure of labour in hauling the larger rocks into position, so is the distance they had to be brought from their site of mining. Other features in the construction of the grave also show unusual, rather elaborate, lengths, for example the rectangular burial chamber dug into the floor in the centre of the (untypically) circular shaft. If we may trust the interpretation of the supervisor, the shape and size of the chamber may have been necessistated by the placing of a wooden box (perhaps a kind of coffin?), within it, covering (or containing?) the body and goods.
In the case of 51U:176 (Fig. 20), the shaft and chamber sunk into its floor did indeed cut into the bedrock. There is no doubt, however, that as with 49V: 18 it would have taken a considerable expenditure of effort to haul the larger capstones into position over the chamber.
It should be no surprise, therefore, to discover that these two graves, with their more labour-intensive method of construction, contained rather more in the way of grave goods and personal dress than the grave types previously discussed. Each had three pottery vessels, and the child of 49V:18 wore a necklace of semi-precious stones, while a copper or bronze pin at the left shoulder presumably fixed her clothing or burial shroud. (Cf. the numbers of vessels and array of personal adornments of the later Tell Jigan burials: M and Kawamata 1984-85: figs. 23-26). Tell Mohammed Arab 51U: 176 also/contained a collection of animal bones, probably an offering of food.
(,.(, Pot-burials
The burial of infants inside big cooking pots is certainly attested for this period at Tell Leilan (Burial 3), Tell Kutan (Forest 1987) and Tel! Melebiya (M. Lebcau, personal communication). At the first of these sites, a large urn, containing the body, was covered with a lid and buried at 0.40 m below the surface (in level 39). At Tell Kutan (and Karrana 3) there were burials of children in vessels sawn horizontally in two.
Infant pot-burials are also attested at Tell Mohammed Arab, Tell Karrana 3, and in the otherwise undistinguished level 3 at Tell Jessary (Fujii et al. 1989-90, with jar, Fig. 15: 6), but their attribution to either the Ninevite 5 or the Late Uruk period is unclear.
A variation of the pot-burial, with horizontally half-sectioned urn containing the body of an adult man and (possibly also originally within the walls of the vessel, although this is uncertain) with male child placed above, is seen in Mohammed Arab grave 50V:28 (Fig. 18), discussed above (6.4).
6.7 Brick constructions in graves
Certain Ninevite 5 burials show mud-brick structures as part of their construction. We have mentioned the partition walls blocking-off the chambers from the shafts in Mohammed Arab graves SOT: 172-173 and possibly 50U:18O, and the partition-wall/platform in 50V28 (6.4). In 49V:20 at Mohammed Arab there was recorded, above the thorax, a row of mud-bricks set on edge, possibly a brick vault of some kind constructed over the body. Brick structures lying at least partly over the body were also encountered in grave 54V:23 at Mohammed Arab (cf 6.3), Burial 1 at Tell Leilan (cf. 6.3) and the grave at Tell Thuwaij (Numoto, this volume).
It is interesting to compare these features with the record of a brick structure in an ED III burial (Grave 26) at Abu Salabikh: "Although it was not clear at the time, it now seems very probable that a single row of plano-convex bricks which lined the NW side of the main grave pit, and separated it from the access shaft, was in fact a deliberate construction related to the grave. ... Very likely related in some way was another row of bricks al right-angles to this but at a lower level, running from the E comer of the access shaft to the SE edge of the grave pit opposite it." (Postgale 1985: 64-65).
7- The lining of graves and coverings for the body
Probably most of the graves were originally lined, and traces of the lining have been observed and recorded in many cases. It seems that the floor (and sometimes the lower walls?) of the burial chamber could be lined with reeds or palm-leaves, perhaps loose but more likely in the form of woven matting (see especially Mohammed Arab 51U:176, Chagar Bazar G65 [Mallowan 1936: 58], Tell Rijim [Bielinski, this volume] and Karrana 3 Burial 10), prior to the placement of the body and grave furnishings, which might be covered with more matting or loose reeds or both (reed impressions have often been observed both below and above the skeletal remains). It may be that on occasion the deceased was actually wrapped in a reed mat.
Evidence of clothing is naturally not normally present, except indirectly by the metal pins positioned on the shoulder or chest, which presumably fastened the garment or burial shroud (see 9). In grave 54V:23 at Mohammed Arab (Fig. 21), however, fragments of textile were recovered, interpreted by the excavators as suggesting that the body had been wrapped in a blanket (Roaf 1983a: 74). Several pieces of cloth found in the area of the chest of the adult skeleton from Tell Rijim have been identified as fragments of silk and linen (Bielinski, this volume).
8. Orientation and disposition of the body
It is well established that the orientation by points of the compass of the body in burials at certain sites at certain periods, at least, can be a constant factor, or one consistently related to other factors, such as the age or sex of the individual deceased. Campbell (in prep.), for instance, has analysed the evidence of orientation in graves of the Samarra, Hassuna, Halaf and Ubaid periods, and demonstrated detlnite, though temporally and spatially varying, preferences for, or avoidances of, particular directions. He notes, however, that "in most cases the preference for a single direction does not mean that no other orientations were found and frequently the sites with smaller samples ... produce no clear orientation."
Inevitably, any attempt to consider the orientation of Ninevite 5 burials will to some extent fall into the traps of the insufficient data sample, namely inability to appreciate patterns with even a few variables or dependent factors, or over-generalisations from superficial or highly localised similarities. Nevertheless, whatever the dangers of formulating more detailed or general descriptions, the conclusion that orientation was, or could be, a relevant factor in Ninevite 5 burial practices seems inescapable even from the limited number of burials excavated. In this sense, our data, though small, are revealingly productive, and may be compared, for example, to the some 74 burials published in the first of the graves volumes for Abu Salabikh (Postgate 1985), only 33 of which (Graves I, 10, 12-14, 16,20, 28,31-34,37.48,52.54,57,61-62, 68, 72, 76, 78, 80-82, 85-87, 89-90, 97-98) contained sufficient in situ skeletal remains for an orientation to be established, of which, for example, only five or six (Graves 13. 16. 57. 761, 68, 82) were well enough preserved for sexing, and do not demonstrate any obvious pattern. II' this seems hardly surprising given the final number, contrast the 12 certainly Ninevite 5 graves from Tell Mohammed Arab alone, containing 14 skeletons, of which the sexes of twelve could be determined (the other two bodies were of babies), a final sample not vastly greater in size but one that certainly does demonstrate constant, if possibly localised, factors.
With the sole exception of G67, in which the body was oriented northeast-southwest, with the head to the northeast (Mallowan 1936: 58), the Ninevite 5 burials from Tell Chagar Bazar had the body oriented in an east-west direction, with head towards either the east or west (Mallowan 1936: 17, for the sondage; for the A.C. Site, cf. Mallowan 1937: 118, with the plan, ibid.: fig. 7: 4. showing head always to the west. NB. Since Mallowan's comments relate to burials at the site in general, it is possible that for the Ninevite 5 burials the alignment might actually have been rather more restricted, for example, with the head always to the west). The burials at Tell Karrana 3, including Burial 10, demonstrated the same alignment, but with the head invariably towards the east, the orientation also of the Ninevite 5 burial at Teli fhuwaij (Fis. 29) (Fujiietal. 1989-90: fig. 4).
The Tell Mohammed Arab burials were oriented as shown on Figs. 13-24 and listed in the table in sec-rion 3.5 above. They show, both in the cemetery site and elsewhere, a strong preference for north-south orientation, with the head normally to the south, usually facing east. Most of these burials are of females, but since 50U:180, 50V:28 and 61W: 17 are the graves of males, this orientation, at variance to Chagar Bazar. Karrana 3 and Tell Thuwaij, does not appear to be dependent upon the sex of the grave occupants. A similar north-south alignment, but with head to the north, facing west, is followed in both of the intact Ninevite 5 burials at Tell Rijim (Bielinski 1987, and this volume).
At Tell Leilan, the body of Burial 1 (Fig. 25) is aligned north-south (head to north facim- west), that of Burial 2 (Fig- 27) east-west (head west facing north).
529
A further point on orientation may be made with respect to the position of the burial chamber, or "dugout" containing the burial, in relation to the shaft. At Tell Mohammed Arab, grave 49V:20 "belled-out" on its western side to contain the burial, and it is on the western side of the shaft that the burial chamber was excavated in SOU: 180, 50V:28, 50V: 123 and 54V:23; grave SOT: 172 has two burial chambers, but that of the adult is also on the western side. Child burials at this site, however, could be placed to the north or south of the grave (50T:J72, 50T:173), and the chamber is to the north of the shaft in Mohammed Arab 54V:23 and the burial from Tell Thuwaij. If our suggested re interpretation of Tell Leilan Burial 1 and Pit 1 should prove correct (above, 6.4), the burial chamber there was to the east of the shaft. Similarly, in the adult burial from Tell Rijim, the body lay at the eastern end of the grave.
Body and grave orientation is usually considered in relation to the cardinal points of the compass, which here is not only an objective criterion but seems also be have been a significant one. It should be remembered, however, that there may be other forms of orientation which are not so easily determined, for example orientation related to the construction of the grave. Any such reasons for varying orientations in Ninevite 5 burials presently elude us, but it does seem that the orientations of the intramural burials of the A.C. site at Chagar Bazar were dictated by the alignments of the walls of the houses into which they were dug (so Mallowan 1937: 118, and fig. 7: 4).
In all Ninevite 5 burials excavated to date the legs of the skeleton were flexed, sometimes lightly but more usually with the knees drawn tightly up to the chest. The arms are bent at the elbows, with the most common, but not invariable, position of the hands placed immediately before the face (Mohammed Arab 49V:I8, 50TT73, 50V: 123, 5111:176, 61W: 17, 62X:08, ?64X:44; Leilan Burial 1; Karrana 3 Burial 10; cf. also the Tell Rijim adult burial, with one of the hands before the face). On occasion this positioning seems to be elaborated, if not "explained", by the presence of a vessel (or two) almost in the grasp of the hands (Mohammed Arab 5 1U:176, ?64X:44; Leilan Burial I). This is a practice paralleled in Early Dynastic burials from southern Mesopotamia. At Kish "the hands were generally placed in front of the face with a copper bowl or small pottery jar or dish between them as if for the dead man to drink" (Mackay 1925: 13), while at Ur "In every undisturbed grave the dead man holds between his hands, in front of his face, a cup or drinking-vessel of metal, stone, or clay" (Woolley 1934: 144). Also at Fara: "Leichen auf der Seite liegend, mit schwach angezogenen Bcinen, eine Hand unter den Kopf gelegt, die andere ein Trinkgefass am Munde haltend" (Heinrich 1931: 17). That in Grave 1 at Abu Salabikh the finger-bones of the right hand were actually found resting inside of the stone bowl placed before the face (Postgale 1985: 21), and that in Grave 82 the small jar placed close to the face contained (unidentified) animal bones {ibid.: 153, No. 8), suggests that the action represented niight be one of eating rather than drinking, and Postgale (1980: 69) notes that in a number of cases when the hands are held before the face "there is no such vessel at all." This is the case also with the Ninevite 5 burials, and we might tentatively suggest, perhaps, that some perishable food had originally been placed in the hands.
<>. Dress and personal items
Although traces of actual cloth arc singularly rare (see 7). the presence of copper or bronze pins, and perhaps also of some jewellery, such as the gold earring from Tell Jigan (3.10) and the pendant from Chagar Bazar grave G67, suggest that the body was usually dressed in some way. Copper or bronze pins were found lying on or near the body in five Mohammed Arab graves; they lay either on the chest (49V:2(), 54V:23), near to the left shoulder (49V: 18, 50T: 172). or before the face, possibly originally holding a garment at the right shoulder (or is this a hairpin?) (50V: 123). Similar pins were also found in grave 54V:23. but among the accumulation of grave goods on the floor of the shaft, and possibly out of position (although cf. I i below). In Burial I at Tell Leilan, a pair of such pins was placed before the face (hairpins, or originally holding a garment at each shoulder?), while in Burial 10 at Karrana 3 a copper or bronze pin was also found in the vicinity of the skull. Copper or bron/.e pins (and needles?) of various types were common in graves at Chagar Bazar (3 in G67 and I in each of G70. G7 I and GI88). At Tell Mohammed Arab, where we have osteological information on the skeletal remains, copper/bronze pins were present only in the graves of women (cf. 3.5, 12).
When present, beads arc found in numbers around the neck (Mohammed Arab 49V: 18. Tell Leilan Burial 1, possibly also Chagar Bazar G177), or individually at the wrists (Mohammed Arab SOT: 173 (?). 54V:23); the position in the infant grave 64X:44 at Mohammed Arab suggests the possibility that they may here have adorned the ankles. No long strings of beads have yet been found around the waist, as in two Early Dynastic graves from Abu Salabikh (Postgate 1980: 69). The recorded materials of beads from Ninevite 5 burials are (unspecified) rough stone (Mohammed Arab 50T:I73). black stone (MA 54V:43). green stone (MA 49V:IS), carnelian {MA 49V;18. Chagar Bazar G177, Leilan I), quartz (Chagar Bazar G190). rock crystal (MA 49V: 18). faience (Chagar Bazar G 177. ?MA S4V:23). silver (Chagar Bazar G68) and simple shells (Chagar Bazar (1177. MA 64X:44. Leilan I). A finely caned shell bull-pendant comes
apparently from a Ninevite 5 grave at Tell ' Atij (3.11). A baked clay pendant in the form of a cylinder with flattened, perforated, end comes from a smashed-up Ninevite 5 burial at Tell Rijim (Bielinski, this volume).
The gold earring from Tell Jigan (3.10) is the only example known from a Ninevite 5 burial.
Grave 54V:23 at Tell Mohammed Arab contained three cylinder seals of calcined stone. For a number of reasons, this is an extremely interesting find, firstly because these seals were worn at the wrists like beads; although clearly personal items, cylinder seals have not normally be found in contexts suggesting that they were worn on the person (cf. the comments of Postgate 1980: 69, on ED burials at Abu Salabikh). Secondly, the very fact that three such seals might belong to a single individual, and a woman at that, is itself a challenge, perhaps, to our preconceived notions, although it should not, in fact, be a surprise: compare, for example, the Early Dynastic graves at Kish, in which one or more cylinder seals were found "in both male and female burials" (Mackay 1929: 90) and the female graves with cylinder seals at Ur (e.g. Moorey 1977: 24-40). It is always possible, of course, that this single Ninevite 5 instance is untypical, and that family heirlooms or cast-off seals were here being retained and worn, not so much as the individual's personal seals, but as a form of jewellery. The glazed steatite cylinder seal from Burial 1 at Tell Leilan was apparently also worn on the wrist (Fig. 25); in this case it does seem to be an "heirloom", in "piedmont Jemdet Nasr" style (Parayre, this volume; and cf. Calderone and Weiss, this volume). A copper or bronze cylinder seal from Burial 10 at Tell Karrana 3 was found in the vicinity of the skull, perhaps also worn at the wrist (given the position of the right hand raised to the mouth: Fig. 31), although it is also possible that this seal was originally strung around the neck. Three Ninevite 5 seals, two of limestone and one of baked clay, derive from smashed-up graves at Tell Rijim, but their original location on the skeletons is, of course, lost to us (Bielinski, this volume).
The statement of Speiser (1931: 13) that in level 7 at Tell Billa "generally a cylinder seal was placed with the dead", a remark which has intrigued some people, must be regarded as inaccurate (cf. Schwartz 1986: 47), although Grave 74 could be a Ninevite 5 burial with a "glazed steatite" cylinder seal (above, 3.3).
A small stone object from Chagar Bazar grave G67 has been described as a "stamp seal" (Mallowan 1936: 58), but is perhaps more simply regarded as a bead or pendant (Fig. 4: 2). Lumps of clay with impressions of stamp seals, one from Chagar Bazar G71 (Mallowan 1936: pi. 1: 9) and two from Mohammed Arab 50V:28, are perhaps accidental inclusions in the grave fill (?).
to. The grave goods
Burials without grave goods are certainly not unknown; not only the infant pot-burials at Tell Leilan (Burial 3) and Tell Kutan, but also Mohammed Arab grave 62X:()8 contained nothing in addition to the body, while Mohammed Arab 50T:173, 501;:! SO and 50V: 123 contained only personal items of dress. Moreover, certain burials without grave goods from Chagar Bazar, excluded From our sample here, probably were of Ninevite 5 date.
Normally, however, some additional items were present. We have considered personal dress and ornamentation (above, 9), but we should now turn briefly to other types of goods placed with the deceased.
io.i Pottery
Pottery vessels are naturally the most common of surviving funerary objects (if they arc always to be considered as such, for in many cases it may be that the offering was not the pot itself but what it contained: cf. below, 11). Pieces in a single grave range from one (or none) to thirty-one (Tell Rijim, Tell Leilan Pit I), thirty-five if we regard Leilan Burial 1 and Pit 1 as a single tomb. Normally, however, the quantity of vessels falls into the range of two to five.
Vessels of differing sizes were sometimes placed one inside another, as in Mohammed Arab 54V:23, Tell Leilan Pit 1 and most elaborately in the adult grave at Tell Rijim. These "sets" of crockery may perhaps be compared to the four- to five-piece "sets" in Early Dynastic III graves at Kish, Mari and Abu Salabikh (Moorey 1976: 163-164, with references: Postgate 1977: 277), interpreted as for food or drink production (Moorey 1976: 164) or possibly for washing (Postgate 1980: 74; cf. below 10.4). The Ninevite 5 examples, however, show greater diversity in the make-up of the "sets", and may represent a convenient way of cramming more ceramics into the grave rather than "sets" of vessels that were necessarily associated (cf. below, 1 i). Presently attested crockery "sets" in Ninevite 5 burials are:
ui A pair of vessels, small cup inside the mouth of a larger jar (Mohammed Arab 49V: 18, 51U: 176. 54V:23; Tell Rijim child burial [Bielinski, this volume|: Tell Thuwaij [see Fujii et al. 1989-90, figs. 5: 8, 12; 13: 8]). Here the contents of the jar may have been poured off, as required, into the cup for consumption, while in the meantime the cup acted as a kind of ltd over the jar. In al! cases these sets are located next to the skul! (cf. below, 11).
531
10.4 Stones
Natural stones are known from four burials, largely with obviously different functions. In Mohammed Arab 50T: 172 a pebble was used as a stopper for a jar, while a pebble found next to the pelvis of the young male in grave 61W:17 may have been accidental (although cf. the "symbolic" pebbles of Abu Salabikh Grave I: Postgate 1980: 75, 1985: 21, 36-37).
Grave 50V:28 at Mohammed Arab and Burial 1 at Leilan contained large stones, the former of rough limestone. Such slabs are common in Early Dynastic burials in southern Mesopotamia, for example at Kish, Abu Salabikh and in the Diyala region. Mackay (1925: 15, 1929: 132, 137) suggested that they were "rubbing stones" or "foot rubbers". "Evidently", he added (1929: 137), "both men and women were unaccustomed to sandals, and required some aid to remove hard skin from the feet." Delougaz (Delougaz, Hill and Lloyd 1967: 59) writes that "such stones were commonly found in graves and are difficult to explain. It seems more probable that they represented grinders or millstones and had some symbolic meaning." Postgate (1980: 74), however, has questioned this interpretation on the grounds that a grindstone for the deceased might well be supplied from the actual grinders of coarse black volcanic rock plentifully available, while at Abu Salabikh most of the stone slabs in the graves were of coarse sandstone, unsuitable for grinding. He goes on to suggest, rather tentatively, that as the stone usually was positioned away from the body, with the other grave goods close by or upon it, and including pottery "sets" thought by him possibly to be associated with washing (cf. 10.1), it could be an ablution slab. The case of Mohammed Arab 50V:28 would seem to add considerable weight to this interpretation: the stone is again set some way from the body, upon a specially constructed brick shelf, and with a jar (for pouring water?) placed conveniently beside. In the Tell Leilan instance, the stone is not set away from the body but is lying over it, yet not perhaps inconveniently for washing, while a jar is also placed close by (and may in fact have originally stood upon the stone?).
10.5 Food offerings
That the position of vessels held up close to the mouth has implied to some the action of drinking or eating, has been considered above, with the suggestion that perishable food may have been held to the mouths of those bodies whose hands are held before the face (above, 8). There is direct evidence for food offerings, however, in the (presently unidentified) animal bones from grave 5 Hi: 176 at Mohammed Arab and in some grains of carbonised cereal ("probably wheat") found lying on the skeleton in grave G65 at Chagar Bazar (Mallowan 1936: 58). Carbonised seeds found in Burial 1 and Pit 1 at Tell Leilan were presumed to be accidental in the fill and to derive from excavation into the earlier stratum (20) of ash and burnt grain, into which the grave was cut (Schwartz 1982: 37-38).
n. "Function" and disposition of the grave goods
There are several possible, not mutually exclusive, functions of grave goods. First, the item may be so intimately associated with the deceased that its loss by burial is natural since it would not be used by anyone else. In a poor society there are not many items which may fit into this category, but possibly some simple ornament, especially clothes, fastenings, simple tools and perhaps a child's toy. Secondly, grave goods may be included for the use of the dead, either for their own personal use or on the journey to the afterlife or in the afterlife itself, or as gifts to the gods and other authorities of the underworld. There is clear textual evidence for the latter use from the Ur III period (Castellino 1957, Kramer 1967) and apparently even already from the Early Dynastic III period (Kramer 1944: cf. Postgate and Moon 1982: 131, n. 17). Lastly, grave goods buried with the dead during a public funeral may be a form of conspicuous consumption on the part of the deceased's family, an indication of wealth or rank or both. According to Tainter's (1978) researches, the amount of energy expended in grave construction is the best indicator of rank. It may be that the nature of the grave goods is the best indicator of wealth. To be sure, the highest rank often goes with the greatest wealth, but what is interesting is how children are buried and to what extent they inherit their families rank in society along with its wealth.
At this temporal and cultural distance it is notoriously difficult to discover the intention and "function" of many of the grave furnishing, especially the various types of vessels, particularly as "function" in the grave is dependent upon prevailing religious belief, of which, especially for a non-literate culture, we may know next to nothing. In discussing the Early Dynastic burial customs at Abu Salabikh, however, Postgate (1980: 77) bases himself on the (large, but perhaps not unreasonable?) assumption that "the grave goods are meant to accompany the dead person into an after-life, and to be at his or her disposal there", and — if only to make some headway in our suggestion of possible functions for the grave furnishings, we will make the same assumption here. We do know, at least, that in third millennium Sumer there seems to have been a belief that the dead were accorded differing treatment upon their arrival in the underworld dependent upon whether they had received proper burial (Pollock 1983: 80. with references).
Postgate (1980: 77) divides the mortuary goods into five general categories, namely personal body-ornaments, other personal items, possessions intended for the use of the deceased but not personal to him or her in life, food and drink for the consumption of the dead person and food and drink for the consumption of "others". In a later work (1982: 131, n. 17), he adds, on the basis of the textual evidence we have noted above, a sixth category, namely "gifts for the denizens of the underworld." For the Ninevite 5 graves the first and second of these categories may not be so distinct, for we have seen that cylinder seals might be actually worn upon the person. The sixth category may help to explain those items apparently of "personal" ornament but which have nevertheless been found at some distance from the body, such as two of the copper or bronze pins from grave 54V:23 at Tell Mohammed Arab (Fig. 21). The third, fourth and fifth of the groups involve in the main pottery, and we shall now turn to a consideration of the numbers and positions within the graves of these vessels as an indication of to which of these groups they belong. In this consideration, we may assume, again with Postgate (1980: 78), that "the more special items were placed near the head, then near the feet, with any remaining grave goods accommodated elsewhere in the grave as space allowed" and that "there is a loose correlation between intended function and position within the grave."
We have discussed already the jars positioned as if held by hands to the mouths of the deceased (above, 8), and these must be classed as food or drink (possibly food) offerings for the personal consumption of the grave occupant; as we have suggested, even where no vessel is present, it is possible that an offering of food was provided. It is not uncommon also for a larger jar, almost always with a small cup stacked within its mouth (an exception is in the Tell Rijim adult burial), to be located close to the head (Mohammed Arab 49V: 18, 51U:I76, 54V:23; Tell Rijim child burial; Tell Thuwaij; cf. above, 10.1). Presumably this was also for personal use, and if the small vessel in the hands was really provided for eating, might this larger jar and small cup be for the provision of liquid to drink? The presence of a second large jar somewhere close to the body (also near the head in Mohammed Arab 49V: 18, 54V:23; on a platform to the side in 50U:28; cf. also, perhaps, 49V:20; Tell Leilan Burial 2, and possibly the two miniature jars in Mohammed Arab 64X:44) might be taken as in someway completing the "set" of items for "personal" use, perhaps not so much this time as a food or drink offering, but as a provision of washing water; in the case of Mohammed Arab grave 50V:28 (and cf. Leilan Burial 1), there is some possible supporting evidence for such a conjecture in the presence of a targe stone (ablution slab?) (see above, 10.4).
The multi-piece crockery "sets" involve larger bowls used as storage vessels for the collections of smaller ciips and jars; a pedestal bowl without smaller vessels stacked inside was found in the fill of the shaft of Mohammed Arab grave 61W: 17, but that is exceptional (cf. above, 10.!). These "sets" of pottery seem to have been positioned more remote from the body, often in the grave shaft (Mohammed Arab 54V:23. Leilan Pit 1, Tell Rijim aduit burial). With these large collections of ceramics, it is difficult to imagine, as Postgate (1980: 76) says for the masses of conical bowls in Early Dynastic graves, that they were required for the personal use of the deceased. For the conical bowls, Postgate (ibid.: 77) suggests two possible explanations, firstly that the numbers of vessels related in some way to the number of mourners bringing offerings as a sign of respect to the dead, or, alternatively (and this is the explanation he favours), (hat the vessels were required for the continuing social commitments of the person after death. In either event, we should perhaps in this case see the pots themselves, rather than their contents, as grave furnishings, and as a general indication of the "esteem" and/or "social status" of the deceased during his or her life (cf. below, 15). It is true that one grave, Mohammed Arab 49V: 18, though appearing to be the burial of a high-ranking individual in other ways (the elaborate and labour-intensive construction of the tomb [6.5] and the necklace of semi-precious stones around the neck of the occupant) contains only three pieces of pottery, but this might be explained in this case by the age of the deceased (a girl of some nine or ten years): could it have been imagined that in death as in life she should not, as a minor, be undertaking the heavy social commitments that her standing might in the case of an adult dictate?
In summary, then, we may indulge our tentative, admittedly speculative, interpretation of the role of the grave furnishings in the land of death. In his or her grasp, directly or in a small vessel, the deceased holds a morsel of food to the mouth, as if to eat. Behind the dead person's head may stand a large jar, full of water or other liquid for drinking, a small cup resting on the rim for use as a drinking mug. Nearby, a second jar may contain water for the deceased's personal washing, sometimes also an ablution slab. So much for the person's basic needs in the afterlife, as in this. If the deceased is a member of more elevated society, however, and is of adult age, he or she will also require a smaller or larger "dinner service" to facilitate the discharge of those weighty social commitments in the aprcs-vie. In contrast, for the poor, such as the woman from Mohammed Arab 62X:08, they may have had to make do with their hands for drinking in the afterlife, just as they probably often did while alive.
12. Secondary deposits
The deliberate breakage of funerary objects as part of the ceremonies attending upon the committal of the body to the grave is well attested across various cultures and periods (Grinsell 1962, 1973), and Postgate (1980: 78) has interpreted broken vessels lying upon layers of reeds within the fill of the shaft of an Early Dynastic grave at Abu Salabikh as reflecting some such ritual: "we had little doubt that they had been thrown into the grave from above, either breaking as a result of the fall, or, it seems likelier, deliberately broken beforehand" (cf. also Postgate 1985: 150). As we have mentioned above (10.1), Mallowan has suggested that a similar rite accompanied one of the Ninevite 5 burials at Tell-Chagar Bazar, and this is not impossible, for unlike the grave with an incomplete vessel at Tell Mohammed Arab, this Chagar Bazar burial (G68) was not otherwise so impoverished that a broken vessel would be expected to have been included (see Fig. 5). On the other hand, it is not difficult to think of other possible explanations, such as a personal attachment of the deceased to that particular vessel, or, simply, a lack of concern, perhaps, for the condition of what may have been regarded only as a carriage vessel for the smaller pieces (cf. 11).
13. Age and sex distinctions
Discussion of patterns of mortuary practice dependent upon the age or sex of the deceased is greatly frustrated by the lack of knowledge of physical anthropology by excavators working in the Near East, and the resultant low level of recording and reporting of osteological data. In the present case we are, to be sure, fortunate in our sample in having been able to make detailed records on this topic for the burials from Tell Mohammed Arab, but a paucity of male burials prevents us determining any likely sex distinctions. For age distinction, of course, the situation is not so hopeless, because anyone can make a rough estimate of the age of a skeleton, and excavators often do distinguish in their reports between the burials of adults, children and infants.
For indication of the sex of grave occupants, a common method has been to rely upon the nature of the grave furnishings. Commenting upon the Early Dynastic cemetery at Al-'Ubaid, Woolley notes that "If the body were that of a man, he might have with him his weapons or tools, if a woman, beads, a spindle-whorl, eye-paint, or rouge" (Hall and Woolley 1927: 173), while for the "A" Cemetery at Kish, Mackay (1925: 13) is even more explicit: 'The identification of sex was based on the grave goods, and nine graves contained bodies which must have been of the male sex, for they were accompanied by such masculine equipment as battle-axes, daggers and adzes. Articles for feminine use, such as spindles, needles, toilet cases, hair-pins, bracelets, etc. were found in sixteen graves." That these assignments of sex on the basis of grave goods have been culturally biased by present norms ornament is proved by Rathbun's (1975) analysis of many of the skeletons from the "A" Cemetery at Kish. Comparing Rathbun's findings with the excavator's lists of grave goods, it is obvious that men were commonly buried with jewellery, hairpins and shells containing cosmetics.
We must therefore confine ourselves in this respect to the data for burials at Tell Mohammed Arab, where osteological determinations are available (above, 3.5). Here we have, probably by accident, a preponderance of female internments, so any real comparison of male and female burials is not presently possible. The most we can do. perhaps, is to reflect upon what items were not inappropriate to the burial of a woman (but which may have been also in the graves of males). They do not, to be sure, contradict what Mackay and Woolley assumed for the Early Dynastic burials of the south (both of them realised that cylinder seals were present in female as well as male burials: cf. above, 9).
Very tentatively, since they are common in the female burials (they are present in five out of seven graves) and are entirely absent from the (four) male burials, we might think that the metal pins are, in this case, a "female" funerary attribute. With such a small sample it is not possible to be sure. If, however, in the light of further evidence, this should prove to be the case, it may be that a number of the Tell Chagar Bazar burials, including some of the "wealthiest", were the graves of women.
Little more positive can be said with regard to age distinctions, even though we have in our sample a number of children's and infants' graves (Chagar Bazar G7i, Leilan Burial 3, Mohammed Arab 49V: 18, 50T:173, 64X:44, Tell Kutan, Kan-ana 3 Burial 10; cf. also Mohammed Arab 50V:28, 50T:172 [shared adult/child graves], SOT: 180 [grave of an adolescent, of about 13-15 years]. Two children's burials without grave goods were also assigned to level 4 at Tell Chagar Bazar.) On the whole, all that can be said about the goods in these graves is that they do not differ from those in adult burials, except that we do not have in our sample infants' or children's graves with large quantities of pottery, even when there are other indications of wealth or social ranking (cf, above, 11; cf. also, perhaps, Chagar Bazar G71, marked out as above average by the presence of a stone bowl). It is of interest to note again, in this context, the pair of miniature vessels in Mohammed Arab grave 64X:44 and in graves at Tell Kutan (cf. 10.2), but the occurrence presently only in graves of children of clay sealings with stamp seal impressions (Chagar Bazar G71,
Mohammed Arab 50V:28) is probably fortuitous, as one imagines that they are accidental inclusions in the grave fill. With the exception of the very individual grave 50V:28 at Mohammed Arab (which also included a child), pot-burials are, not surprisingly, restricted to babies (6.6). See also under 15 below.
14. Family relationships and pathology
Detailed osteological data are presently available only for the Tell Mohammed Arab burials, but they do yield some interesting information about family relationships at the cemetery site and pathology. A number of diseases ranging from mild arthritis and dental disease to possible tuberculosis and calcium-deficiency related disorders give a picture of the state at death of this portion of the Ninevite 5 population. Full details will be published in the excavation reports, but in the meantime it is worth mentioning those instances where there was evidence for family relationships between the deceased.
Graves 50T.T72 and 50V:28 were both "dual" burials of an adult and child (Figs. 13, 17). The former burial contained a 23-27 year old woman and an 8-10 year old girl, while the other burial was of a 33-37 year old man and an 8-10 year old boy. In 50TT72, both skeletons displayed abnormal pitting on the bones of the pelvic region. The cause for this is unknown; it may have been congenital in which case this would be a good argument for a close kin relationship between the two, perhaps mother and child. Similarly, for 50V:28, the child's eye orbits exhibited a mild degree of cribra orbitalia (a condition of uncertain origin, but probably related to childhood malnutrition, although there is evidence for a hereditary propensity) and the man also had slight cribra orbitalia. Both skeletons exhibited slight congenital abnormalities in the vertebrae; unfortunately, due to poor preservation, it was impossible to make direct comparisons. This is some evidence, however, that the two were related.
Of course, the really interesting question with these double burials is whether the individuals buried together died at the same time and if so were they deliberately sacrificed or did they die by accident or of disease? There does not seem to be any clear evidence, however, as to whether the burials were simultaneous.
is. Wealth and social ranking
There has been a revolution in mortuary studies in the last twenty-five years "by the realization that varied treatments accorded individuals in death are related systematically to social differences that existed among the living ... This new interest has stressed the use of mortuary evidence, not as a basis for fanciful speculation on primitive beliefs and religion, but as a means of specifying in detail the nature of extinct social systems" (O'Shea 1984: xi, 1). As we have already stressed, however, the application of this theory to Ninevite 5 burial practices is largely frustrated by the very small size of the sample. Schwartz (1986: 56) concluded that "The mortuary evidence from northern Mesopotamia is at present too sparse to resolve the question of the presence or absence of social stratification. The development of stratified, urban society in the early third millennium can neither be definitely attested nor disproved with mortuary data."
However, that a certain social stratification did exist is evident from a comparison of the more meagre graves with those rarer cases of above-average wealth, such as some of the Chagar Bazar burials (perhaps G67, G68), Mohammed Arab grave 54V:23/Tell Leilan Burial 1/Pit 1. and the adult burial at Tell Rijim. In all of these cases the quantities of pottery are large enough to suggest, perhaps, either great devotion from mourners or, more directly, that the social ranking of the individual entombed was at a level to require such goods for "entertainment" purposes (cf. above, 11). Moreover, some of the personal items in these burials, the three cylinder seals, stone vessel and beads (including one silver) of the woman of Mohammed Arab grave 54V:23, the cylinder seal, pins and beads in Tell Leilan Burial 1, the pins and pendant in Chagar Bazar G67 and the silver bead and copper sickle in G68, would seem to mark out their occupants as above the average in wealth or adoration. Furthermore, some of the graves of children, which appear never to contain much in the way of pottery, can still be marked out as the burials of socially elevated individuals (or the children of socially high ranking families?) in the comparatively wealthy personal items they display and, perhaps, in the degree of elaboration and effort of labour in the construction of the tomb (Karrana 3 Burial 10, Mohammed Arab 49V: 18) (see 11). Interestingly, as well as children, deceased women might be well provided for in the afterlife (see especially Tell Mohammed Arab 54V:23 and cf. the "rich" possibly female graves from Chagar Bazar, 13) while impoverished men might take with them to the next life little or nothing (Mohammed Arab SOU: 180, 61W:17). Social status, therefore, does not appear to have been rooted primarily in sex or age differentiation, and that a woman could possess cylinder seals may be an indication of female property rights (?) (cf. above, 9).
This is the most that can be said: suggestions of a high degree of social stratification based upon the evidence of the supposed Ninevite 5 vaulted tombs at Nineveh can. for the present at any rate, be discounted (above. 3.1).
Schwartz (1986: 53-55) compared the picture from Ninevite 5 excavated burials with that known from southern Mesopotamia in the Ubaid to Early Dynastic periods. His conclusion was that "There are no Ninevite V graves with wealth to match those of the richer Ur Royal Cemetery tombs or the unusually wealthy graves of Kish A Cemetery or Khafajah. Further there are no graves with the large concentration of metal vessels and objects of precious stone or metal found in the majority of datable Ur Royal Cemetery graves. The richest excavated tombs of the Ninevite V period contain beads, copper pins, and perhaps [sic] a cylinder seal, but not the rich collection of metal containers or stone vessels from wealthy southern tombs." He goes on (ibid.: 55-56) to suggest, however, that "It may be that richer graves exist in the north but have yet to be excavated. ... considerably different results could be obtained from the excavation of burials in an urban site." This is a possibility, and it might be noted that even the picture from the south can be distorted by consideration only of published burials at the "richer" sites (for a brief general review of the archaeological evidence, with comparisons of the grave goods, see Pollock 1983: 82-86): if one looks at a relatively impoverished urban site like Abu Salabikh (which Schwartz did not consider in his review), the number of graves with metal or stone vessels is, as Martin (in Postgate 1985: 16-17, with n. 19) has stressed, very few (she even suggests that there may have been a lack of publication of the robbed or disturbed graves in earlier reports).
Nevertheless, in the early third millennium B.C. of northern Mesopotamia, it is perhaps unlikely that there would be such large concentrations of wealth and power as in the southern cities of this time. The northern cities were not and could not be as large as the southern ones because the rainfed agriculture of the north would not support such big conglomerations of people. Therefore, we think it unlikely that we will ever find "royal" graves like those at Ur. Although potentially rich trading depots of the north, such as Gawra and Nineveh, could have had a considerable amount of social stratification in terms of wealth ■— ranging from poor beggars to wealthy merchants — even so their administrative/ governmental structures, including state supported religious cults, were very likely not nearly as elaborate as those in the southern cities. This is why one may expect quite a variety in type of grave goods, ranging from nothing to ornaments made of precious metals and semi-precious stones, but one does not expect, perhaps, a vast difference in grave construction or placement.
Two possibilities remain: either that Ninevite 5 burials are, in comparison to those of southern Mesopotamia, relatively impoverished because they reflect a society and culture that itself was so (in which case the very title of this conference volume might be doubted), or that the present limited sample does not reflect the true picture of levels of wealth and poverty in the Ninevite 5 period. Only further work, or evidence from sources other than the funerary data, can hope to find a conclusive answer to this important and intriguing question.
Nevertheless, we hope that what suggestions are made in this paper on various matters concerning the burial practices of the Ninevite 5 people, dependent upon a degree of assumption and speculation though some may be, will offer the possibility of new ways of looking at the scanty but not entirely uninformative evidence available, and we therefore offer these thoughts for consideration.
Burial plans and grave groups: captions to illustrations
NB. In the case of Tell Mohammed Arab (Figs. 13-24), numbers used for the identification of objects from the graves (marked on the plan and beneath the object sketches) are, at the request of the excavators, the original field numbers; in the case of burials from other sites, however, these identification numbers are arbitrary designations peculiar to this publication and used for convenience only: No. 1 is in these cases reserved for the skeletal remains (even where no plan is available), with other numbers following for the grave goods, etc.
TELL BILLA
a b c
- Fig. 1: Published vessels from graves at Teil Billa assigned by Ihe excavators to Ninevite 5 level 6.
Atter Speiser 1933: pi. LIU: 4, 6, 7.
537
- Fig. 2: Grave groups from burials assigned to T. Chagar Bazar levels 5 (a, G72) and 4 (b, G48; c, G50) which do not appear distinctively Ninevite 5. \fter Mallowan 1936: (a) fig. 10: 15; (b) figs. 7: 24, 7: 25, 11: 13, 15: II; (c) figs. 7: 5, 9: 2, 13: 12; (d) fig. 15: 2
TELL CHAGAR   BAZAR G60
2 3 A 5 6
- Fig. 3: After Mallowan Í 936: figs. 10:4, 10: 11, 10: 10, 11: 1, 19:5.
TELL CHAGAR  BAZAR G67
7 8 9 10
11 12 13 %
- Fig. 4; After Mallowan I936: figs. 7: 10. K-: 2, K: 7. K: 20. It): 16, 10: 17. 9: 15. 28; 6, 18: 4. 18: 3. IS: i. 19: 2.
538
TELL CHAGAR BAZAR G68
TELL CHAGAR  BAZAR G70
w Í
Fig. 6: After Muliowan 1936: figs. 8: 5, 15: 8, 19: 7.
539