No cultus image has been found. In many temples the postaments are supported on gigantic and deep foundations although their height above the flooring is invariably very slight. We may conclude from this that they were intended to bear heavy weights. Herodotus (i. 183) states that the seated statue of Marduk in the temple Esagila with its accessories weighed 800 talents of gold, and speaks of another sacred statue 12 ells high in massive gold. It is obvious that such costly statues could not survive to a later period. Their immense value was their certain ruin. Thus if we attempt to form an idea of the appearance of a temple statue we must have recourse to the terra-cottas. They are found by many thousands over the entire city area. Only a few of these are uninjured, by far the largest number are in small fragments. These, however, even when they are very small, can be recognised as belonging to a well-known type. Great as is the number of these terra-cottas, the number of different types is proportionately small. They appear to have been used as a species of household gods, and they are all of the same modest size, about 12 centimetres high. They are moulded, and the design is only on the front, the back is smooth and merely rounded; thus they are absolutely full face. The men are clothed, but the women are nude until the Greek times, when the woman with a child in her arms appears for the first time draped. All other female types remain unclothed up to the latest period. With regard to technique, in the later Greek period a slight change was introduced, and a mould was made for the back as well as for the front of the figure; the two edges must have been fastened together, leaving the inside hollow. These terra-cottas now show only the yellowish, or occasionally reddish, colour of the burnt clay, but originally they were painted, as we learn from some few better-preserved specimens. Of the time of Nebuchadnezzar and earlier there are some that appear to be glazed in one colour; but the glaze is always so much decayed that it is impossible to say whether or not the figures were originally glazed in a variety of colours. The characteristic form of each of these somewhat rare types of divinities occurs with such convincing similarity in the numerous examples of each type that the cultus image of the respective gods in their temples must have had the same form. Now, if we find in, or near, one temple a considerable number identical in type, we are, in some measure, justified in forming from them a conjectural restoration of the divine image. We must bear in mind, however, that coincidence may here play a part. In any case, I am quite prepared later to modify the conclusions here put forward with regard to each temple, in favour of what may be thought more solid and more probable considerations.