In planning a Babylonian private house a square principal chamber on the south side of a court appears under all circumstances to have been indispensable. Everything else might vary according to circumstances and temporary requirements; the side-chambers might be more or less numerous, several courts with the chambers connected with them might be added to the house, but the court and the principal chamber are always there. Before the introduction of Greek art there were no pillars either in the court or in the house.

The largest house (Fig. 236) that we have yet found in Merkes possesses three courts (4, 19, 26), each with its principal chamber on the south (12, 23, 27), which corresponds in size with the court to which it is attached. The wide doorway of the house on the north is in a flat length of wall which has no toothed projections, such as all the other walls have. Through this we enter the vestibule (1), and can turn either left to the main portion with the large court, or right to the private or secondary portion with two courts. The former part of the house was certainly consecrated to business and to intercourse with the general public. This is indicated by the fact that in this part only there was a second outer door on the south side, which later was walled up. This opened on a small room (13) that communicated immediately with the principal chamber, and may have served as a shop. In any case, the owner could here communicate with the outside world without being obliged to use the ceremonious northern entrance. On entering by the latter, one passed a very small room (2), the entrance chamber and porter’s lodge, the cloak- or waiting-room (3) before reaching the court (4). To the east of this lay the servants’ apartment (5), and to the south the stately principal chamber, about 14 by 7 metres in size; with a smaller series of four chambers to the right (17, 14, 15, 16) and a larger one of six chambers (6–11) to the left of it. Both these series of rooms communicated with the principal chamber by a corridor (14, 8) and with the court by their most northerly chamber (17, 6), which was perhaps a merchant’s office. The inner rooms (15, 16, 10, 11) must have been perfectly dark unless they were lighted by windows on the street, which is very improbable. In one of them (15) there was a well, constructed as usual of pottery cylinders. They may have been store-rooms or sleeping- and living-rooms for the people employed there. It is scarcely necessary to warn our readers that all these suggestions as to the purpose of the various rooms rest entirely on supposition. We have no other authority for them than the arrangement of the ground-plan appears to afford.

The secondary group of chambers was reserved apparently for the private life of the owner. The rooms are grouped round two smaller courts (19 and 26) which communicated with the principal chamber of the northern one (23), and with each other by means of a corridor (25). From this corridor a door on the west led to an adjoining house, which had been built previously, and of which, on the whole, the great house represented an extension. The entrance chamber (18) and the two principal chambers (23, 27) are also easily recognised. It is not necessary at present to hazard conjectures as to the purpose of the other rooms.

The original pavement of the house has twice undergone restoration (Fig. 237). Between the layers of brick, most of which bear Nebuchadnezzar stamps, only a little earth is laid. No one was buried in the house while it was occupied; the 21 graves that occur on the site are all of the period when the building lay in ruins. This is shown by the way in which the walls and pavement were cut through, and by the fact that the pavements were not repaired in any way after the burials had taken place. The graves are chiefly of brick, as they are exclusively of the Parthian period. It is quite possible that the house was built during the reign of Nebuchadnezzar; no difficulty is involved by the occurrence of the bricks bearing his stamp, as it does not force us to infer any complete destruction of one of Nebuchadnezzar’s buildings. The bricks may very well be older material offered for sale by the king on the occasion of one of his rebuildings. It is impossible to say how late into Persian or Greek times the house existed; a poorer house was built on its ruins after the heap of rubbish had reached a height of about 2 metres.

Before the main house was built the site must long have remained unoccupied. Under the pavement lay 4 metres of rubbish above the floor of an earlier house. Three metres deeper again there were tablets of the time of Kadashmanturgu, Kadashmanbel, and Kurigalzu; and again, 2 to 3 metres deeper, were some of Samsuiluna, Ammiditana, and Samsuditana.

The mud-brick walls were plastered with mud, and over this was a wash of white gypsum mortar.

Not one of the chambers showed any traces from which we could infer the existence of a stairway to an upper storey. If there were steps, which we cannot doubt, they were certainly of wood, something like the simple stairways to the roof that are used at the present time by the people of Kweiresh (Fig. 238).

When the house was built, the entire area was first surrounded by a sloping wall without any toothed projections, filled up inside with earth, this forming a substantial terrace on which the actual building stood (Fig. 239). The top of the terrace was 1½ metres higher than the brick pavement of the street on the north. The terrace wall is not so thick as the outer walls of the superstructure, but it projects out on the outer side about as far as the toothed projections above it stand out, and thus forms a kind of plinth. Owing to the constant raising of the street level this is little observable; the plinth disappeared with the subsequent heightening of the street. The outer wall itself had more than 90 of those toothed projections, to which we have frequently referred, and is provided with a system of wooden braces, intended to strengthen the projections. A beam lies on the outside, parallel with each wall face, about the length of one projection, in the next brick course this is gripped at one end by a beam placed more or less at right angles to it. The outside must have appeared very much as it is figured in the reconstruction (Fig. 235). The frontage of another house in Merkes is given in Fig. 240.

For comparison we also give a ground-plan from Fara of about the fifth millennium (Fig. 241). It will show how few changes the internal arrangements of a Babylonian house underwent during the lapse of thousands of years. Nothing shows more conclusively than these ground-plans the immense age of Babylonian civilisation; for even in this remote period, which is in part prehistoric, they give clear indications of a yet earlier development from a presumably simpler and more primitive building.

The original Babylonian house, as we may assume it to have been from the present state of our knowledge, was probably a rectangular roofed-in space within a walled court. It is most desirable that we should obtain explicit evidence as to the form of the early Babylonian house in one of the prehistoric sites, but to do this is attended with difficulties. They occur generally in narrow crosscuts, or in deep trenches where the limited space renders the following up of these ancient sites very difficult. It would be necessary to open up a much wider area down to a considerable depth to afford sufficient material for arriving at conclusions, and at Surgul and El-Hibba, as well as at Fara, there was not time to do this.

In strange contrast to these Babylonian ground-plans is the palace of Telloh. The reason why the account given of it by de Sarzec is so difficult to understand, is because it was built at three different periods, which should be clearly differentiated from each other, but which are all placed together and attributed to Gudea as the builder. Only a small part, on the contrary, the inner part B (Fig. 242), which is not organically connected with the building as a whole, belongs to Gudea. All the rest is later, most of it very much later. In 1886 I examined and surveyed all that then survived of the palace. The dotted portion of the plan I give here was then no more to be seen; these walls had already been carried away by brick robbers. At my second visit in 1898 the work of destruction had not been carried much further. The ancient portion, marked black on the plan, represents part of the facing wall of a zikurrat that lay behind it to the south-east, with a stepped and grooved façade and a large gutter for water, such as is usually found in ancient zikurrats. This portion is built of Gudea bricks laid in asphalt and mud. The grooved façade of a lower-lying wall that belongs to it, which formed part of a lower floor, a terrace, or a later kisu, is given by de Sarzec in the court (B); on the north-east various chambers abut on it, the walls of which are built with re-used Gudea bricks. The asphalt still clings in many places to the lower side of the bricks, and the drops of asphalt which naturally when the bricks were first used fell on the outer face of the bricks and left slight traces pointing downwards, in their later use point upwards.

The north-western outer front of rooms 31, 29, show simple grooved work, which disappeared behind the walls of the later building round court C, and were cut off by the surrounding wall. In our plan these portions are heavily scored. Of the third later building, lightly scored in the plan, which was also built partly of re-used Gudea bricks, and partly of unstamped bricks, laid in mud mortar, two courts can be recognised (C and B). Here we do not find the unmistakably important principal chamber, which is so remarkable a feature of genuine Babylonian buildings. In chambers 11, 35, and 18 de Sarzec reports table-shaped fireplaces, such as I have never found either in Old or Neo-Babylonian buildings, while, on the contrary, such a flat raised hearth is found in chamber XXXV of an unmistakably Parthian house in Nippur that has a peristyle (Fisher, Journal of the Archaeological Institute of America, vol. viii., 1904, No. 4, p. 411). In the pavement of the court adjoining it, the well-known bricks of Adadnadinakhe are said to have been found. An examination of the south-eastern quarter, which must evidently have been already much destroyed at the time of de Sarzec, furnishes the strongest evidence against his representations. Thus in front of 23, he represents a door as constructed of a thick and a very thin wall, and at 24 and 25 he reports a door embrasure actually standing opposite a door-opening. We are therefore forced to the conclusion that here also buildings of entirely different and disconnected periods have been erroneously placed together by the modern draughtsman as having formed one complete building. The peristyle that we expect to find in connection with the two courts (C and B) should be placed in A.