Jars from Arqa Phase P

The capacity diagram (Fig. 4) shows five distinct groups, probably related mainly to functional specialization. The distribution of jars with (black and grey dots) and without handles (white dots) is especially remarkable.

Group 3 is comprised of jars with handles only, and capacities ranging from 20 to 25, exceptionally 30 litres (Fig. 3: 5). Such jars weighing about 30 kilos when full would probably not be difficult to carry, although the position of the handles or the large flat bases are not well suited for this use. The vertical handles are attached low on the body, i. e. close to the center of gravity of the vessel, as is ususal in the Levant during the EB (Amiran 1969: 59, 63, 66, 67) and still at the beginning of the MB period (Amiran 1969: 104; Aston 2002: figs. 1-4). This makes it easy to move the jar or to tilt it for pouring when it rests on the ground, but does not insure verticality when carrying it by hand. such jars would therefore be better considered as short-term storage vessels, which for this reason had to be frequently displaced within the dwelling area, rather than as transport or “commercial” containers.

In a significant manner, during MB and LB, handles will be attached higher and higher on the shoulder, well above the center of gravity, and pointed or stump bases will replace the flat ones. Capacities are then frequently more important; the jars can be carried in a vertical position by two men or handled for pouring, using one handle and the stump base, by

Fig. 4 Capacity diagram for Phase P

One. One of the earliest examples of this morphological adaptation of jar handles to the logics of transport, and especially of maritime trade, is to be found in the group of jars from the Royal Tombs at Byblos (Tufnell 1969: fig. 6); their average capacity is 40 to 45 litres, nearly twice as much as the jars of Arqa group 3.

Groups 4 and 5, with capacities ranging respectively from 55 to 75 and 90 to 120 litres, are clearly distinct, although the sheer consideration of general size and proportions, as noted above, would not allow to set them apart. Because of their weight, all such jars (Fig. 3: 6-8) must be non-movable storage vessels, and indeed most of them were found filled with cereals in both destructions layers of Phase P. The two groups may correspond to the storage of different kinds of products (liquid/solid) or to different conditions (middle-range or long-term storage).

Surprisingly, handles are occasionally found on jars from these groups. A few jars of group 5, all above 100 litres, have a small loop handle from the top of the shoulder to the rim (Fig. 3: 8): while unpractical to tilt such large vessels for pouring, the handle would be well suited to attach with a rope a wooden stopper for instance, if the jars had to be frequently opened and closed. For this reason, they could be interpreted as water containers - the type is rare, and one or two such jars only were necessary in each house.

Some jars of group 4 (grey dots) have vertical handles, similar in position and shape to those of group 3 jars (Fig. 3: 6): for reasons given above, they were certainly very ill-suited for carrying the jars when full. The group is very homogeneous, characterized by the profiling of the rim, but above all by the incised and impressed decoration on the upper part of the body (Fig. 3: 6). They are found at Arqa from level 16B to level 15A, a period of about two centuries. With the exception of a few fragments from sites close to Arqa and a unique fragment from Byblos (BYBLOSII: 16572), I know of no parallels to this type of decoration and consider it most probably the work of one or two families of local potters over a few generations.

Apart from the possible symbolic connotations of a part of the decoration (“suns” or “stars” and stylized vegetal elements), the overall pattern seems to be derived from a practical device of ropes or basketry, comparable to our modern dames-jeannes. The wide “arches” of impressed dots are “attached” to a row of similar impressions on the maximum diameter of the body, to the handles and to small lugs which are otherwise part of the applied decorative elements. It is impossible to know whether such a system, which would permit to carry easily the large jars and may also have been used on their smaller counterparts of group 3, was actually used in the Levant by the end of the IIIrd millenium. It is however such a simple device that the probability is high; it would have been later abandoned with the above-mentioned evolution of pottery containers specifically adapted to trade. Representations of such devices, probably on handleless jars only, exist on cylinder-seals at the end of the Uruk period (e. g. Le Brun 1978: fig. 8: 5; Le Brun and Vallat 1978: fig. 6: 4, 9, fig. 7: 12), but I know of no later ones.

Finally, some puzzling questions arise from the capacity diagram. None of the smaller jars of group 1 (5 to 13 litres, Fig. 3: 1, 2), either because they are handleless or have too wide and short necks, appear to be adapted to the carrying of water for daily use. Everywhere in the Middle East, and especially at Arqa, where the river flows in a deep gorge at the foot of the tell some 40 to 50 m below the settlement, this was a painstaking but important task, for which one would expect to find specially designed containers. In the whole assemblage of Phase P, only very large jugs (9 to 15 litres, triangles on the diagram, Fig. 3: 3, 4), with restricted neck and trefoil mouth, meet the necessary requirements. They are too large for pouring water for individual use into the small cups and goblets which are the standard drinking vessels of the period. But they could be easily carried on the shoulder or on the head, while the restricted neck prevented the spilling out of water. This type is frequent, as can be expected for vessels with a high probability of breakage, and very many of them were probably necessary in every single house.

Jars from Arqa Phase N

The capacity diagram for Phase N shows a very different picture (Fig. 6). It is obvious that pottery production is much more specialized and standardized: only three groups. Jars with (black dots) and without handles (white dots) are represented in the first (12 to 17 litres) and the second one (20 to 30 litres), while all larger jars in group 3 (40 to 60 litres) are handleless; the very large containers with capacities of 90 litres and more seem to have disappeared. This is probably in part due to the fact that the Phase N assemblage is derived from the potters’ quarter and workshop (Thalmann 2000: 47-50; 2002: 368-369), not from an ordinary dwelling quarter. It is however noteworthy that, if large jars of about 100 litres were in use elsewhere in the settlement during Phase N, they were manufactured in a different location and by different potters; this was certainly not the case in the preceding period, when all groups of jars exhibited a strong technical homogeneity (Thalmann 2000: 44).

The jars with handles of group 2 (Fig. 5: 3) differ in shape from their counterparts of Phase P, group 3, and with the same range of capacities may have served the same purposes for short-term storage. But

They are also probably more specifically designed for transport, because there now exists a number of handleless jars in the same group, which are equally well suited for short-term domestic storage, but of course not for transport. A further indication is that we have from Phase N a number of non-local jar sherds, probably originating from the Byblos area or southern Lebanon, what can be inferred from their limestone tempering vs. the strong basaltic component in the temper of all local wares.

Handleless storage jars are more or less evenly distributed (which can also be checked from fragments) between groups 2 and 3, and two morphological types only: a plump one with rounded body (Fig. 5: 1, 2) and a tall slender one (Fig. 5: 4, 5). Probably as for the jars of Phase P, this corresponds to different products and conditions of conservation, but the repertoire of Phase N and the capacity groups show a much higher degree of standardization and specialization.

Jars with handles in group 1 (ca. 15/16 litres, Fig. 5: 7) are equally interesting, as they probably are the counterparts of the large jugs of phase P for carrying water. Jug types in the assemblage of Phase N are numerous (Thalmann 2000: figs. 44, 46b; 2002: fig. 8), but they are all of small size and belong to what may properly be termed “tableware”. On the contrary, the jars of group 1, with their moderate capacity and their shape (rounded body, “low” handles and tall restricted neck) meet all the requirements for easy carrying on the head or shoulder; the handles attached at center of gravity level are especially well suited for pouring when holding the vessel with both hands.

Only one large vessel with a capacity of 75 litres coud be reconstructed (Fig. 5: 6), but fragments of rims of a similar shape are numerous. The wide opening and the two strong handles do not match the usual types of contemporary storage jars, handleless and with a restricted neck, easy to seal with a clay stopper. This could well be also a specialized shape, new to Phase N, for storing domestic water: the wide opening allows for drawing up water with pots of all shapes and sizes, and the vessel can be held by the handles and tilted down for pouring when it is half or nearly empty.

Evolution of the repertoire in later periods

For later periods, the number of complete shapes available from Arqa is too low for significant capacity calculations. But new trends in the production and use of medium and large size containers are nevertheless apparent.

Very large storage vessels or pithoi (150 litres and more) appear only with Phase M (MB II) and different types of similar capacity are also produced during Phase L (early LB). Medium-sized jars are numerous and, much more than during the preceding phase N, some of them are of non-local origin.

Most of these jars, so far as can be inferred from fragments, probably fall within a capacity range between ca. 30 and 40 litres. Is this possibly a more or less standard capacity for many MB II jars, as noted above in the case of the Byblos jars? It should be necessary to accurately measure a wide sample from many Levantine sites to get the beginning of an answer. One of the few complete jars from Arqa, very similar in shape to the Byblos specimens and ascribed to the very beginning of phase M, holds 33 litres. On the other hand, it is probably not by chance that very few rims or large body fragments can be ascribed to the “intermediate” capacity group of 60/80 litres, which was well represented, although in somewhat different ways, in both pre-ceeding phases P and N.

It seems therefore that the specialized production of different types of storage vessels and the standardization so apparent in the Phase N repertoire was not continued during phase M. This may be related to the widespread circulation along the Levantine coast, especially from the beginning of MB II, of the true “commercial” jars, well adapted as noted above to the constraints of maritime trade, but which could also be re-used when empty for all kinds of local storage. At least for the manufacture of medium and large size containers, a trend in the de-specialization of local pottery manufacture, vs. the higher specialization of fewer workshops which produced the “ commercial” jars, probably began at Arqa during MB II with the wider availability of imported vessels; it becomes more visible in later periods, in all classes of pottery including tableware.