¹ George, A. R., Babylonian Topographical Texts (OLA 40; = Topog. Texts), pp. 13–29 Google Scholar; idem, “Babylon Revisited: Archaeology and philology in harness”, Antiquity 67 (1993), pp. 734–46.CrossRefGoogle Scholar

² In Sumerian é. s a . í 1, “House whose Top is High” ( George, , Topog. Texts, pp. 294–8Google Scholar; idem, House Most High, pp. 139 f.), pronounced Esangil with nasal // even in the first millennium, as is shown by the pseudo-etymological spellings of the E-sagil Commentary (e.g., ll. 15: [é. s] a. a n. g i. í l; 17: [é. s]a₆. an. gil) and the Aramaic transcription yysngl(see George, , Topog. Texts, p. 296 Google Scholar).

³ Ibid., No. 13.

⁴ Ibid., No. 14.

⁵ George, “86.11.12: Measurements of the Interior of the Temple E-sagil”, in Spar, I. et al, CTMMA II Google Scholar (forthcoming).

⁶ George, , Topog. Texts, pp. 9–11 and 43–55.Google Scholar

⁷ Ibid., No. 6.

⁸ Ibid., No. 7.

⁹ BM 119282, cited in Antiquity 67, p. 740 Google Scholar, as unpublished but now copied and edited by Pongratz-Leisten, B., Ina Šulmi Īrub (Mainz, 1994)Google Scholar, No. 6: pp. 218 ff., p. 285 and Pl. 1.

¹⁰ Read of course ṣal-mu, not er-mu, at the beginning of these lines (coll.).

¹¹ Weiher, Egbert von, Uruk: Spätbabylonische Texte aus dem Planquadrat U 18 (Deutsches Archäologisches Institut, Abteilung Baghdad, Ausgrabungen in Uruk-Warka, Endberichte, Band 12), Mainz am Rhein, 1993, pp. 133–4 and 225.Google Scholar

¹² Ibid., pp. 133–4. I am most grateful to Professor von Weiher for his kindness in supplying photographs of the tablet, W 22656/14, and to Professor R. M. Boehmer and the Deutsches Archäologisches Institut, Abteilung Baghdad, for granting me permission to publish collations from them. Where new readings have been obtained signs so collated are marked in the transliteration given below with an asterisk.

¹³ Note, however, that at places in the text measurements are given in cubits, i.e. iv′ 13 (for a possible reading see footnote 79); and v′ 14–15: […(*)]+14 ammat(kùš) šiddu(ús) / [… amma]t(kùš) pūtu(sag).

¹⁴ Von Weiher's kišib.lá, “ein (auf Vorrat angefertigtes?) Bauelement”, does not convince. He rejects the reading dub. l á, “was wohl tublu 'Fundamentgrube' wäre,” apparently in the belief that such a thing is unsuited to the context. However, the new evidence that the text presents makes it imperative that the nature of the dub. l á be re-examined to see whether it really is part of the foundations. This is done below.

¹⁵ Restoring [x bab bit] ˹d˺ 30; that this line gives the number of bricks across the space of a gateway is clear from the presence in the lines immediately before and after of its jambs (i′ 3 // 5: [x sip-p]e-e, “[x (bricks): the] sippu's”); cf. the discussion of sippu below.

¹⁶ For documentation see George, , Topog. Texts, p. 392.Google Scholar

¹⁷ Ibid., p. 96,9′ cf. p. 126,10–13. The “chapel of Bēltīya” (bīt Bēltīya) is the ancient term for that part of E-sagil which contained Bēltīya's cella, courtyard and other chambers. This complex was more than a “chapel”, of course, but less than a “temple”.

¹⁸ Ibid., p. 206, 4′.

¹⁹ Ibid., pp. 477 f., sub index entries “Court of Bēltīya” and “E-dara-anna”.

²⁰ See ibid., p. 279. I have restored bītu in the new text also — see footnote 15 — since bīt DN is the common usage in it, while šubat DN occurs only once, if at all.

²¹ Ibid., pp. 44 if.; p. 94,31.

²² BRMIV 25, 46 // SBH No. VII, 22′.

²³ Topog. Texts, p. 126, 9.

²⁴ Ibid., pp. 281 f., on Tintir II 2″, where it is listed as a šubtu.

²⁵ TCL VI 32 = Topog. Texts, No. 13.

²⁶ On the advice of M. Civil and E. Reiner, von Weiher read the fraction, passim, as BAR, i.e., “aḫû/aḫītu, 'äußere (Seite)'.” However, these figures do refer to numbers of bricks, as will be shown, and in bricklaying half-bricks are of course an unavoidable necessity wherever a corner is turned.

²⁷ The writing sip-pe-e expresses the plural, as also in NB royal inscriptions; contrast the situation at the gate of the cella of Bēltīya, which is flanked by double sippu's, each listed individually, and written sip-pi (§3′, edited below).

²⁸ George, , Topog. Texts, pp. 400 f.Google Scholar; Antiquity 67, pp. 738 ff.Google Scholar

²⁹ As we also know from the cultic compendium catalogued above as No. 7, a stone (marḫušu) statue of Marduk as Asarre was located in the chapel of Ninurta, and no doubt sat or stood on this platform.

³⁰ SpTUIV, p. 134.

³¹ In fact, as will be shown below, the damaged or illegible line can be reconstructed to supply exactly the figure by which the total appears to fall short.

³² See my article on MM A 86.11.12 in Spar, CTMMA II (forthcoming).

³³ ZA 72 (1982), p. 110 Google Scholar; RLA VII, p. 471.Google Scholar

³⁴ The figure of 32 cm square is given for the baked bricks (“Barnsteine”) of the temple's abutment wall (kisû) by Wetzel, F. and Weissbach, F. H., Das Hauptheiligtum des Marduk in Babylon (WVDOG 59), p. 4.Google Scholar The date of this wall is uncertain, since no stamped bricks were found in it, but it was certainly in place by the time of Neriglissar, whose inscription indicates that it already existed (I R 67, i 21: kisè-e bābāt(ká.ká) é.sag.íl). Though such a reinforcement may not have been originally planned for the building finished by Aššurbanipal, and it may have post-dated his reign, there is a letter of Aššurbanipal's superintendent of works, Urad-aḫḫēšu, which might anticipate the use of baked brickwork (if epertu is only used of bricks that are fired) in the precinct of E-sagil (ABL 119, rev. 12–15): re-˹eḫ-te˺ ^lú ú-ra-se e-pèr-[tú] ša tarbāṣē(tùt)^meš ša é.sag.[il] li-iš-hu-tu, “let the rest of the building labourers mould the (baked) bricks for the exterior courtyards of E-sagil.” The dictionaries are divided as to whether the Neo-Assyrian phrase eperta šaḫāṭu means to glaze baked bricks (CAD, s.v. šaḫātu A 4, following an idea of A. Salonen, last presented in Ziegeleien, pp. 67 ff.) or to smooth brickwork over with mud (AHw, s.v. šaḫātu IV, 4). Drawing attention to an archaeological problem, namely the lack of glazed bricks at Mari where the idiom libittam šaḫātum is attested nevertheless, M. Sauvage has proposed that this refers to the painting of bricks (NABU 1994/43). However, the common-sense approach of Postgate, J. N., JRAS 1974, pp. 52 f.Google Scholar, revealed twenty years ago that šaḫāṭu means simply “to mould” bricks. Such bricks could be treated to a subsequent process, as in Parpola, LAS I 283 = CT 53 106, rev. 6′-7′: ^lú.uru ak-kad-u-a e-pèr-t[ú] i-šáḫ-ḫu-] ˹ṭu˹ i-sa-ak-ki-[ru]. In this context the NA verb sakāru, “to heat (in an oven)”, refers to firing, and not, as proposed by Parpola, to gilding (LAS II, p. 278; also CAD S, s.v. sekēru B).

³⁵ Koldewey, R., Die Tempel von Babylon und Borsippa (WVDOG 15), p. 44.Google Scholar

³⁶ See Weissbach, , Hauptheiligtum, p. 39 Google Scholar, a. A different impression is given by Koldewey, , Tempel, p. 44 Google Scholar; also idem, Das wieder erstehende Babylon (5th edition, Munich 1990; = WEB ⁵), p. 204 Google Scholar, according to whom the bricks of Aššurbanipal's floor were a uniform 37 cm square.

³⁷ Weissbach, , Hauptheiligtum, p. 38 Google Scholar, a–b. According to Koldewey, op. cit., the dimension of the brick from below the middle floor of E-sagil was 40 cm square. This brick, BE 8084, specifically refers to the adornment of the pavements (tal-lak-ti) of E-sagil and Babylon with kiln-fired bricks (Borger, Esarh., §13). Among the bricks of Esarhaddon that are 30cm square is BE 39840 ( Koldewey, , WEB ⁵, p. 206 Google Scholar, Fig. 127), which according to Koldewey was found “in der Umgegend” (ibid., p. 205), sc. of the floors of E-sagil. These stray bricks of Esarhaddon may have been displaced from the lowest floor of the temple (level n). If we accept as fact Sennacherib's account of the temple's destruction and Esarhaddon's report of the radical extent of the rebuilding (ṣēr uššīšu maḫrûti… attadi temmenšu, “I laid its foundation platform directly on top of its ancient footings”: Borger, , Esarh., p. 21, 42–6Google Scholar) this floor is not older than the destruction of 689 B.C. and so dates to the beginning of Esarhaddon's work on the temple. In as far as the excavators reached it this floor was otherwise made of unstamped bricks, the dimensions of which are not revealed by Koldewey.

³⁸ The inscriptions of Esarhaddon and Aššurbanipal give the impression that Esarhaddon completed the structure of the temple and Aššurbanipal decorated it and fitted it out ready for the return of Marduk's statue ( Borger, , Esarh., pp. 21–4Google Scholar; Streck, , VAB VII Google Scholar, passim, note especially p. 230, 12–14: ši-pir é.sag.íl la qa-ta-a ú-šak-lil ina kaspi ḫurāṣi ni-siq-ti abnī ^meš é.sag.íl az-nun-ma ki-ma ši-ḫir bu-ru-mu ú-nam-mir é.umuš.a, “I completed the unfinished work on E-sagil: I decorated E-sagil with silver, gold and precious gemstones and I made E-umuša (Marduk's cella) sparkle like the 'writing of the firmament' (i.e., stars)”). Although the rebuilding of E-sagil is assumed to have begun quite early in the reign of Esarhaddon, eleven years after its destruction by Sennacherib, that is, in 678 B.C., it is now considered probable that the work did not make much progress until the conquest of Egypt in 671 (see Parpola, S. in Alster, B. (ed.), Death in Mesopotamia (Mesopotamia 8 = CRRA 26), pp. 179 f.Google Scholar; Frame, G., Babylonia 689–627 B.C., pp. 68, 77 f.Google Scholar). Even so, most, if not all, of the basic work must have been completed by the time that the cult-statues eventually returned to Babylon, at the accession of Šamaš-šuma-ukīn in 668 B.C., although some furnishings, notably Marduk's bed and chariot, were not installed until much later (654 and 653 B.C. respectively). Though six months elapsed between the death of Esarhaddon and Šamaš-šuma-ukīn's arrival in Babylon with the cult-statue of Marduk, it remains unlikely that the walls of the central courtyard and other structural parts of the main building had yet to be built at the time of Aššurbanipal's accession. What is probable, however, is that some, if not all, of the secondary brickwork known to have been the work of Aššurbanipal, rather than his father — the raising and repaving of the floors, and maybe the addition of the kisû on the exterior walls — dated to this time. The raising of the floor, by nearly half a metre, was very likely occasioned by damp rising from the water table, which suggests that Esarhaddon's architects failed to make high enough the mud-brick platform (temmennu = level p on Andrae's section) on which they built the temple. (Nebuchadnezzar's later raising of the floor by over a metre must have resulted from the same problem, a perennial difficulty at Babylon.) As is well known, Marduk's return to E-sagil was anticipated by Esarhaddon's inscription which purports to record the event, but in reality it had to be postponed, a change of plan that was put down to bad omens (see Lambert, W. G., “Esarhaddon's Attempt to Return Marduk”, AOAT 220 = Fs Deller, pp. 157–74Google Scholar; Frame, op. cit., pp. 77 f.). An inadequately waterproof floor would have been a sound practical reason for postponement of the ceremony, and might have been exactly what encouraged the king's experts to seek an excuse for the change of schedule in their divination.

³⁹ As I am advised by Professor David Oates, with whom I discussed privately the technology of building with mud brick. On this subject see further Salonen, A., Die Ziegeleien im Alten Mesopotamien (AASF B 171)Google Scholar; Gasche, H., “Lehm als Baumaterial”, RLA VI, pp. 550–6Google Scholar; Oates, D., “Innovations in Mud-Brick: Decorative and structural techniques in ancient Mesopotamia”, World Archaeology 21 (1990), pp. 388–406 CrossRefGoogle Scholar, with bibliography; Moorey, P. R. S., Ancient Mesopotamian Materials and Industries (Oxford, 1994), pp. 302 ff.Google Scholar, with references there cited.

⁴⁰ Wetzel, , Hauptheiligtum, p. 7.Google Scholar

⁴¹ Ibid., p. 7 and Pl. 4b. Wetzel remarks: “der Unterschied zwischen der Summe der Einzelmaße und der Durchmaße erklärt sich aus den schon erwähnten Schwierigkeiten der Messung” (p. 7), sc. “aus den kleinen Unregelmäßigkeiten der einzelnen Einheiten, von denen meist nur wenige Schichten freigelegt wurden, und vor allem aus der großen Behinderung der Maßaufnahme in den engen Stollen, die oft eine genaue Bestimmung sehr erschwerten” (p. 5). Even though the north wall of the courtyard was laid bare almost in its entirety, with only the outer jambs of Gates j and 1 inside tunnels, the implication is still that the aggregate measurement of the individual sections is likely to be less accurate than the overall measurement.

⁴² In traditional building with mud brick a layer of rough mud mortar, typically about 2 cm thick, separates each course (according to the modern evidence collected by Salonen, , Ziegeleien, pp. 47 ff.Google Scholar). The question of how much mortar separated the individual bricks of a course in a Neo-Babylonian mud-brick building is less well documented. However, it may not be irrelevant to quote evidence gleaned from measurements taken of the baked brickwork of E-sagil. In the abutment wall, which was constructed out of standard 32 cm bricks, mortar (Wetzel's “Lehmzwischenlagen”) could, by my calculations, add between 0.67 and 2.41 cm to the length of a baked brick. These figures are computed from the dimensions given for this brickwork in Wetzel, and Weissbach, , Hauptheiligtum, p. 5 Google Scholar, where some sections of the temple's façade are given in both metres and brick-lengths. The width of the westernmost section of “Kurtine” on the north front of the temple is given at 5.85 m (17 bricks), which produces a length of 34.41 cm per brick inclusive of mortar. The lowest rate of centimetres per brick inclusive of mortar yielded by this set of figures is 32.67.

⁴³ The copy reads so, against von Weiher's transliteration, which has only BAR.

⁴⁴ The unpublished photograph confirms these signs, which remain obscure.

⁴⁵ See the survey of Muayad Said Damerji, Basim, The Development of the Architecture of Doors and Gates in Ancient Mesopotamia, transi. Takase, Tornio and Okada, Yasuyoshi (Tokyo, 1987), pp. 68–70.Google Scholar Though the stepped brickwork is ornamental rather than functional I have retained the term “rabbeting” as the most convenient.

⁴⁶ See Wetzel and Weissbach, Hauptheiligtum, Pl. 4b, where the individual figures for the three steps of rabbeting that adorn the west jamb of Gate k are 40, 32 and 32cm = 1.04 m. The combined measurements of the rabbeting of the jambs of all three gates can be discovered by subtracting the widths of the individual gateways from the figures given under “Kurtine” in the table on ibid., p. 7. Using these figures the two jambs of Gate j together occupy a width of 2.22 m, i.e., 1.11m each if the construction was absolutely symmetrical; those of Gate k measure 1.17 m each, or, if the figures given in Wetzel's plan really belong here, an asymmetrical 1.04 m and 1.30 m; those of Gate 1 work out at 1.045 m each. Since this last figure is almost exactly the dimension recorded in Wetzel's plan for Gate k, one wonders whether the individual widths of the three steps of rabbeting given there are misplaced, and really belong to Gate 1. Comparison with other sections of the text shows that the rabbeting of the gateways on a given stretch of wall was considered to be uniform in width. The excavated remains reveal that the builders did not achieve quite such a consistency of measurement. On this wall the text assumes the maximum measurement achieved at the rabbeting to be universal; excavation shows that in practice it was not.

⁴⁷ Muayad, Said, Doors and Gates, pp. 71 ff.Google Scholar

⁴⁸ Reference works in English traditionally favour the term “buttress” (e.g. Frankfort, H., The Art and Architecture of the Ancient Orient, p. 18 Google Scholar; Leick, G., A Dictionary of Ancient Near Eastern Architecture, p. 39 Google Scholar). Neither “buttress” nor even “pilaster” seems quite the right word when the projection is simply a flat surface proud of the face of the wall, not load-bearing like a buttress or ornamented like a pilaster. According to Harris, Cyril M. (ed.), Illustrated Dictionary of Historic Architecture (New York, 1977), p. 420 Google Scholar, the correct term for such a projection should be pilaster mass: “an engaged pier built up with the wall, usually without the capital and base of a pilaster.” A less wide structure is known as a pilaster strip or lesene. The thickened section of wall that is produced by the use of a pilaster mass or strip may properly be called a pier: “a member, usually in the form of a thickened section, which forms an integral part of a wall; usually placed at intervals along the wall to provide lateral support or to take concentrated vertical loads” (ibid., p. 417).

⁴⁹ The projecting features are described by Koldewey, as “schwach vortretende Pfeiler-Türme” (Tempel, p. 42 Google Scholar). Wetzel termed niche and projection respectively “Kurtine” and “Turm” (Hauptheiligtum, pp. 5 and 7).

⁵⁰ Hauptheiligtum, p. 7: “die Hofwände sind durch einfache Türme, die je l Stein vorspringen, gegliedert.” This Statement holds for the north, east and south walls of the courtyard, but according to the plans the outward projection of the pilaster masses of the west wall, which fronted the cella of Marduk, was much greater.

⁵¹ In the following I render dublu as “pilaster” instead of “pilaster mass” or “strip” for the reason that the ancients are unlikely to have distinguished between decorated, wide and narrow: if one such feature was a dublu, probably they all would have been.

⁵² With dublu āṣû in this meaning cf. Nebuchadnezzar II's di-ma-a-tim a-ṣa-a-tim, “projecting towers” — i.e., buttresses? — built on the moat wall of Babylon (CT 37 12, 27; not a-sà-a-tim, “turrets”, as erroneously read in Topog. Texts, p. 346). It may be that in § 1′ a [dublu] ṣīru, “massive/huge/tall pilaster”, should be restored (i′ 2). This would be a more substantial pilaster mass that perhaps protruded above the level of the roof, like a tower.

⁵³ For the jambs see above, footnote 46. In the text the four pilasters that relieve the face of the wall, measuring between 2.62 and 2.71 m, would all have been dublu's of a uniform width, i.e., 7 brick-lengths. The two recessed faces, at 3.34 and 3.38 m, are an average 68 cm or 2 standard brick-lengths wider than the average pilaster. So the ḫibšu's of this wall were both 9 bricks wide.

⁵⁴ This figure is not only suggested by the emendation of 4 to 7 in iii′ 3 but actually demanded by the need to maintain regular widths sippu, dublu and ḫibšu. Since the combined width of the sippu's must be 21 bricks, the total width of the dublu's 28 bricks, and the gateways, after emendation, 19 bricks, which all add up to 68 bricks, the tablet's total of 83 can only stand if the two hibšu's measure 7½ bricks each. However, as seen in the preceding footnote, Wetzel's figures show that the recesses were of 9 bricks' length, and thus the total is also best raised by three to 86.

⁵⁵ To the references for city gates add Topog. Texts, p. 140, 3.6. The translation “(door)-sill” used in that book follows Salonen's “Schwelle” (cf. Landsberger's rendering “threshold” in early MSL and CAD; also Hebrew sap), but is not entirely satisfactory, for the sill is only the bottom of the jamb, not the whole thing.

⁵⁶ In Tintir II 29′–30′ the “stations” (manzāzu, i.e., socles for statues) of Ababa and Antadurunnu, the twin gatekeepers of E-sagil, are situated at a pair of sippu's in the temple.

⁵⁷ In ancient Mesopotamia wooden doorframes comprised various constituent parts, including the ḫittu (Sum. ^gišḫé . du₇), “lintel”, the giškanakku (g i š . k á . na), literally the “wooden part of the gate”, and the gištallu (giš . dal), “wooden cross member”. The equation giš. ˹ká˺ . na = sí-ip-pu-um in proto-Kagal ( MSL XIII, p. 88 Google Scholar, 68) need not be taken as an exact equation, but may only mean that the giškanakku was a wooden item that was fixed to the sippu and could fulfil the same function as the upright edge of the wall, i.e. it was the vertical member of the doorframe.

⁵⁸ “Palace” is certainly out of place. Corruption of original dub. lá = dublu?

⁵⁹ Tempel, p. 43; cf. WEB ⁵, p. 205.

⁶⁰ Or NS 54 (1985), pp. 189–202.Google Scholar

⁶¹ “Exit Talim! Studies in Babylonian Demonology, I”, JEOL 27 (1981–1982), pp. 90–105 Google Scholar; idem, Mesopotamian Protective Spirits. The Ritual Texts (Groningen, 1992), pp. 164–6.Google Scholar

⁶² Lambert, , Or NS 54, pp. 192, 194–6Google Scholar; Wiggermann, , JEOL 27, p. 95 ²⁷ Google Scholar, cites the earlier literature, to which add Cooper, Jerrold S., The Curse of Agade, p. 248.Google Scholar

⁶³ Lambert, , Or NS 54, p. 191 Google Scholar; Wiggermann, , JEOL 27, pp. 101 ff.Google Scholar

⁶⁴ E.g. Falkenstein, Å., Heimpel, W., Nissen, H., as cited by Wiggermann, , JEOL 27, p. 95 ²⁷ Google Scholar; also Soden, W. von, CRRA 20, p. 141 Google Scholar; D. O. Edzard, ibid., p. 156; and now too Jerrold S. Cooper, op. cit.

⁶⁵ Specifically entries in which du-ub-lu (An IX 44) and its variant du-bur (LTBA II 2, 327) = išdu, “foundation”, and d u b . l á = dur-zv-um, i.e. dur(uš)šu, “foundation” ( Pettinato, G., MEE IV, p. 324, 1162 Google Scholar; Ebla). The latter would seem to confirm the previously questionable identity of d u b . l á and dublu. Note also ^du-burdubur, dúbur = iš-[du], Ea V 104–5, = iš-d[u], A V/2 126-7. Lambert proposes that d u b. l á is a loan from dublu (Or NS 54, p. 193). It could also be argued that the Sumerian word was properly dubur, which was taken over in Semitic as dublu and borrowed back again as d u b . lá (such a development is not as unlikely as it might at first appear: cf. the reversely analogous sequence of loans šāpiru > š a b r a > šabrû and bēlūtu > b i ll u d a > pilludû). On d ú b u r see recently Alster, B., RA 85 1991), pp. 9 f.Google Scholar

⁶⁶ E.g. Sjöberg, Å. W., as cited by Wiggermann, , following CAD D, p. 168 Google Scholar; W. von Soden, AHw, s.v. tublu, “etwa ‘Fundamentgrube’,” followed by Weiher, von, SpTU IV, p. 134.Google Scholar Note also Jacobsen, T., The Harps that Once…, p. 419 Google Scholar, “socle” (translating d u b . lá in Gudea, , Cyl. A, xxiv 18 and 26 Google Scholar).

⁶⁷ Or NS 54, pp. 193 f.Google Scholar

⁶⁸ George, , House Most High, p. 79, 203–5Google Scholar, where the name is translated “House, Exalted Door-Socket”, following Lambert.

⁶⁹ As noted by Gadd, C. J., UET I, p. 22 ¹² Google Scholar, the ceremonial name also appears in an Ur III offering list as the recipient of a sheep ( Langdon, S., RA 19 (1922), p. 192 Google Scholar, No. 4, rev. 6; from Adab or Umma?), and in an OB adoption contract, where a priest of dub.lá.maḫ acts as witness (BIN II 75, 34; from Larsa?). Cf. Falkenstem, , AnOr 30, p. 124 ³.Google Scholar

⁷⁰ See the photographs in, e.g., Woolley, Leonard, UE VIII, pp. 4 ff.Google Scholar; Excavations at Ur, Pl. 29a.

⁷¹ As discussed in Topog. Texts, pp. 416 f. My conclusion there, that “no recess in the exterior walls of E-sagil occupies a space large enough to be identified as the ‘suḫātu adjacent to the arkabinnu-doof’,” left unresolved the location of the recess, the arkabinnu-door and the adjoining courtyard, which is called the Court of Ištar and Zababa. The new information regarding the nature of the suḫātu led me to look again, and I discovered that the size of the suḫātu given there, 2½ mušaru “in the large cubit standard”, fits the recess south of Gate H if in the text “large cubit standard” is an error for the ordinary cubit standard. 2½ mušaru in the smaller, Neo-Babylonian standard represents an area of about 90 m². The area of the recess south of Gate H is 90·1 m² if measurements are taken from the abutment walls (kisû), or about 95.2 m² if one ignores them. In this analysis the arkabinnu-door of the E-sagil Tablet will have been hung in Gate H, and will indeed be the same as the arkapinnu-Gate listed as one of the principal gates of the temple in the list catalogued above as text No. 6 (Topog. Texts, p. 96, 6′). The Court of Ištar and Zababa will be the area outside this gate, to the south of the main building and west of the protruding façade of the eastern annexe. See Fig. 1.

⁷² Von Weiher read k á, but in both copy and photograph the sign is different from the one in the adjacent line. Multiple jambs of equal size are elsewhere grouped together, and followed by the notation malmališ. Here the individual jambs are described differently, and are thus kept separate. In a jamb with two steps of rabbeting, as here, the “jamb of the interior” (or, reading sip-pi qablî, “the middle jamb”) is clearly the step sandwiched between the flat face of the wall on the one side, and the jamb at the very edge of the gateway on the other.

⁷³ Clearly not SI]G₄ (SO von Weiher).

⁷⁴ Combining the figures given in Wetzel, , Hauptheiligtum, p. 7 Google Scholar, for the three central stretches of “Kurtine”.

⁷⁵ The sign is k i s a 1, definitely not ḪÚL (as read by von Weiher).

⁷⁶ The “cella of Nabû on the courtyard” is the phrase used by Nebuchadnezzar II to describe é . z i. d a in E-sagil ( Weissbach, , WVDOG 5 Google Scholar, Wadi Brisa B ii 2: pa-pa-ḫi ^dnabû(nà) ša ki-sa-al-lum; CT 37 7, 33: pa-pa-ḫi ^dna-bi-um ša ki-sa-al-lam). The restoration of the location of the wall as a heading must be provisional, however, since the same information is given at the end of the section. Note that there is room at the end of § 6′ for that section, too, to have been equipped with a full description of the wall in both heading and subscript.

⁷⁷ The text appears to be in slight disorder here as a result of the scribe's decision to squash ll. 37–40 on to the lower edge of the tablet. In so doing he has placed ll. 39–40 in the vacant space he left after the ruling at the bottom of col. ii′, and has allowed 1. 39 to run over into the gap between ll. 37 and 38 at the bottom of col. iii′ proper.

⁷⁸ Reading with the copy and photograph; von Weiher's transliteration: 32.

⁷⁹ From the unpublished photograph I read ˹i-gar˺ šu-bat, “the wall of the shrine”, at the beginning of the line and ˹9⅓˺ ammat(kùš) pūtu(sag), “9⅓ cubits wide”, at the end. What is missing can hardly be other than a divine name, since šubat DN is so common in E-sagil. Since this part of the text deals with the walls of the cella of Nabû and associated chambers, it would seem likely that this šubtu is in fact the cella of Nabû, é. z i. d a. This sanctuary is termed a šubtu in Tintir II 2″, where its occupant, presumably Nabû, bears the name Lugaldimmerankia. That name is much too long to be restored in the middle of the line, of course, and one thinks instead of ˹^d nabû(nà˺, but this is not confirmed by the photograph and needs collation. Since the walls of the two immediately preceding sections, §§ 7′-8′, are the long walls of a Breitraum cella and the wall opposite, it would not be inappropriate to include the width of the cult-chamber also.

⁸⁰ Von Weiher reads: 1-at-ma ^? pa ^?-paḫ ḫa GIŠ(+)UD.

⁸¹ Some jambs of interior doorways are rabbeted, though with a single step only, notably those at the back of Rooms 12 and 15. These chambers were both more than usually sacred.

⁸² For the architecture and lay-out of E-zida see now J. E. Reade's plan, after Rassam, and Koldewey, , in Iraq 48 (1986), p. 107.Google Scholar

⁸³ For the western side, which fronted Marduk's cella, see above, the discussion on §3′. The eastern side shows a variation on the articulation used in the north and south walls, in that the stretches of walls between the gateways are decorated with three pilaster masses rather than two.

⁸⁴ See footnote 20.

⁸⁵ It is not easy at first to reconcile the heading of § 6′, “the wall of Ka-Lamma-rabi inside the courtyard of Bēltīya”, with the known fact that Ka-Lamma-rabi is Gate D, the north gate of the main building, not an interior gateway. What the heading must mean is that the wall in question was an interior wall of the courtyard, and it was in some way connected with Ka-Lamma-rabi. In this analysis the gateway at one end of the wall, ká ^dl a m m a (iii′ 17), will be the gate that gave access from the lobby of Ka-Lamma-rabi to the courtyard of Bēltīya. Whether the name represents an abbreviation of k á. ^dl a m m a . r a. b i or is a hitherto unknown bāb lamassi, it is appropriate to a gate in such a position. The implication is that this end of the wall ran into the corner of the courtyard that was nearest Ka-Lamma-rabi.

⁸⁶ See Topog. Texts, p. 126, 8–11. In my attempt to reconcile the measurements given in the metrological text from Aššur with the known ground-plan of E-sagil I suggested exchanging ll. 8 and 9 of this text, so that the chapel of Ea fell south of the Court of Bēl (ibid., p. 121). The new text demonstrates that this solution was wrong, and this problem, like others posed by the metrological text, remains unsolved. It should also be noted that one of the šubtu's of Ea listed in Tintir II, namely é . ḫ a 1. a n . k i (l. 20), occurs immediately after the two šubtu's of the chapel of Ninurta and another in the “courtyard of the chapel of [Ninurta or Bēltīya]” (ll. 17–19). There are traces of a topographical order in the list (cf. ibid., p. 10), and so one might expect Ea's é. ḫ a l . a n . k i to have been located somewhere near Ninur-ta's chapel. Very probably it was in fact inside the chapel of Ea, though the cultic focus of this chamber would have been another statue of Marduk, this time of taskarinnu-wood (BM 119282, obv. 1b; cf. Pongratz-Leisten, B., Ina Šulmi Īrub, p. 218 Google Scholar, 3).

⁸⁷ It will be noted that what we are reconstructing as the opposing pairs of sides of a courtyard are not of exactly the same lengths, but one or two bricks' length out. This kind of irregularity appears to have happened in reality: the west side of the central courtyard of the temple, at 37.37 m, was 33 cm, i.e., one brick's length, shorter than the east, at 37.70 m: Wetzel, , Hauptheiligtum, p. 7.Google Scholar

⁸⁸ The width of this space was just over 15 m: Room 5 is 9.95 m long, Room 11 3.0 m wide ( Wetzel, , Hauptheiligtum, p. 6 Google Scholar and Pl. 4b). The width of the wall that separates them, one of the narrowest excavated in the temple, I read from the plan as 2.15 m. In bricks of E-sagil 46 bricks represents between 13.8 m (using bricks of 30 cm) and 17.02 m (using bricks of 37 cm), but the standard brick of ⅔ cubit, i.e., 32 cm, yields the figure 14.72 m, not including mortar.

⁸⁹ For E-zida see Reade's, plan, Iraq 48, p. 107.Google Scholar Hormuzd Rassam's plan of E-babbarra is presented in modernized form by L. De Meyer (ed.), Tell ed-Dēr III, plan 3, and George, , Topog. Texts, p. 220.Google Scholar The courtyard and rooms north-west of Samaš's cella have been further explored by the University of Baghdad, whose plan of this part of the temple is reproduced by al-Jadir, W. in Meyer, L. De and Gasche, H. (eds.), Mésopotamie et Elam = CRRA 36, p. 195.Google Scholar

⁹⁰ Topog. Texts, p. 126, 11.

⁹¹ Ibid., p. 94, 34, describes a dais of the goddess Nissaba, “on the north wall, behind Bēltīya”. Elsewhere “behind” a deity means at the rear of the cult-statue, by implication behind the cella that housed that statue (e.g., ibid., p. 96, 10′, where the west gate of E-sagil, Ka-ḫegal, is described as the gate “behind Bēl”, i.e., at the rear of his cult-rooms). Whether Nissaba's dais was inside the temple, in a chamber beyond Bēltīya's cella, or outside, against the temple's exterior façade, this description would seem to signify that Bēltīya's cult-statue stood in front of the temple's north wall. In a Breitraum cella the statue rests on a dais against a niche in the back wall. If north lay to Bēltīya's rear, since the back wall must be one of the cella's long sides, this wall would have to be parallel with the temple's north front, and thus would have lain east–west, on an axis perpendicular to that of Marduk's cult-rooms.

⁹² Thus, at E-zida, the secondary cellae (of Tašmētum and Nanāy, or of Tašmētum and Mār-bīti?) are situated either side of Nabû's cult chambers, each off its own courtyard, but with short sides and long sides along the same axes as those of Nabû's cella. At Sippar in the great temple of Šamaš, the arrangement is similar: Samaš's cella lies in the middle, between the central courtyard and the temple's south-west façade. To one side, on the north-west, is a second set of cult-rooms reached from its own courtyard and probably dedicated to Aya, Šamaš's wife. On the other side, less well explored, there was a similar arrangement, probably for Šamaš's vizier, Bunene. All the cellae and their antechambers are aligned on the same axes.

⁹³ Published in TCL VI and Dijk, J. van and Mayer, W. R., Texte aus dem Rēš-Heiligtum in Uruk-Warka. Bagh. Mitt., Beiheft 2 (Berlin, 1980).Google Scholar See further McEwan, G. J. P., BiOr 38 (1981), 639 Google Scholar, who identifies the two men.

⁹⁴ Hunger, H., SpTU I 94 Google Scholar, 56.

⁹⁵ See ibid., p. 13.

⁹⁶ See Beaulieu, P.-A., “Antiquarian theology in Seleucid Uruk”, Acta Sum 14 (1992), pp. 47–75.Google Scholar

⁹⁷ On the Rēš Temple in general see Falkenstein, A., Topographie von Uruk, pp. 4–26.Google Scholar

⁹⁸ This, too, was a feature of other temples, which everywhere very likely represents a conscious imitation of the arrangement in Enlil's E-kur at Nippur. There was certainly a Dais of Destinies at Uruk long before the foundation of the Rēš Temple (as listed in KAR 142, ii 13), no doubt housed in an Ubšu-ukkinna in E-anna. See Falkenstein, op. cit., p. 82.

⁹⁹ BRMIV 7, 24: bīt(é) á-ki-tum bīt(é) ik-rib.

¹⁰⁰ On the seven stages of the akītu procession at Babylon and Uruk see Pongratz-Leisten, B., Ina Šulmi Īrub (Mainz, 1994), pp. 40–3.Google Scholar It may be noted that the five incantations chanted by the exorcists as Anu takes up residence in the akītu house, as given in BRM IV 7 28–31, include the three known to have been used when refurbished divine statues were taken back to their temples, as prescribed in the mīs pî rituals of Babylon (Sidney Smith, , JRAS 1925, pp. 40 Google Scholar f. = Ebeling, E., TuL, p. 107, 60–1Google Scholar). However, this fact is not necessarily evidence for borrowing, for these incantations may have served this general purpose in all Babylonian temples.

¹⁰¹ See, e.g., Lambert, W. G., Iraq 45 (1983), p. 86 CrossRefGoogle Scholar; George, A. R., Iraq 48 (1986), pp. 143 f.CrossRefGoogle Scholar; BSOAS 52 (1989), p. 119 Google Scholar; Pongratz-Leisten, , Ina Sulmi Īrub, p. 63.Google Scholar

¹⁰² To judge by the colophon of AO 6451 (RAcc, p. 80, 46–9), a tablet prescribing the meals offered to “the gods resident in the Rēš Temple, the Ešgal and E-šarra, the exalted dais, the ziqqurrat of Anu”, later tradition had it that the daily rites, at least, of Anu's temple were preserved on old ritual tablets that had been looted from Uruk by Nabopolassar and fortuitously discovered and copied in Elam by the chief priest of the Rēš Temple in the reign of Seleucus and Antiochus. The anachronism of this claim provokes suspicion that the priests of Anu legitimized his reformed cult in the normal way, by resort to fiction, in this case the production of what were passed off as copies of ancient — and conveniently inaccessible — documents. See Falkenstein, , Topographie, pp. 8 f.Google Scholar

¹⁰³ I.e., dublū? Von Weiher read K]IŠIB ma? lu-ú