North-east of the Ninib temple we have cut four trenches through the hill to the plain beyond. Here we found the same strata of private houses and streets that we shall meet with again in Merkes.
Here, at the depth of the water-level, were some small plano-convex clay tablets with carefully modelled reliefs of lions, fabulous creatures, etc., on the flat side, as well as some figures in the round, also worked with great minuteness. Among these there is a fine bearded head with the hair tied up in a napkin, as, beside others, it is worn by Marduk on the piece of lapis lazuli described above. They appear to be working models for a large statue.
Beside the numerous scantily ornamented pottery vases, there were some decorated in coloured glazes with concentric lines, rosettes, and plaited bands (Fig. 152—Frontispiece). They come from the lower levels, which apparently date back to the time of the Assyrian domination. In one place where rubbish had been thrown, there were numerous tablets containing business, literary, or scientific inscriptions. It is possible that they came from the temple and formed part of the temple library, which, as is generally supposed, every temple possessed. No systematic storing of inscriptions has yet been discovered in any temple, including those of Babylon, Khorsabad, and Assur, all of which have been completely excavated. It is true that these were buried under a proportionately shallow covering of earth, while Esagila lay protected under fully 20 metres of untouched accumulations, and is still unexcavated.
The mound itself proves to be thickly strewn throughout with potsherds, and the mud-brick walls of the houses lie close below the surface. They are only thinly covered by a uniform layer of dust. In the plain, on the contrary, as our trenches at the Ninib temple have shown, the house ruins lie under a layer more or less high, of drifted sand, and the surface contains exceedingly few potsherds. All this is explained if we take the trouble to realise the antecedents of the formation of these ruins. At the time when the site was deserted and fell into ruins the surrounding contours were far more marked than they are at present. The heights were higher and the depths were deeper. The mud-brick walls, which at first stood out above the soil, crumbled away after they lost their roofs into dusty heaps of clay, which accumulated against the walls and covered the pavement higher and higher, while the walls themselves, so far as they over-topped these heaps, disappeared, and thus all was levelled to an irregular undulating surface.
Fig. 153.—Schematic diagram of the transfer of the upper levels (A, B, left) of a mound of debris to lower-lying region (A, B, on the right).
But the process of destruction of the city did not end here. Every winter, however short, with its frost and rain, and the long summer with the torrid heat of the sun, split, shattered, and pulverised all that still clung together and turned it to a light powdery dust, which was easily whirled away by the strong recurrent summer winds and deposited in the lower-lying parts. Thus the heights were continually denuded and lowered and the depths were gradually raised (Fig. 153). The heavier objects, such as pieces of burnt brick and fragments of pots and sarcophagi, were thus sifted as it were and left exposed on the surface, and the higher the mound had been in which they lay scattered, the closer they would now lie together. Thus on the surface of ancient mounds that were not inhabited later we find small objects in very large numbers. Clay coffins, which at the time of burial were laid deep in the ground, are now on the surface, and as the process continues they form a small heap of sherds. A specially striking example is the appearance of the wells and sunk shafts, which consist of pottery rings placed one above another. Originally, of course, they all ended at the level of the pavement of the buildings to which they belonged. When these fell to pieces and were blown away and disappeared with a large part of the earth on which they stood, the lower part of the well which was in the ground was covered over with a small heap of fragments from the broken upper part, which stood out above the surrounding ground as an exposed drum (Fig. 154).
Fig. 154.—Schematic diagram of section through Babylonian house ruins, with wells.
The longer the ruin as such had remained fallow the more marked are the traces of this abrasion of the fallen material and the emergence of the harder objects. In Merkes and in Ishin aswad we can, on the whole, scarcely count on more than one wind-swept stratum of habitations. At Fara (Shuruppak) there were more of them, and at Surgul and El-Hibbah there were many. Every new inhabited stratum, so long as the mounds rose, joined on new wells to the old ones as the latter disappeared from sight, while on every denuded dwelling site the well appeared on the surface together with those of the preceding layer. This is the reason why the well rings visible on very ancient ruins, such as Surgul and El-Hibbah, are so exceedingly numerous, a fact which is unintelligible to those who do not understand their origin. Many erroneous explanations have been given, among others that they were drains intended to keep the hill dry, whereas they had absolutely nothing to do with that purpose.