Questioning the BI-Ware - article ; n°1 ; vol.18, pg 303-334


33 Pages
Read an excerpt
Gain access to the library to view online
Learn more


Travaux de la Maison de l'Orient - Année 1990 - Volume 18 - Numéro 1 - Pages 303-334
32 pages
Source : Persée ; Ministère de la jeunesse, de l’éducation nationale et de la recherche, Direction de l’enseignement supérieur, Sous-direction des bibliothèques et de la documentation.



Published by
Published 01 January 1990
Reads 28
Language English
Document size 3 MB
Report a problem

Jean-Francois Salles
Questioning the BI-Ware
In: Failaka, fouilles françaises 1986-1988. Sous la direction de Yves Calvet et Jeacqueline Gachet. Lyon : Maison de
l'Orient et de la Méditerranée Jean Pouilloux, 1990. pp. 303-334. (Travaux de la Maison de l'Orient)
Citer ce document / Cite this document :
Salles Jean-Francois. Questioning the BI-Ware. In: Failaka, fouilles françaises 1986-1988. Sous la direction de Yves Calvet et
Jeacqueline Gachet. Lyon : Maison de l'Orient et de la Méditerranée Jean Pouilloux, 1990. pp. 303-334. (Travaux de la Maison
de l'Orient)
Jean-François SALLES
In the very beginning of her book about the Hellenistic pottery from Failaka, Lise
Hannestad refers to the BI-Group which she defines as follows :
« a special group, characterized by a pale yellow clay (Munsell 2.5 Y8/4.8/2) and a heavy,
dark green glaze which decays into a dark brownish-yellow. To a large extent the group
uses other shapes than those seen in the majority of glazed ware, and very characteristic
are also the more elaborate forms with pronounced ring feet and richly profiled runs »
(1983: 14).
The various pots and sherds which belong to this group are described and discussed
at length in each section of L. Hannestad's study, and the conclusions are stated :
« The BI-Group of glazed ware, which clearly represents a late and limited habitation on
the site, has so far been reported only from Susiana where it is supposed by Haerinck
to belong to the late Parthian period (1st cent. A.D. to ca 225 A.D.). On Failaka, a date
no earlier than the later part of the Ist' cent. B.C. and probably in the Is' cent. A.D. is
confirmed by the find of Nabataean bowls together with this ware, whereas the two
imitations of ESA bowls form 19 (cat. nos 36-36) found together with the BI-Group rather
point towards the Ist cent. B.C. than later. A dating about the beginning of our era is
thus the most probable.
Period II : probably very late Is' cent. B.C. into the first cent. A.D. » (ibid. : 78).
After three seasons of excavation on about 450 m2, which are nearly a third of the
surface excavated by the Danes, the French archaeological mission was unable to identify
a single sherd of the BI-Group pottery. We discussed the problem several times with
Lise Hannestad who kindly provided us with some samples of BI-Ware in order to get
true and physical references1. However, our investigations proved fruitless. On the other
1. I would like to express my deepest gratitude to Professor Lise Hannestad, University of Aarhus, for her
tireless encouragements in my approach of the BI-Ware. We have been discussing the question (and many
others, too !) lor several years ; she kindly allowed me to use all the drawings of her publication. Overall,
she corrected and commented at length the first version of this paper, letting me avoid a lot of mistakes
and misunderstanding. I have thought useful to introduce most of her recent remarks in the present paper;
they are being referred to by her initials, L.H. I am pleased to dedicate this paper to Lise Hannestad as
a token of friendly regard.
During a meeting organized by Professor Uwe Finkbeiner in Tübingen in June 1987, a preliminary
account of the present paper was presented and discussed with several colleagues working in the Gulf
area R. Boucharlat, K. Ciuk, J. Gachet, E. Haerinck, E. Keall, M. Kervran, O. Lecomte, B. and T. Leisten,
D. Potts and E. Valtz, and Lise Hannestad as well. Although it benefited very much of various interventions,
the present paper does not present the final results of the meeting on this special subject, and reflects
my own assumptions only ; therefore, possible mistakes or misinterpretations are under my unique
M. Kervran and D.T. Potts were kind enough in allowing me to publish drawings of some of their
unpublished material : my warm thanks to both of them. R. Boucharlat, Y. Calvet and P. Lombard read
the first version of this paper and enriched it with their comments, as well as D.T. Potts and E. Valtz who
corrected some of my misinterpretations.
Failaka: 1986-1988
TMO 18, Lyon, 1990 304 J.-F. SALLES
Figure 1. Catalogue of the Bl-group, after Hannestad 1983 : A. The Glazed BI-Ware.
Cat. Shape Glazed Usual glazed References
BI-Ware ware in of location
31 bowl with angular X WE 62-70
profile and BI and BJ
outturned rim
rounded bowl X 35 AC LP
36 id. X WE 82-90, BY + WE, 62-
70 FN, C
53 id. X WE 62-70 BI
54 id. X WE 62-70 BI and BX,
WE 82-90 BY
55 id. X WE 62-70 BI
56 id. X WE 62-70 BI and DF, WE
82-90 BY
57 id. X WE 62-70 BI + WE 82-90
58 id. X WE 82-90 BY
125 bowl with X WE 64, 38 AV
flaring rim
and offset lip,
variant 2
126 id. X WE 52-60 FE
Illustration non autorisée à la diffusion 141 X id., variant 3 0 AIP
175 large rounded X WE 62-65 BI
176 id. X WE 62-70 BI + C
177 id. X WE 62-70 BI
179 id. X WE 62-70 BI
180 id. X WE 62-70 BI
181 id. X WE 62-70 BI + D
182 id. X WE 62-70 Β I
183 id. X WE 62-70 BI
184 id. X WE 62-70 BI
185 bowl X WE 52-60 DN,
WE 82-90 BY,
WE 62-70 BX
187 beaker X D-F ext. s.
189 X WE 65, 90-70 BK + C + A-C drinking cup
190 id. X ? A-C + C + C ext. n.
204 X fish-plate C
205 id. X WE 62-70 BI
Figure 1. Catalogue of the BI-Group, after Hannestad 1983
WE 62-70 BI-BX, WE 82- 207 id. X
90 BY
X C + F5K BI 272 amphora
287 id. X C
X C + C ext. n. 290 id.
WE 82-90 BY + WE 62- 291 id. X
70 BX + AC + C
X WE 62-70 BI 315 pilgrim flask
handle ? X C 333
vol. 1, p. 44
X heavy foot
vol. 1, p. 26
B. Related
400 Arabian WE 62-70 BI
WE 62-70 BI 419 Nabatean
Illustration non autorisée à la diffusion Ware
WE 62-70 BI 420 id.
421 id. WE 62-70 BI
422 id. J
id. C ext. n. 423
424 id.
425 id. Q
426 id. Q
499 WE 62-70 BI Commonware,
closed shapes
WE 62-70 BI 510 storage jar
WE 82-90 BY + C 537 glazed
WE 82-90 BY + AC 538 id.
red WE 82-90 BY 553
cooking-ware J.-F. SALLES 306
hand, the new stratigraphie sequence described in the previous chapter by Jacqueline
Gachet shows a late level (Period 4 /Stage V) during which a scattered re-occupation
of the site was found on the abandoned ruins of the earlier levels : it may well be the
« limited settlement set up in front and in the pronaos of Temple A, which was by now in
ruins » (ibid. : 78), which is characterized by the BI-Ware. It might as well be something
different, since the pottery from our Period 4 is similar in all its aspects to the corpus
of the earlier levels, without any BI-Ware, and although no stratified finds could help
us in dating this late level, two Characenian coins point to a date in the second half
of the 1st cent. B.C. The question, thus, can be formulated as follows :
- a) our Period 4 is not L. Hannestad's Period II2, and the stratigraphy of the site includes :
Hannestad's Period I (our 1 to 3)
our Period 4, late 1st cent. B.C. Period II, end of 1st cent. B.C. into 1st cent. A.D.
- b) our Period 4 is L. Hannestad's Period II, and we have to slightly revise the datation
proposed for the BI-Group of glazed ware, which cannot exceed the beginning of our
era. It may also imply a revision of its interpretation (origin, comparisons, etc.), especially
in the light of new research at Seleucia-on-the-Tigris3. It is the purpose of this paper.
L.H. rightly comments on this perhaps 'fallacious alternative. In case a), our Period 4
could well fall into her Period I, since the pottery is identical in both. If so, the end of
Period I, which L.H. assigned to the early 1st cent. B.C. on the basis of a Hyspaosines
n° 67), coin and the terracotta figurines of so-called Parthian kings (Mathiesen 1982,
should be redated to the end of the 1st cent. B.C., in accordance with the abovementioned
Characenian coins - which, I should stress it, are not stratified. Consequently, L.H.'s
Period II should be definitely pushed into the 1st cent. A.D. In case b), why should our 4 be identical with L.H.'s Period II since pottery types are different, the BI-Ware
being absent from our Period 4 ?
Without. developing counter-arguments, I would like to schematically sum up the two
stratigraphies :
French excavations Danish excavations
Period I (mid-3rd cent. Period 1 to 3 (early 3rd cent.
early 1st cent.) B.C. mid-2nd cent. B.C.)
hiatus hiatus
Period II (late 1st cent. B.C. Period 4 (probably second half
into the 1st cent. A.D.) of 1st cent. B.C.)
I would be tempted to identify both hiatus as a single period. However, I understand
L.H.'s arguments, and we could suggest the following sequence :
Period I (more or less Periods 1 to 3)
Period 4
Period II
Actually, the elucidation of this puzzle needs further investigation, both in archaeological
and historical fields. It does not bias against the comments I will develop about the BI-
Ware, which are committed to ceramics more than to stratigraphy.
2. Periods I and II, in Roman characters.belong to Lise Hannestad's chronology ; Periods 1 to 4, in so-called
Arabic characters, belong to the French stratigraphy.
3. My deepest thanks to Dr Elizabetta Valtz, Museo Egizio, Torino, who gave me access to her latest discoveries
at Seleucia-on-the-Tigris and allowed me to mention and illustrate them. :
The criteria which allow us to describe a pottery or a sherd as belonging to the BI-
Group, and more specifically to the Bl-type of glazed ware, are its fabric and its shape ;
furthermore, this group was found on a very precise spot of the site, although it cannot
be considered as a criterion. The catalogue of potteries explicitely assigned to the BI-
Ware by L. Hannestad includes 28 numbers (Fig. I4 and Fig. 2).
The location on the site (Fig. 1 and Fig. 3)
The term BI-Group comes from as grid reference of the I960' excavation and refers
to a place where most of this pottery was identified, in the long West-East central trench
T, in front of Temple A (WE 62-70 BI, and also BK and BX) ; a few pieces were found
in Trenches C and D as well, which correspond to the area extending from Temple A
to the rectangular altar to the East (176, 187, 204, 272 287 333). This type of pottery
rarely comes from inside Temple A but is not unknown there : 35, 189, 190, 291. The
mending of some vases associates various trenches, even « north extension of Trench
C » which is a dwelling area (190). Such a distribution leads L. Hannestad to conclude :
« The area in front of Temple A seems to be the onlv one left open during the many
rèbuildings of the fortress » (1983 : 14) and : « By the time of the BI-Group, the temple had
clearly fallen into disrepair ; the roof had probably fallen in » (ibid. : 123, note 10). Another
place was rich in Bl-Ware, in the eastern end of the central Trench T, grid reference
WE 82-90 BY : it is located on top of the moat, « which seems to imply that at this time
the moat had more or less been filled in » (ibid. 14).
Only two noticeable exceptions to this limited distribution can be traced. The bowl
141, which is clearly identified as Bl-Ware although a very common shape in earlier
levels, was found in Trench 0, near the circular altar facing Temple B. On the other
hand, three fragments of Nabataean bowls, a ware which is a mark of the BI-Group,
came from Trench Q, that is the filling of the well in the southern part of the fortress.
Lastly, we must mention that several sherds with a BI or BY grid reference are not
labelled as Bl-Ware and represent the « normal » glazed ware : 36 ?, 55, 57, 58, 181, 185, 315.
This restricted distribution of the BI-Group may explain why we did not recover it
in the upper layers of our excavation, in the north-eastern quarter of the fortress.
However, if we assume that L. Hannestad's Period II equates with our Period 4/Stage
V - what is not argumented yet -, the very limited location of the BI-Group may appear
somehow perplexing : why it is impossible to find Bl-Ware fragments anywhere else than
in front of Temple A despite the fact that several other places of the fortress should
have been settled at the same time requires an explanation.
Reconsidering the whole Hellenistic material stored in Failaka Museum's reserves, I
discovered recently a handful of presumably Bl-Ware sherds which were misinterpreted
at the time of their discovery5 they all come from the beach sanctuary Β 6, and although
4. In Fig. 1, Catalogue of the BI-Group, I have included all the potteries which were labelled as Bl-Ware in
L. Hannestad's descriptive catalogue (vol. 2), as well as all fragments with a reference of location « BI »
or « BY ». We must stress that the BI-Group includes glazed Bl-Ware and usual glazed ware as well, such
as 185, 290, 315, etc. On the other hand, fragments of Nabataean Ware, Arabian Red-Washed Ware,
Commonware and Cooking-Wares are part of the BI-Group according to their location on the site.
5. Caubet and Salles 1983: cat. 145, 158, 282, 358 and 359. « The only missing shapes are the ones of the BI-
Ware and the large bowls with rounded base which only appear in the later levels of the fortress » (ibid : 126).
The datation of the site was not based on the unique evidence of the pottery, and there is no positive
reason for correcting the date of Β 6 sanctuary, well established by the coins and other finds. Sherds 145
and 158 were found in Β 6-Period 1 level (early 2nd cent. B.C.), and 282, 358 and 359 in Β 6-Period 2
horizon (first half of the 2nd cent. B.C.). Therefore, we are facing an alternative
- these sherds might come' from unstratified contexts, what seems refuted by the excavators, and belong
to a later re-occupation of the site, hardly elicited and undated (ibid. : 127) ;
- these Bl-Ware fragments really belong to stratified contexts, and we have to assume an earlier date for
the arrival of this ware on Failaka. At the same time, we have to justify why this type of pottery occurs
in the beach sanctuary Β 6 and is absent in the contemporary levels of the fortress. 308 J.-F. SALLES
Illustration non autorisée à la diffusion
Figure 2. Pottery Shapes of the BI-Group, after Hannestad 1983. See also Fig. 6, Fig. 7 and Fig. 8. QUESTIONING THE BI-WARE 309
□ □
mo m
isolated finds associated with BI-Ware
main concentration of BI-Ware
scattered sherds of BI-Ware
Figure 3. Schematic distribution of the BI-Group inside the F 5 Hellenistic fortress (Danish excavations,
1958-1963) 310 J.-F. SALLES
a: F83.1501 (FFF 1983: 359), «terre d'ombre brûlée» corroded glaze, gold iridescence, traces of
dark green glaze, Period 2 (see Hannestad 1983, 204)
b: F83.1459 (unpublished), id.
c: F83.1540 (FFF 1983: 358), light brownish corroded glaze, darker on the rim, Period 2 (see
Hannestad 1983, 54-57?)
d : F83.2105 (FFF 1983 : 158), dark green glaze with iridescence, Period 1 (see Hannestad 1983, 190)
e: F83.1529 (unpublished), «terre d'ombre brûlée» corroded glaze, with iridescence, traces of
dark green glaze
Figure 4. Bl-Ware fragments from the Β 6 beach sanctuary (French excavations, 1983) THE BI-WARE 311 QUESTIONING
their shapes do not refer to the most significant ones among the BI-Ware repertory, glaze and fabric clearly attest to their belonging to the BI-Group (Fig. 4). This
group, thus, is not limited to a precise location inside the fortress, but can be found
elsewhere on the island. I should add that Jawad al-Najjar was kind enough to show
me a few sherds among thousands from the material he excavated in 1978 on Akhaz
Island, in the Kuwait Bay (Najjar 1980) : pending a more detailed analysis, some of them
might belong to the BI-Group.
The fabric (Fig. 5)
As mentioned above, L. Hannestad describes the fabric as « a pale yellow clay, and a
heavy dark green glaze which decays into a dark-brownish yellow » (Hannestad 1983 : 14).
When going through the individual descriptions of each vase (ibid., vol. 2), we find the
following occurences when describing the clay :
- pale yellow to white, white : 5 quotations ;
- yellow, pale yellow, very pale yellow : 10 quotations ;
- very pale brown, very pale brown to pale yellow: 12 quotations;
- black and reddish-brown impurities are mentioned twice.
However, similar descriptions are frequently used in the catalogue for non-BI-Ware
glazed vases of the BI-Group, or even for potteries coming from other areas of the
fortress ; sometimes, the Munsell Chart reference is the same. The colour of the clay
taken alone, thus, cannot be considered as a safe criterion for identification of the BI-
Ware, as claimed by L. Hannestad in several circumstances.
The dark-green glaze could be a better clue. It is well preserved in a few occasions
only : large rounded bowls 176 and 182 (181 is not explicitely referred as BI-Ware), or
amphora 287. Sometimes, the covering glaze has partly disappeared, and scanty traces
only can indicate the original colour: «small spots of dark green iridescent» (175), « few
traces of the green glaze left » (178), « a few traces of dark green still left » (179, 180, 184,
etc.), « originally dark green, now corroded to dark yellow » (205, 206, 207), « the green
glaze is darker in the incised circles around the shoulder» (291). In one case, the glaze is
lighter green and shows a darker line on the rim (141). In several examples, iridescence
is visible, usually brownish. Nevertheless, apart from explicit references to a dark green
colour of the glaze, we must admit that the description of the BI-Group sherds does
not differ noticeably from the usual descriptions of other glazed wares.
Another specific feature of the BI-Ware could be a dark yellow or brownish corrosion
of the glaze when decayed, which is mentioned in most cases. When corroded, the glaze
shows a dark yellow colour, much darker than the usual glazed pottery (L.H.). However,
similar descriptions can be found in the rest of the corpus of Failaka pottery as well,
and recent excavations have yielded many sherds with a dark yellow, orange or brownish
decayed glaze, found in various levels and good stratigraphie contexts which exclude the
proposed date of the BI-Group. This kind of corrosion might be due to the chemical
conditions of the soil (Bourgeois 1983) and cannot be a specific definition of the BI-Ware.
L.H. maintains, and I do agree with her, that there is a clear general difference in the
clay between the BI-Ware and the usual glazed pottery : the clay of the first one is pale
yellow in comparison with the paler brown of the second. I quote here L.H.'s letter :
« But I would not assign any sherd to the BI-Ware on the appearance of the clay only.
In conclusion, I have identified a group of pottery in the fortress differing from the rest
of the glazed pottery in its use of a different type of glaze, different shapes and a clay
which is generally pale ». My own comments will never dispute such a basic evidence.
Truly enough, effective identification of pottery presupposes a long physical contact with
the sherds which teaches more than drawings and descriptions in the best publications ;
at Failaka, a lengthy relationship with Hellenistic sherds is not even sufficient to identify
the BI-Ware by its fabric only, as acknowledged by L. Hannestad herself, and we have
to refer to the shapes.