The treatment of the walls is similar to that of Emach in an intensified form. Here every tower is placed between two flanking towers, thus forming a unit of three towers. This also occurs in the great temple of Nebo in Borsippa. Exactly in the middle of each side there is a great gateway elaborated with massive projecting towers. Paved ramps, with side balustrades, lead up to the three gateways on the north, west, and south. All is on a larger scale than in other temples. The symmetrical planning which in other temples leaves much to be desired, is here remarkably accurate, and here alone is an entrance to be found on each side. [Illustration: FIG. 127.—Esagila and Babylon brick of Esarhaddon. ] Although from the outside these gateways all appear to be alike, the east gate must have been the principal entrance, as it has a passage through a magnificent vestibule that leads direct to the court, while the entrance through the north and south doors leads first into a small vestibule and then through a corridor that runs by the side of it. On the walls of the court also doorways and towers are symmetrically alternated. [Illustration: FIG. 128.—Terra-cotta figure from brick casket at Esagila. ] A considerable similarity exists between our temple and the description of the “temples” that lay near the zikurrat given in Smith’s summary of an inscription (see p. 192 _et seq._). Smith was not then aware of the difference between Esagila and the Bel sanctuary of Herodotus. His “temples” have measurements and proportions which, on account of their disproportionate length, are entirely impossible as such. For enclosed chambers they are far too widely spanned. They can therefore only be measurements of the area of separate parts of the temples, including the adjacent walls. If all of these are added together we find that they amount almost exactly to the occupied area of Esagila. Furthermore, these areas can with ease be arranged so as to fill in the ground plan of Esagila with very few discrepancies. Then again the principal cellae are here, that of Marduk and of Zarpanit in the west, and that of Ea in the north, while that of Anu and Bel may at least find its counterpart in the double cella in the south of Esagila. Thus the measurements of Smith’s “temples” must have been taken either directly from Esagila or perhaps from the temple on the top of the zikurrat, which must then have had much the same dimensions and arrangements as Esagila. It is to be expected that the further excavations of Esagila will throw light on these most interesting questions.