SIR LEONARD WOOLLEY (1880-1960), the great archaeologist whose most significant work was probably the revelations about the Sumerians arising out of the excavations at Ur, began his career as Assistant Keeper of the Ashmolean Museum at Oxford (1905-1907). From 1912-14 and again in 1919 he worked with T. E. Lawrence to clear the Hittite city of Carchemish and explore surface remains in Sinai. For two seasons he worked with the Egypt Exploration Society at Tell el Amarna, and from 1922-1934 he directed the joint British Museum and University of Pennsylvania Expedition working at Ur of the Chaldees. His great interest in discovering the connections between the civilizations of the Aegean and Mesopotamia led him to the later excavations at the Syrian seaport of al-Mina (1936) and Tell Aichana-the ancient Alalakh-(1937-39, 1946-49).

In the largest of all the stone-built royal tombs, which had been entered by robbers and most thoroughly plundered, there remained only one corner of the last chamber to be cleared, and we had given up expectation of any 'finds' when suddenly a loose bit of shell inlay turned up, and the next minute the foreman's hand, carefully brushing away the earth, laid bare the corner of a mosaic in lapis lazuli and shell.

... at the time we had very little idea of what it might be: the wooden background had perished entirely, and the tiny pieces of inlay, though they kept their relative positions in the soil, were all quite loose; falling stones had bent and twisted the once flat panel, while as the wood decayed and the fragments sank back into the empty space behind, their different thickness made the surface of them rough and uneven. So delicate was the task of removing the dirt without further disturbing the mosaic that only about a square inch could be dealt with at a time--each section was waxed as soon as cleared, but so much of the surrounding dirt mingled with the hot wax that the face of the panel became invisible. When at last it could be lifted from the earth, I knew that we had found a very fine thing, but should have been hard put to it to say exactly what it was.

Now, it would have been perfectly feasible to take the mosaic to pieces, bit by bit, and re-make it on a new background, and the task might have been done as well by the modern craftsman as by the old, but the panels would have been the work of a modern craftsman.

What was done was this. The two sides of the panel were separated, and waxed cloth was fixed to the back of the inlay and the face of it was roughly cleaned; it was then laid face downwards on glass and warmed until the wax was soft, and it was pressed with the fingers from behind until by looking underneath one could be sure that each fragment of the inlay was in direct contact with the glass. The panel was now flat, but the pattern was much distorted; the edges of the mosaic fragments had lost contact in the ground and earth and powdered bitumen had filtered between them, and now wax as well, so that while some overlapped, others were widely apart. The next stage was to remove the cloth from the back, leaving the mosaic virtually loose on the glass, and to pick out all foreign matter, and then by sideways pressure with the fingers coax the pieces together. When this was done, fresh wax and cloth were applied behind and a proper backing fixed on.

The result of this is that the mosaic is not nearly so regular or smooth as the Sumerian artist made it, but what we possess is the work of that artist uninterfered with except by the accidents of time; the pieces of shell and lapis which he put together no one else has taken apart and re-set.

... the labour of restoration was at the same time a process of discovery; the work in the field had really been done in the dark, and it was only when the panels were cleaned and had begun to take shape in the laboratory that their importance could be recognised. There are two main panels, rectangular and measuring 22 inches long by 9 inches high, and two triangular pieces which formed the ends; these were fixed together so that the larger sides sloped inwards and the whole thing was fastened on to the end of a pole...