CHAPTER II

THE SITES OF EARLY CITIES AND THE RACIAL CHARACTER OF THEIR INHABITANTS

The excavations which have been conducted on the sites of early Babylonian cities since the middle of last century have furnished material for the reconstruction of their history, but during different periods and for different districts it varies considerably in value and amount. While little is known of the earlier settlements in Akkad, and the very sites of two of its most famous cities have not yet been identified, our knowledge of Sumerian history and topography is relatively more complete. Here the cities, as represented by the mounds of earth and _débris_ which now cover them, fall naturally into two groups. The one consists of those cities which continued in existence during the later periods of Babylonian history. In their case the earliest Sumerian remains have been considerably disturbed by later builders, and are now buried deep beneath the accumulations of successive ages. Their excavation is consequently a task of considerable difficulty, and, even when the lowest strata are reached, the interpretation of the evidence is often doubtful. The other group comprises towns which were occupied mainly by the Sumerians, and, after being destroyed at an early date, were rarely, or never, reoccupied by the later inhabitants of the country. The mounds of this description, so far as they have been examined, have naturally yielded fuller information, and they may therefore be taken first in the following description of the early sites.

The greater part of our knowledge of early Sumerian history has been derived from the wonderfully successful series of excavations carried out by the late M. de Sarzec at Tello,[1] between 1877 and 1900, and continued for some months in 1903 by Captain (now Commandant) Gaston Cros. These mounds mark the site of the city of Shirpurla or Lagash, and lie a few miles to the north-east of the modern village of Shatra, to the east of the Shatt el-Hai, and about an hour's ride from the present course of the stream. It is evident, however, that the city was built upon the stream, which at this point may originally have formed a branch of the Euphrates,[2] for there are traces of a dry channel upon its western side.

The name of the city is expressed by the signs _shir-pur-la_ (_-ki_), which are rendered in a bilingual incantation-text as Lagash.[3] Hitherto it has been generally held that Shirpurla represented the Sumerian name of the city, which was known to the later Semitic inhabitants as Lagash, in much the same way as Akkad was the Semitic name for Agade, though in the latter case the original name was taken over. But the prolonged excavations carried out in the mounds of Tello have failed to bring to light any Babylonian remains later than the period of the kings of Larsa who were contemporaneous with the First Dynasty of Babylon. At that time the city appears to have been destroyed, and to have lain deserted and forgotten until it was once more inhabited in the second century B.C. Thus it is difficult to find a reason for a second name. We may therefore assume that the place was called Lagash by the Sumerians, and that the signs which can be read as Shirpurla represent a traditional ideographic way of writing the name among the Sumerians themselves. There is no difficulty in supposing that the city's name and the way of writing it were preserved in Babylonian literature, although its site had been forgotten.

The group of mounds and hillocks which mark the site of the ancient city and its suburbs form a rough oval, running north and south, and measuring about two and a half miles long and one and a quarter broad. During the early spring the limits of the city are clearly visible, for its ruins stand out as a yellow spot in the midst of the light green vegetation which covers the surrounding plain. The grouping of the principal mounds may be seen in the accompanying plan, in which each contour-line represents an increase of one metre in height above the desert level. The three principal mounds in the centre of the oval, marked on the plan by the letters A, K, and V,[4] are those in which the most important discoveries have been made. The mound A, which rises steeply towards the north-west end of the oval, is known as the Palace Tell, since here was uncovered a great Parthian palace, erected immediately over a building of Gudea, whose bricks were partly reused and partly imitated. In consequence of this it was at first believed to be a palace of Gudea himself, an error that was corrected on the discovery that some of the later bricks bore the name of Hadadnadinakhe in Aramean and Greek characters, proving that the building belonged to the Seleucid era, and was probably not earlier than about 130 B.C. Coins were also found in the palace with Greek inscriptions of kings of the little independent province or kingdom of Kharakene, which was founded about 160 B.C. at the mouth of the Shatt el-'Arab. But worked into the structure of this late palace were the remains of Gudea's building, which formed part of E-ninnû, the temple of the city-god of Lagash. Of Gudea's structure a gateway and part of a tower are the portions that are best preserved,[5] while under the south-east corner of the palace was a wall of the rather earlier ruler Ur-Bau.[6]

In the lower strata no other earlier remains were brought to light, and it is possible that the site of the temple was changed or enlarged at this period, and that in earlier times it stood nearer the mound K, where the oldest buildings in Tello have been found. Here was a storehouse of Ur-Ninâ,[7] a very early patesi of the city and the founder of its most powerful dynasty, and in its immediate neighbourhood were recovered the most important monuments and inscriptions of the earlier period. Beneath Ur-Ninâ's storehouse was a still earlier building,[8] and at the same deep level above the virgin soil were found some of the earliest examples of Sumerian sculpture that have yet been recovered. In the mound V, christened the "Tell of the Tablets," were large collections of temple-documents and tablets of accounts, the majority of them dating from the period of the Dynasty of Ur.

The monuments and inscriptions from Tello have furnished us with material for reconstructing the history of the city with but few gaps from the earliest age until the time when the Dynasty of Isin succeeded that of Ur in the rule of Sumer and Akkad. To the destruction of the city during the period of the First Dynasty of Babylon and its subsequent isolation we owe the wealth of early records and archaeological remains which have come down to us, for its soil has escaped disturbance at the hands of later builders except for a short interval in Hellenistic times. The fact that other cities in the neighbourhood, which shared a similar fate, have not yielded such striking results to the excavator, in itself bears testimony to the important position occupied by Lagash, not only as the seat of a long line of successful rulers, but as the most important centre of Sumerian culture and art.

The mounds of Surghul and El-Hibba, lying to the north-east of Tello and about six miles from each other, which were excavated by Dr. Koldewey in 1887, are instances in point. Both mounds, and particularly the former, contain numerous early graves beneath houses of unburnt brick, such as have subsequently been found at Fâra, and both cities were destroyed by fire probably at the time when Lagash was wiped out. From the quantities of ashes, and from the fact that some of the bodies appeared to have been partially burnt, Dr. Koldewey erroneously concluded that the mounds marked the sites of "fire-necropoles," where he imagined the early Babylonians burnt their dead, and the houses he regarded as tombs.[9] But in no period of Sumerian or Babylonian history was this practice in vogue. The dead were always buried, and any appearance of burning must have been produced during the destruction of the cities by fire. At El-Hibba remains were also visible of buildings constructed wholly or in part of kiln-baked bricks, which, coupled with the greater extent of its mounds, suggests that it was a more important Sumerian city than Surghul. This has been confirmed by the greater number of inscriptions which were found upon its site and have recently been published.[10] They include texts of the early patesis of Lagash, Eannatum and Enannatum I, and of the later patesi Gudea. A text of Gudea was also found at Surghul proving that both places were subject to Lagash, in whose territory they were probably always included during the periods of that city's power. That, apart from the graves, few objects of archaeological or artistic interest were recovered, may in part be traced to their proximity to Lagash, which as the seat of government naturally enjoyed an advantage in this respect over neighbouring towns.

During the course of her early history the most persistent rival of Lagash was the neighbouring city of Umma,[11] now identified with the mound of Jôkha, lying some distance to the north-west in the region between the Shatt el-Hai and the Shatt el-Kâr. Its neighbourhood and part of the mound itself are covered with sand-dunes, which give the spot a very desolate appearance, but they are of recent formation, since between them can still be seen traces of former cultivation. The principal mound is in the form of a ridge over half a mile long, running W.S.W. to E.N.E. and rising at its highest point about fifteen metres above the plain. Two lower extensions of the principal mound stretch out to the east and south-east.

No excavations have yet been conducted on this site, but it was visited by Dr. Andrae in the winter of 1902-3. He noted traces of a large building on a platform to the north of the principal ridge, marked A on the plan. It appears to have formed a square, its sides measuring seventy metres in length, and a small mound rises in the centre of it. Quantities of square, kiln-burnt bricks are scattered on the mound which covers it, and on the south side traces of a rectangular chamber are visible.[12] Numerous fragments of diorite also suggest the presence of sculptures, and at the south corner of the building, at the spot marked with a cross on the plan, the Germans found a fragment of diorite with part of a carefully chiselled inscription in archaic characters. The occurrence of unglazed potsherds, flint implements, and plano-convex bricks on other parts of the mound are an indication that, like Fâra, the site contains relics of still earlier habitation. Moreover, it is said that for years past Arab diggings have been carried out there, and early tablets and three cones of the patesi Galu-Babbar have reached Europe from this site. In view of the promising traces he noted and of the important part which the city played in early Sumerian history, it is almost to be regretted that Dr. Andrae did not substitute Jôkha for Abû Hatab as a site for his subsequent excavation.

Other mounds in the same neighbourhood also suggest prospects of success for the future excavator. One of these is Hammâm, which lies about seven and a half miles W.S.W. of Jôkha and close to the bed of the Shatt el-Kâr. It consists of a group of separate mounds, on one of which are the remains of a rectangular building resembling a ziggurat or temple-tower. Its side measures thirty metres, and it rises to a height of twelve metres above the surface of the mound, which in turn is three metres above the plain. Clay, in which layers of reeds are embedded, has been spread between the bricks as at Warka. More to the north of it in the same mound are traces of another building, possibly the temple of which it formed a part. To the south of Hammâm, and a little over three miles to the west of the Shatt el-Kâr is Tell 'Îd, another site which might repay excavation. It consists of a well-defined mound, about thirty metres high at the summit, and is visible from a considerable distance. Unlike Hammâm and Jôkha, however, it shows no trace upon its surface of any building, and there are no potsherds, bricks, or other objects scattered on the mound to afford an indication of its date. Both Tell 'Îd and Hammâm stand on a slightly elevated tract of desert soil, some ten miles broad, which raises them above the marshes caused by the inundations of the Euphrates. On the same tract farther to the south are Senkera and Warka, which were examined by Loftus in the early fifties.[13]

Of the early sites in the region of the Shatt el-Kâr the mounds at Fâra have been the most productive of remains dating from the prehistoric period of Sumerian culture. Systematic excavations were begun here by Dr. Koldewey in 1902,[14] and were continued in the following year by Drs. Andrae and Noeldeke.[15] The accompanying plan will give some idea of the extensive area occupied by the mounds, and of the method adopted for ascertaining their contents without too great an expenditure of time. The Arabic numerals against the contour lines indicate their height in metres above the level of the plain. Roman figures are set at each end of the trenches in the order in which they were cut. Thus the first two trenches (I. and II.), running from north to south and from east to west respectively, were cut across the mounds by Dr. Koldewey to gain some idea of their general character. The subsequent trenches were all cut parallel to the second through the higher portions of the site, a few of them being extended so as to cover the lower detached mounds to the east. In the plan the trenches are marked as continuous, but actually each consists of a series of short sections, divided by bands of soil left uncut. These hold up the sides of the trench and leave passages for crossing from one side to the other.[16] Whenever a trench discloses the remains of a building it can be completely uncovered and the trench afterwards continued until another building is disclosed. In the plan the principal cleared areas are outlined, and the position of walls which were uncovered within them is indicated by fine lines.

In the course of the systematic excavation of the site, it was clearly established that all the mounds at Fâra belong to a very early period. In many places the trenches cut through thick strata of ashes and charred remains, and it was seen that the whole settlement had been destroyed by fire, and that the greater part of it had never been reoccupied. All trace of buildings practically ceased at a depth of more than two metres beneath the present surface, and those that were excavated appear to belong to a single epoch. Their early period is attested by the fact that they are all built of plano-convex bricks,[17] both baked and unbaked, with thumb-marks or lines impressed by the finger on their upper surface. Many of them were clearly dwelling-houses, consisting of chambers grouped around a rectangular court; others are of circular form, measuring from two to five metres across, and their use has not been determined.[18] It has been suggested that the latter may have served as wells, and it is true that they generally descend to a depth of about four metres below the level of the plain. But they are scattered so thickly in the mound that this explanation of their use is scarcely adequate; moreover each was roofed in with an arch of overlapping bricks laid horizontally. They may have been cisterns, or designed for receiving refuse-water from the houses, but against this view is to be set the fact that they are not connected in any way with the numerous brick channels and clay drains that were discovered. Similar constructions were found at Surghul, and nothing in the _débris_ which filled them, either there or at Fâra, has thrown light upon the purpose which they served.

The most interesting discoveries at Fâra were the graves. These consist of two classes, sarcophagus-graves and mat-burials. The sarcophagi are of unglazed clay, oval in form, with flat bottoms and upright sides, and each is closed with a terra-cotta lid. In the mat-burials the corpse with its offerings was wrapped in reed-matting and placed in a grave dug in the soil. The bodies were never buried at length, for in both classes of graves the skeletons are found lying on their sides with their legs and arms bent. The right hand usually holds a drinking-cup, of clay, stone, copper or shell, which it appears to be raising to the mouth; and near the skull are often other vessels and great water-pots of clay. In the graves the weapons of the dead man were placed, and the tools and ornaments he had used during life. Copper spear-heads and axes were often found, and the blades of daggers with rivets for a wooden handle, and copper fish-hooks and net-weights.

The ornaments were very numerous, the wealthy wearing bead-necklaces of agate and lapis-lazuli, the poorer contenting themselves with paste or shell, while silver finger-rings and copper arm-rings were not uncommon. A very typical class of grave-furniture consisted of palettes or colour-dishes, made of alabaster, often of graceful shape, and sometimes standing on four feet. There is no doubt as to their use, for colour still remains in many of them, generally black and yellow, but sometimes a light rose and a light green. Since all other objects in the graves were placed there for the personal use of the dead man, we may infer that colour was employed at that period for painting the body.

No difference in age appears to have separated the two classes of burial, for the offerings are alike in each, and the arrangement of the bodies is the same. Why there should have been a difference in custom it is difficult to say. It might be inferred that the sarcophagus was a mark of wealth, were it not that the offerings they contain are generally more scanty than in the mat-burials. Whatever may be the explanation there is little doubt that they belong to the same race and period. Moreover, we may definitely connect the graves with the buildings under which they are found, for in some of them were seal-cylinders precisely similar to others found in the _débris_ covering the houses, and the designs upon them resemble those on sealings from the strata of ashes in the upper surface of the mounds. The seals are generally of shell or limestone, rarely of harder stone, and the designs represent heroes and mythological beings in conflict with animals. The presence of the sealings and seal-cylinders, resembling in form and design those of the early period at Tello, in itself suggests that Fâra marks the site of an early Sumerian town. This was put beyond a doubt by the discovery of clay tablets in six of the houses,[19] where they lay on the clay floor beneath masses of charred _débris_ which had fallen from the roof; beside them were objects of household use, and in one room the remains of a charred reed-mat were under them. The tablets were of unbaked clay, similar in shape to early contracts from Tello, and the texts upon them, written in extremely archaic characters, referred to deeds of sale.

There is thus no doubt as to the racial character of the inhabitants of this early settlement. The discovery of a brick inscribed with the name of Khaladda, patesi of Shuruppak, proved that Fâra was the site of the ancient city which later tradition regarded as the scene of the Deluge. Khaladda's inscription is not written in very archaic characters, and he probably lived in the time of the kings of Sumer and Akkad. We may thus infer that Shuruppak continued to exist as a city at that period, but the greater part of the site was never again inhabited after the destruction of the early town by fire. We have described its remains in some detail as they are our most valuable source of information concerning the earliest Sumerians in Babylonia. Until the objects that were found have been published it is difficult to determine accurately its relation in date to the earlier remains at Tello. A few fragments of sculpture in relief were discovered in the course of the excavations, and these, taken in conjunction with the cylinder-seals, the inscribed tablets, and the pottery, suggest that no long interval separated its period from that of the earliest Sumerians of history.

A less exhaustive examination of the neighbouring mounds of Abû Hatab was also undertaken by Drs. Andrae and Noeldeke. This site lies to the north of Fâra, and, like it, is close to the Shatt el-Kâr.[20] The southern part of the tell could not be examined because of the modern Arab graves which here lie thick around the tomb of the Imâm Sa'îd Muhammad. But the trenches cut in the higher parts of the mound, to the north and along its eastern edge, sufficed to indicate its general character.[21] Earlier remains, such as were found at Fâra, are here completely wanting, and it would appear to be not earlier than the period of the kings of Sumer and Akkad. This is indicated by bricks of Bûr-Sin I., King of Ur, which were discovered scattered in _débris_ in the north-west part of the mound, and by the finding of case-tablets in the houses belonging to the period of the dynasties of Ur and Isin.[22] The graves also differed from those at Fâra, generally consisting of pot-burials. Here, in place of a shallow trough with a lid, the sarcophagus was formed of two great pots, deeply ribbed on the outside; these were set, one over the other, with their edges meeting, and after burial they were fixed together by means of pitch or bitumen. The skeleton is usually found within lying on its back or side in a crouching position with bent legs. The general arrangement of drinking-cups, offerings, and ornaments resembles that in the Fâra burials, so that the difference in the form of the sarcophagus is merely due to a later custom and not to any racial change. Very similar burials were found by Taylor at Mukayyar, and others have also been unearthed in the earlier strata of the mounds at Babylon.

The majority of the houses at Abû Hatab appear to have been destroyed by fire, and, in view of the complete absence of later remains, the tablets scattered on their floors indicate the period of its latest settlement. It thus represents a well-defined epoch, later than that of the mounds at Fâra, and most valuable for comparison with them. At neither Fâra nor Abû Hatab were the remains of any important building or temple disclosed, but the graves and houses of the common people have furnished information of even greater value for the archaeologist and historian. Another mound which should provide further material for the study of this earliest period is Bismâya, the site of the city at Adab, at which excavations were begun on December 25, 1903 by the University of Chicago and continued during the following year.[23] The mound of Hêtime to the west of Fâra, may, to judge from the square bricks and fragments of pot-burials that are found there, date from about the same period as Abû Hatab. But it is of small extent and height, the greater part being merely six or seven feet above the plain, while its two central mounds rise to a height of less than fourteen feet.

Such are the principal early Sumerian mounds in the region of the Shatt el-Kâr and the Shatt el-Hai. Other mounds in the same neighbourhood may well prove to be of equally early dates; but it should be noted that some of these do not cover Sumerian cities, but represent far later periods of occupation. The character of the extensive mound of Jidr to the east of Fâra and Abû Hatab is doubtful; but the use of lime-mortar in such remains as are visible upon the surface indicates a late epoch. A number of smaller tells may be definitely regarded as representing a settlement in this district during Sassanian times. Such are Dubâ'i, which, with two others, lies to the south of Fâra, and Bint el-Mderre to the east; to the same period may be assigned Menêdir, which lies to the north-east, beyond Deke, the nearest village to Fâra. This last mound, little more than a hundred yards long, covers the site of a burial-place; it has been completely burrowed through by the Arabs in their search for antiquities, and is now covered with fragments of sarcophagi. The mounds of Mjelli and Abû Khuwâsîj to the west of Fâra are probably still later, and belong to the Arab period.

It will have been noted that all the Sumerian mounds described or referred to in the preceding paragraphs cover cities which, after being burned down and destroyed in a comparatively early period, were never reoccupied, but were left deserted. Lagash, Umma, Shuruppak, Kisurra, and Adab play no part in the subsequent history of Babylonia. We may infer that they perished during the fierce struggle which took place between the Babylonian kings of the First Dynasty and the Elamite kings of Larsa. At this time city after city in Sumer was captured and retaken many times, and on Samsu-iluna's final victory over Rîm-Sin, it is probable that he decided to destroy many of the cities and make the region a desert, so as to put an end to trouble for the future. As a matter of fact, he only succeeded in shifting the area of disturbance southwards, for the Sumerian inhabitants fled to the Sea-country on the shores of the Persian Gulf; and to their influence, and to the reinforcements they brought with them, may be traced the troubles of Samsu-iluna and his son at the hands of Iluma-ilu, who had already established his independence in this region. Thus Samsu-iluna's policy of repression was scarcely a success; but the archaeologist has reason to be grateful to it. The undisturbed condition of these early cities renders their excavation a comparatively simple matter, and lends a certainty to conclusions drawn from a study of their remains, which is necessarily lacking in the case of more complicated sites.

Another class of Sumerian cities consists of those which were not finally destroyed by the Western Semites, but continued to be important centres of political and social life during the later periods of Babylonian history. Niffer, Warka, Senkera, Mukayyar, and Abû Shahrain all doubtless contain in their lower strata remains of the early Sumerian cities which stood upon their sites; but the greater part of the mounds are made up of ruins dating from a period not earlier than that of the great builders of the Dynasty of Ur. In Nippur, during the American excavations on this site, the history of Ekur, the temple of the god Enlil, was traced back to the period of Shar-Gani-Sharri and Narâm-Sin;[24] and fragments of early vases found scattered in the _débris_ beneath the chambers on the south-east side of the Ziggurat, have thrown valuable light upon an early period of Sumerian history. But the excavation of the pre-Sargonic strata, so far as it has yet been carried, has given negative rather than positive results. The excavations carried out on the other sites referred to were of a purely tentative character, and, although they were made in the early fifties of last century, they still remain the principal source of our knowledge concerning them.

Some idea of the extent of the mounds of Warka may be gathered from Loftus's plan. The irregular circle of the mounds, marking the later walls of the city, covers an area nearly six miles in circumference, and in view of this fact and of the short time and limited means at his disposal, it is surprising that he should have achieved such good results. His work at Buwârîya, the principal mound of the group (marked A on the plan), resulted in its identification with E-anna, the great temple of the goddess Ninni, or Ishtar, which was enormously added to in the reign of Ur-Engur. Loftus's careful notes and drawings of the facade of another important building, covered by the mound known as Wuswas (B), have been of great value from the architectural point of view, while no less interesting is his description of the "Cone Wall" (at E on the plan), consisting in great part of terra-cotta cones, dipped in red or black colour, and arranged to form various patterns on the surface of a wall composed of mud and chopped straw.[25] But the date of both these constructions is uncertain. The sarcophagus-graves and pot-burials which he came across when cutting his tunnels and trenches are clearly contemporaneous with those at Abû Hatab, and the mound may well contain still earlier remains. The finds made in the neighbouring mounds of Senkera (Larsa), and Tell Sifr, were also promising,[26] and, in spite of his want of success at Tell Medîna, it is possible that a longer examination would have yielded better results.

The mounds of Mukayyar, which mark the site of Ur, the centre of the Moon-god's cult in Sumer, were partly excavated by Taylor in 1854 and 1855.[27] In the northern portion of the group he examined the great temple of the Moon-god (marked A on the plan), the earliest portions of its structure which he came across dating from the reigns of Dungi and Ur-Engur. Beneath a building in the neighbourhood of the temple (at B on the plan) he found a pavement consisting of plano-convex bricks, a sure indication that at this point, at least, were buildings of the earliest Sumerian period, while the sarcophagus-burials in other parts of the mound were of the early type. Taylor came across similar evidence of early building at Abû Shahrain,[28] the comparatively small mound which marks the site of the sacred city of Eridu, for at a point in the south-east side of the group he uncovered a building constructed of bricks of the same early character.

At Abû Shahrain indeed we should expect to find traces of one of the earliest and most sacred shrines of the Sumerians, for here dwelt Enki, the mysterious god of the deep. The remains of his later temple now dominates the group, the great temple-tower still rising in two stages (A and B) at the northern end of the mound. Unlike the other cities of Sumer, Eridu was not built on the alluvium. Its situation is in a valley on the edge of the Arabian desert, cut off from Ur and the Euphrates by a low pebbly and sandstone ridge. In fact, its ruins appear to rise abruptly from the bed of an inland sea, which no doubt at one time was connected directly with the Persian Gulf; hence the description of Eridu in cuneiform literature as standing "on the shore of the sea." Another characteristic which distinguishes Eridu from other cities in Babylonia is the extensive use of stone as a building material. The raised platform, on which the city and its temple stood, was faced with a massive retaining wall of sandstone, no doubt obtained from quarries in the neighbourhood, while the stairway (marked D on the plan) leading to the first stage of the temple-tower had been formed of polished marble slabs which were now scattered on the surface of the mound. The marble stairs and the numerous fragments of gold-leaf and gold-headed and copper nails, which Taylor found at the base of the second stage of the temple-tower, attest its magnificence during the latest stage of its history. The name and period of the city now covered by the neighbouring mound of Tell Lahm, which was also examined by Taylor, have not yet been ascertained.

It will thus be seen that excavations conducted on the sites of the more famous cities of Sumer have not, with the single exception of Nippur, yielded much information concerning the earlier periods of history, while the position of one of them, the city of Isin, is still unknown. Our knowledge of similar sites in Akkad is still more scanty. Up to the present time systematic excavations have been carried out at only two sites in the north, Babylon and Sippar, and these have thrown little light upon the more remote periods of their occupation. The existing ruins of Babylon date from the period of Nebuchadnezzar II., and so thorough was Sennacherib's destruction of the city in 689 B.C., that, after several years of work, Dr. Koldewey concluded that all traces of earlier buildings had been destroyed on that occasion. More recently some remains of earlier strata have been recognized, and contract-tablets have been found which date from the period of the First Dynasty. Moreover, a number of earlier pot-burials have been unearthed, but a careful examination of the greater part of the ruins has added little to our knowledge of this most famous city before the Neo-Babylonian era. The same negative results were obtained, so far as early remains are concerned, from the less exhaustive work on the site of Borsippa. Abû Habba is a far more promising site, and has been the scene of excavations begun by Mr. Rassam in 1881 and 1882, and renewed by Père Scheil for some months in 1894, while excavations were undertaken in the neighbouring mounds of Deir by Dr. Wallis Budge in 1891. These two sites have yielded thousands of tablets of the period of the earliest kings of Babylon, and the site of the famous temple of the Sun-god at Sippar, which Narâm-Sin rebuilt, has been identified, but little is yet accurately known concerning the early city and its suburbs. The great extent of the mounds, and the fact that for nearly thirty years they have been the happy hunting-ground of Arab diggers, would add to the difficulty of any final and exhaustive examination. It is probably in the neighbourhood of Sippar that the site of the city of Agade, or Akkad, will eventually be identified.

Concerning the sites of other cities in Northern Babylonia, considerable uncertainty still exists. The extensive mounds of Tell Ibrâhîm, situated about four hours to the north-east of Hilla, are probably to be identified with Cutha, the centre of the cult of Nergal, but the mound of 'Akarkûf, which may be seen from so great a distance on the road between Baghdad and Falûja, probably covers a temple and city of the Kassite period. Both the cities of Kish and Opis, which figure so prominently in the early history of the relations between Sumer and Akkad, were, until quite recently, thought to be situated close to one another on the Tigris. That Opis lay on that river and not on the Euphrates is clear from the account which Nebuchadnezzar II. has left us of his famous fortifications of Babylon,[29] which are referred to by Greek writers as "the Median Wall" and "the Fortification of Semiramis."

The outermost ring of Nebuchadnezzar's triple line of defence consisted of an earthen rampart and a ditch, which he tells us extended from the bank of the Tigris above Opis to a point on the Euphrates within the city of Sippar, proving that Opis is to be sought upon the former river. His second line of defence was a similar ditch and rampart which stretched from the causeway on the bank of the Euphrates up to the city of Kish. It was assumed that this rampart also extended to the Tigris, although this is not stated in the text, and, since the ideogram for Opis is once rendered as Kesh in a bilingual incantation,[30] it seemed probable that Kish and Opis were twin cities, both situated on the Tigris at no great distance from each other. This view appeared to find corroboration in the close association of the two places during the wars of Eannatum, and in the fact that at the time of Enbi-Ishtar they seem to have formed a single state. But it has recently been shown that Kish lay upon the Euphrates,[31] and we may thus accept its former identification with the mound of El-Ohêmir where bricks were found by Ker Porter recording the building of E-meteursagga, the temple of Zamama, the patron deity of Kish.[32] Whether Opis is to be identified with the extensive mounds of Tell Manjûr, situated on the right bank of the Tigris in the great bend made by the river between Samarra and Baghdad, or whether, as appears more probable, it is to be sought further down stream in the neighbourhood of Seleucia, are questions which future excavation may decide.[33]

The brief outline that has been given of our knowledge concerning the early cities of Sumer and Akkad, and of the results obtained by the partial excavation of their sites, will have served to show how much still remains to be done in this field of archaeological research. Not only do the majority of the sites still await systematic excavation, but a large part of the material already obtained has not yet been published. Up to the present time, for instance, only the briefest notes have been given of the important finds at Fâra and Abû Hatab. In contrast to this rather leisurely method of publication, the plan followed by M. de Morgan in making available without delay the results of his work in Persia is strongly to be commended. In this connection mention should in any case be made of the excavations at Susa, since they have brought to light some of the most remarkable monuments of the early Semitic kings of Akkad. It is true the majority of these had been carried as spoil from Babylonia to Elam, but they are none the less precious as examples of early Semitic art. Such monuments as the recently discovered stele of Sharru-Gi, the statues of Manishtusu, and Narâm-Sin's stele of victory afford valuable evidence concerning the racial characteristics of the early inhabitants of Northern Babylonia, and enable us to trace some of the stages in their artistic development. But in Akkad itself the excavations have not thrown much light upon these subjects, nor have they contributed to the solution of the problems as to the period at which Sumerians and Semites first came in contact, or which race was first in possession of the land. For the study of these questions our material is mainly furnished from the Sumerian side, more particularly by the sculptures and inscriptions discovered during the French excavations at Tello.

It is now generally recognized that the two races which inhabited Sumer and Akkad during the early historical periods were sharply divided from one another not only by their speech but also in their physical characteristics.[34] One of the principal traits by which they may be distinguished consists in the treatment of the hair. While the Sumerians invariably shaved the head and face, the Semites retained the hair of the head and wore long beards. A slight modification in the dressing of the hair was introduced by the Western Semites of the First Babylonian Dynasty, who brought with them from Syria the Canaanite Bedouin custom of shaving the lips and allowing the beard to fall only from the chin; while they also appear to have cut the hair short in the manner of the Arabs or Nabateans of the Sinai peninsula.[35] The Semites who were settled in Babylonia during the earlier period, retained the moustache as well as the beard, and wore their hair long. While recognizing the slight change of custom, introduced for a time during the West Semitic domination, the practice of wearing hair and beard was a Semitic characteristic during all periods of history. The phrase "the black-headed ones," which is of frequent occurrence in the later texts, clearly originated as a description of the Semites, in contradistinction to the Sumerians with their shaven heads.

Another distinctive characteristic, almost equally striking, may be seen in the features of the face as represented in the outline engraving and in the sculpture of the earlier periods. It is true that the Sumerian had a prominent nose, which forms, indeed, his most striking feature, but both nose and lips are never full and fleshy as with the Semites. It is sometimes claimed that such primitive representations as occur upon Ur-Ninâ's bas-reliefs, or in Fig. 1 in the accompanying block, are too rude to be regarded as representing accurately an ethnological type. But it will be noted that the same general characteristics are also found in the later and more finished sculptures of Gudea's period. This fact is illustrated by the two black diorite heads of statuettes figured on the following page. In both examples certain archaic conventions are retained, such as the exaggerated line of the eyebrows, and the unfinished ear; but nose and lips are obviously not Semitic, and they accurately reproduce the same racial type which is found upon the earlier reliefs.

A third characteristic consists of the different forms of dress worn by Sumerians and Semites, as represented on the monuments. The earliest Sumerians wore only a thick woollen garment, in the form of a petticoat, fastened round the waist by a band or girdle. The garment is sometimes represented as quite plain, in other cases it has a scolloped fringe or border, while in its most elaborate form it consists of three, four, or five horizontal flounces, each lined vertically and scolloped at the edge to represent thick locks of wool.[36] With the later Sumerian patesis this rough garment has been given up in favour of a great shawl or mantle, decorated with a border, which was worn over the left shoulder, and, falling in straight folds, draped the body with its opening in front.[37] Both these Sumerian forms of garment are of quite different types from the Semitic loin-cloth worn by Narâm-Sin on his stele of victory, and the Semitic plaid in which he is represented on his stele from Pir Hussein.[38] The latter garment is a long, narrow plaid which is wrapped round the body in parallel bands, with the end thrown over the left shoulder. It has no slit, or opening, in front like the later Sumerian mantle, and, on the other hand, was not a shaped garment like the earlier Sumerian flounced petticoat, though both were doubtless made of wool and were probably dyed in bright colours.

Two distinct racial types are thus represented on the monuments, differentiated not only by physical features but also by the method of treating the hair and by dress. Moreover, the one type is characteristic of those rulers whose language was Sumerian, the other represents those whose inscriptions are in the Semitic tongue. Two apparent inconsistencies should here be noted. On the Stele of the Vultures, Eannatum and his soldiers are sculptured with thick hair flowing from beneath their helmets and falling on their shoulders. But they have shaven faces, and, in view of the fact that on the same monument all the dead upon the field of battle and in the burial mounds have shaven heads, like those of the Sumerians assisting at the burial and the sacrificial rites, we may regard the hair of Eannatum and his warriors as wigs, worn like the wigs of the Egyptians, on special occasions and particularly in battle. The other inconsistency arises from the dress worn by Hammurabi on his monuments. This is not the Semitic plaid, but the Sumerian fringed mantle, and we may conjecture that, as he wrote his votive inscriptions in the Sumerian as well as in the Semitic language, so, too, he may have symbolized his rule in Sumer by the adoption of the Sumerian form of dress.

It is natural that upon monuments of the later period from Tello both racial types should be represented. The fragments of sculpture illustrated in Figs. 6 and 7 may possibly belong to the same monument, and, if so, we must assign it to a Semitic king.[39] That on the left represents a file of nude captives with shaven heads and faces, bound neck to neck with the same cord, and their arms tied behind them. On the other fragment both captive and conqueror are bearded. The latter's nose is anything but Semitic, though in figures of such small proportions carved in relief it would perhaps be rash to regard its shape as significant. The treatment of the hair, however, in itself constitutes a sufficiently marked difference in racial custom. Fig. 8 represents a circular support of steatite, around which are seated seven little figures holding tablets on their knees; it is here reproduced on a far smaller scale than the other fragments. The little figure that is best preserved is of unmistakably Semitic type, and wears a curled beard trimmed to a point, and hair that falls on the shoulders in two great twisted tresses; the face of the figure on his left is broken, but the head is clearly shaved. A similar mixture of types upon a single monument occurs on a large fragment of sculpture representing scenes of worship,[40] and also on Sharru-Gi's monument which has been found at Susa.[41]

At the period from which these sculptures date it is not questioned that the Semites were in occupation of Akkad, and that during certain periods they had already extended their authority over Sumer. It is not surprising, therefore, that at this time both Sumerians and Semites should be represented side by side upon the monuments. When, however, we examine what is undoubtedly one of the earliest sculptured reliefs from Tello the same mixture of racial types is met with.

The object is unfortunately broken into fragments, but enough of them have been recovered to indicate its character. Originally, it consisted of two circular blocks, placed one upon the other and sculptured on their outer edge with reliefs. They were perforated vertically with two holes which were intended to support maces, or other votive objects, in an upright position. The figures in the relief form two separate rows which advance towards one another, and at their head are two chiefs, who are represented meeting face to face (Fig. 9). It will be noticed that the chief on the left, who carries a bent club, has long hair falling on the shoulders and is bearded. Four of his followers on another fragment (Fig. 10) also have long hair and beards. The other chief, on the contrary, wears no hair on his face, only on his head, and, since his followers have shaven heads and faces,[42] we may conjecture that, like Eannatum on the Stele of the Vultures, he wears a wig. All the figures are nude to the waist, and the followers clasp their hands in token of subordination to their chiefs.

The extremely rude character of the sculpture is a sufficient indication of its early date, apart from the fact that the fragments were found scattered in the lowest strata at Tello. The fashion of indicating the hair is very archaic, and is also met with in a class of copper foundation-figures of extremely early date.[43] The monument belongs to a period when writing was already employed, for there are slight traces of an inscription on its upper surface, which probably recorded the occasion of the meeting of the chiefs. Moreover, from a fifth fragment that has been discovered it is seen that the names and titles of the various personages were engraved upon their garments. The monument thus belongs to the earliest Sumerian period, and, if we may apply the rule as to the treatment of the hair which we have seen holds good for the later periods, it would follow that at this time the Semite was already in the land. The scene, in fact, would represent the meeting of two early chieftains of the Sumerians and Semites, sculptured to commemorate an agreement or treaty which they had drawn up.

By a similar examination of the gods of the Sumerians, as they are represented on the monuments, Professor Meyer has sought to show that the Semites were not only in Babylonia at the date of the earliest Sumerian sculptures that have been recovered, but also that they were in occupation of the country before the Sumerians. The type of the Sumerian gods at the later period is well illustrated by a limestone panel of Gudea, which is preserved in the Berlin Museum. The sculptured scene is one that is often met with on cylinder-seals of the period, representing a suppliant being led by lesser deities into the presence of a greater god. In this instance Gudea is being led by his patron deity Ningishzida and another god into the presence of a deity who was seated on a throne and held a vase from which two streams of water flow. The right half of the panel is broken, but the figure of the seated god may be in part restored from the similar scene upon Gudea's cylinder-seal. There, however, the symbol of the spouting vase is multiplied, for not only does the god hold one in each hand, but three others are below his feet, and into them the water falls and spouts again. Professor Meyer would identify the god of the waters with Anu, though there is more to be said for M. Heuzey's view that he is Enki, the god of the deep. We are not here concerned, however, with the identity of the deities, but with the racial type they represent. It will be seen that they all have hair and beards and wear the Semitic plaid, and form a striking contrast to Gudea with his shaven head and face, and his fringed Sumerian mantle.[44]

A very similar contrast is represented by the Sumerian and his gods in the earlier historical periods. Upon the Stele of the Vultures, for instance, the god Ningirsu is represented with abundant hair, and although his lips and cheeks are shaved a long beard falls from below his chin.[45] He is girt around the waist with a plain garment, which is not of the later Semitic type, but the treatment of the hair and beard is obviously not Sumerian. The same bearded type of god is found upon early votive tablets from Nippur,[46] and also on a fragment of an archaic Sumerian relief from Tello, which, from the rudimentary character of the work and the style of the composition, has been regarded as the most ancient example of Sumerian sculpture known. The contours of the figures are vaguely indicated in low relief upon a flat plaque, and the interior details are indicated only by the point. The scene is evidently of a mythological character, for the seated figure may be recognized as a goddess by the horned crown she wears. Beside her stands a god who turns to smite a bound captive with a heavy club or mace. While the captive has the shaven head and face of a Sumerian, the god has abundant hair and a long beard.[47]

Man forms his god in his own image, and it is surprising that the gods of the Sumerians should not be of the Sumerian type. If the Sumerian shaved his own head and face, why should he have figured his gods with long beards and abundant hair and have clothed them with the garments of another race? Professor Meyer's answer to the question is that the Semites and their gods were already in occupation of Sumer and Akkad before the Sumerians came upon the scene. He would regard the Semites at this early period as settled throughout the whole country, a primitive and uncultured people with only sufficient knowledge of art to embody the figures of their gods in rude images of stone or clay. There is no doubt that the Sumerians were a warrior folk, and he would picture them as invading the country at a later date, and overwhelming Semitic opposition by their superior weapons and method of attack. The Sumerian method of fighting he would compare to that of the Dorians with their closed phalanx of lance-bearing warriors, though the comparison is not quite complete, since no knowledge of iron is postulated on the part of the Sumerians. He would regard the invaders as settling mainly in the south, driving many of the Semites northward, and taking over from them the ancient centres of Semitic cult. They would naturally have brought their own gods with them, and these they would identify with the deities they found in possession of the shrines, combining their attributes, but retaining the cult-images, whose sacred character would ensure the permanent retention of their outward form. The Sumerians in turn would have influenced their Semitic subjects and neighbours, who would gradually have acquired from them their higher culture, including a knowledge of writing and the arts.

It may be admitted that the theory is attractive, and it certainly furnishes an explanation of the apparently foreign character of the Sumerian gods. But even from the archaeological side it is not so complete nor so convincing as at first sight it would appear. Since the later Sumerian gods were represented with full moustache and beard, like the earliest figures of Semitic kings which we possess, it would naturally be supposed that they would have this form in the still earlier periods of Sumerian history. But, as we have seen, their lips and cheeks are shaved. Are we then to postulate a still earlier Semitic settlement, of a rather different racial type to that which founded the kingdom of Kish and the empire of Akkad? Again, the garments of the gods in the earliest period have little in common with the Semitic plaid, and are nearer akin to the plainer form of garment worn by contemporary Sumerians. The divine headdress, too, is different to the later form, the single horns which encircle what may be a symbol of the date-palm,[48] giving place to a plain conical headdress decorated with several pairs of horns.

Thus, important differences are observable in the form of the earlier Sumerian gods and their dress and insignia, which it is difficult to reconcile with Professor Meyer's theory of their origin. Moreover, the principal example which he selected to illustrate his thesis, the god of the central shrine of Nippur, has since been proved never to have borne the Semitic name of Bêl, but to have been known under his Sumerian title of Enlil from the beginning.[49] It is true that Professor Meyer claims that this point does not affect his main argument;[50] but at least it proves that Nippur was always a Sumerian religious centre, and its recognition as the central and most important shrine in the country by Semites and Sumerians alike, tells against any theory requiring a comparatively late date for its foundation.

Such evidence as we possess from the linguistic side is also not in favour of the view which would regard the Semites as in occupation of the whole of Babylonia before the Sumerian immigration. If that had been the case we should naturally expect to find abundant traces of Semitic influence in the earliest Sumerian texts that have been recovered. But, as a matter of fact, no Semitism occurs in any text from Ur-Ninâ's period to that of Lugal-zaggisi with the single exception of a Semitic loan-word on the Cone of Entemena.[51] In spite of the scanty nature of our material, this fact distinctly militates against the assumption that Semites and Sumerians were living side by side in Sumer at the time.[52] But the occurrence of the Semitic word in Entemena's inscription proves that external contact with some Semitic people had already taken place. Moreover, it is possible to press the argument from the purely linguistic side too far. A date-formula of Samsu-iluna's reign has proved that the Semitic speech of Babylonia was known as "Akkadian,"[53] and it has therefore been argued that the first appearance of Semitic speech in the country must date from the establishment of Shar-Gani-sharri's empire with its capital at Akkad.[54] But there is little doubt that the Semitic kingdom of Kish, represented by the reigns of Sharru-Gi, Manishtusu and Urumush, was anterior to Sargon's empire,[55] and, long before the rise of Kish, the town of Akkad may well have been the first important centre of Semitic settlement in the north.

It would thus appear that at the earliest period of which remains or records have been recovered, Semites and Sumerians were both settled in Babylonia, the one race in the north, the other southwards nearer the Persian Gulf. Living at first in comparative isolation, trade and war would gradually bring them into closer contact. Whether we may regard the earliest rulers of Kish as Semites like their later successors, is still in doubt. The character of Enbi-Ishtar's name points to his being a Semite; but the still earlier king of Kish, who is referred to on the Stele of the Vultures, is represented on that monument as a Sumerian with shaven head and face.[56] But this may have been due to a convention in the sculpture of the time, and it is quite possible that Mesilim and his successors were Semites, and that their relations with the contemporary rulers of Lagash represent the earlier stages in a racial conflict which dominates the history of the later periods.

Of the original home of the Sumerians, from which they came to the fertile plains of Southern Babylonia, it is impossible to speak with confidence. The fact that they settled at the mouths of the great rivers has led to the suggestion that they arrived by sea, and this has been connected with the story in Berossus of Oannes and the other fish-men, who came up from the Erythraean Sea and brought religion and culture with them. But the legend need not bear this interpretation; it merely points to the Sea-country on the shores of the Gulf as the earliest centre of Sumerian culture in the land. Others have argued that they came from a mountain-home, and have cited in support of their view the institution of the ziggurat or temple-tower, built "like a mountain," and the employment of the same ideogram for "mountain" and for "land." But the massive temple-tower appears to date from the period of Gudea and the earlier kings of Ur, and, with the single exception of Nippur, was probably not a characteristic feature of the earlier temples; and it is now known that the ideogram for "land" and "mountain" was employed in the earlier periods for foreign lands, in contradistinction to that of the Sumerians themselves.[57] But, in spite of the unsoundness of these arguments, it is most probable that the Sumerians did descend on Babylonia from the mountains on the east. Their entrance into the country would thus have been the first of several immigrations from that quarter, due to climatic and physical changes in Central Asia.[58]

Still more obscure is the problem of their racial affinity. The obliquely set eyes of the figures in the earlier reliefs, due mainly to an ignorance of perspective characteristic of all primitive art, first suggested the theory that the Sumerians were of Mongol type; and the further developments of this view, according to which a Chinese origin is to be sought both for Sumerian roots and for the cuneiform character, are too improbable to need detailed refutation. A more recent suggestion, that their language is of Indo-European origin and structure,[59] is scarcely less improbable, while resemblances which have been pointed out between isolated words in Sumerian and in Armenian, Turkish, and other languages of Western Asia, may well be fortuitous. With the Elamites upon their eastern border the Sumerians had close relations from the first, but the two races do not appear to be related either in language or by physical characteristics. The scientific study of the Sumerian tongue, inaugurated by Professors Zimmern and Jensen, and more especially by the work of M. Thureau-Dangin on the early texts, will doubtless lead in time to more accurate knowledge on this subject; but, until the phonetic elements of the language are firmly established, all theories based upon linguistic comparisons are necessarily insecure.

In view of the absence of Semitic influence in Sumer during the earlier periods, it may be conjectured that the Semitic immigrants did not reach Babylonia from the south, but from the north-west, after traversing the Syrian coast-lands. This first great influx of Semitic nomad tribes left colonists behind them in that region, who afterwards as the Amurru, or Western Semites, pressed on in their turn into Babylonia and established the earliest independent dynasty in Babylon. The original movement continued into Northern Babylonia, and its representatives in history were the early Semitic kings of Kish and Akkad. But the movement did not stop there; it passed on to the foot of the Zagros hills, and left its traces in the independent principalities of Lulubu and Gutiu. Such in outline appears to have been the course of this early migratory movement, which, after colonizing the areas through which it passed, eventually expended itself in the western mountains of Persia. It was mainly through contact with the higher culture of the Sumerians that the tribes which settled in Akkad were enabled later on to play so important a part in the history of Western Asia.

[1] In point of time, the work of Loftus and Taylor (see below, pp. 32 ff.) preceded that of De Sarzec, but the results obtained were necessarily less complete. It would be out of place in the present volume to give any account of excavations in Assyria, as they have only an indirect bearing on the period here treated. For a chronological sketch of the early travellers and excavators, see Rogers, "History of Babylonia and Assyria," vol. i. pp. 100 ff., who also gives a detailed account of the decipherment of the cuneiform inscriptions; cf. also Fossey, "Manuel," I., pp. 6 ff. For a similar chronological treatment, but from the archaeological side, see the sections with which Hilprecht prefaces his account of the Nippur excavations in "Explorations in Bible Lands," pp. 7 ff.

[2] See above, p. 11.

[3] Cf. "Cun. Texts in the Brit. Mus.," Pt. XVI., pl. 36, l. 4 f.; as written here the name might also be read Lagarum or Lagadil. That Lagash is the correct reading is proved by the fragment of a duplicate text published in Reisner, "Sum.-Bab. Hymnen," pl. 126, No. 81, where the final character of the name is unmistakably written as _ash_; cf. Meissner, "Orient. Lit.-Zeit.," 1907, col. 385.

[4] Separate mounds in the group were referred to by De Sarzec under the letters A-P, P', and V. For the account of the diggings and their results, see E. de Sarzec and Léon Heuzey, "Découvertes en Chaldée" ("Description des fouilles" by De Sarzec; "Description des monuments" by Heuzey; "Partie épigraphique" by Amiaud and Thureau-Dangin), Paris, 1884-1906; see also Heuzey, "Une Villa royale chaldéenne," and "Revue d'Assyriologie," _passim_.

[5] The plate opposite p. 20 illustrates the way in which Gudea's gateway has been worked into the structure of the Parthian Palace. The slight difference in the ground-level of the two buildings is also clearly shown.

[6] See the plate opposite p. 26.**

[7] From the nature of this building Amiaud christened the mound the "Tell de la Maison des Fruits."

[8] A description of these buildings is given in Chap. IV., pp. 90 ff.**

[9] Cf. "Zeits. für Assyr." II., pp. 406 ff.

[10] Cf. Messerschmidt, "Vorderasiatische Schriftdenkmäler," p. v. f., pl. 1 ff.

[11] The name is still often transcribed as Gishkhu or Gishukh; for the reading Umma, supplied by a Neo-Babylonian vocabulary, see "Cun. Texts," XII., pl. 28, Obv., l. 7, and cf. Hrozný, "Zeits. für Assyr." XX. (1907), pp. 421 ff. For its identification with Jôkha, see Scheil, "Rec. de trav.," XIX., p. 63; cf. also XXI., p. 125.

[12] Cf. "Mitteil. der Deutsch. Orient-Gesellschaft," No. 16, p. 20 f. Dr. Andrae adds valuable notes on other mounds he visited during this journey.

[13] See below, p. 33 f.

[14] See "Mitteil. der Deutsch. Orient-Gesellschaft," No. 15, p. 9 ff.

[15] _Op. cit._, No. 17, p. 4 ff.

[16] Each section of a trench is also given a letter, so that such a symbol as IV. _b_ or XII. _x_ indicates within very precise limits the _provenance_ of any object discovered. The letter A on the plan marks the site of the house built by the expedition.

[17] This form of brick is characteristic of the Pre-Sargonic period; cf. p. 91.

[18] The positions of some of the larger ones, which were excavated in the northern part of the mounds, are indicated by black dots in the plan.

[19] The houses with the clay tablets were found in trenches VII., IX., XIII., and XV.

[20] In the folding map Fâra has been set on the right bank of the Shatt el-Kâr, in accordance with Loftus's map published in "Travels and Researches in Chaldaea and Susiana." From Andrae's notes it would seem that Abû Hatab, and probably Fâra also, lie on the east or left bank. But the ancient bed of the stream has disappeared in many places, and is difficult to follow, and elsewhere there are traces of two or three parallel channels at considerable distances apart, so that the exact position of the original bed of the Euphrates is not certain at this point.

[21] In the plan the trenches and excavated sites are lettered from A to K. The figures, preceded by a cross, give in metres and centimetres the height of the mound at that point above the level of the plain.

[22] Itûr-Shamash, whose brick-inscription furnished the information that Abû Hatab is the site of the city Kisurra, is to be set towards the end of this period; see below, Chap. XI., and cf. p. 283 f., n. 1.

[23] See the extracts from the "Reports of the Expedition of the Oriental Exploration Fund (Babylonian Section) of the University of Chicago," which were issued to the subscribers.

[24] See below, Chap. IV., pp. 85 ff.

[25] See "Chaldaea and Susiana," pp. 174 ff. and 188 f.

[26] Op. cit., pp. 244 ff., 266 ff.

[27] See his "Notes on the ruins of Mugeyer" in the _Journal of the Royal Asiatic Society_, 1855, pp. 260 ff., 414 f.

[28] See his "Notes on Abû Shahrein and Tel el-Lahm," _op. cit._, p. 409. The trench which disclosed this structure, built of uninscribed plano-convex bricks laid in bitumen, was cut near the south-eastern side of the ruins, between the mounds F and G (see plan), and to the north-east of the gulley.

[29] See Weissbach, "Wâdī Brîssā," Col. VI., ll. 46 ff., and cf. pp. 39 ff.

[30] The incantation is the one which has furnished us with authority for reading the name of Shirpurla as Lagash (see above, p. 17, n. 3). It is directed against the machinations of evil demons, and in one passage the powers for good inherent in the ancient cities of Babylonia are invoked on behalf of the possessed man. Here, along with the names of Eridu, Lagash, and Shuruppak, occurs the ideogram for Opis, which is rendered in the Assyrian translation as _Ki-e-shi_, _i.e._ Kesh, or Kish (cf. Thompson, "Devils and Evil Spirits," vol. i., p. 162 f.)

[31] See above, p. 9.

[32] See George Smith, "Trans. Soc. Bibl. Arch.," III., p. 364, and cf. Thureau-Dangin, "Orient. Lit.-Zeit.," 1909, col. 205 f.

[33] The fact that in an early Babylonian geographical list ("Cun. Inscr. West. Asia," Vol. IV., pl. 36, No. 1) the name of Opis is mentioned after a number of Sumerian cities, is no indication that the city itself, or another city of the same name, was regarded as situated in Sumer, as suggested by Jensen (cf. "Zeits. für Assyr.," XV., pp. 210 ff.); the next two names in the list are those of Magan and Melukhkha.

[34] For the fullest treatment of this subject, see Meyer, "Sumerier und Semiten in Babylonien" (Abh. der Königl. Preuss. Akad. der Wissenschaft., 1906).

[35] Cf. Herodotus, III., 8.

[36] The women of the earlier period appear to have worn a modified form of this garment, made of the same rough wool, but worn over the left shoulder (see below, p. 112, Fig. 43). On the Stele of the Vultures, Eannatum, like his soldiers, wears the petticoat, but this is supplemented by what is obviously a separate garment of different texture thrown over the left shoulder so as to leave the right arm free; this may have been the skin of an animal worn with the natural hair outside (see the plate opposite p. 124).

[37] A very similar fringed mantle was usually worn by the Sumerian women of the later period, but it was draped differently upon the body. Pressed at first over the breasts and under each arm, it is crossed at the back and its ends, thrown over the shoulders, fall in front in two symmetrical points; for a good example of the garment as seen from the front, see below, p. 71.

[38] See below, p. 245, Fig. 59.

[39] Remains of an inscription upon Fig. 6 treat of the dedication of a temple to the god Ningirsu, and to judge from the characters it probably does not date from a period earlier than that of Gudea.

[40] See the plate facing p. 52, and cf. p. 68 f.

[41] See below, Chap. VIII., pp. 220 ff.

[42] According to the traces on the stone the figure immediately behind the beardless chief has a shaven head and face, like his other two followers in Fig. 3. The figure on the right of this fragment wears hair and beard, and probably represents a member of the opposite party conducting them into the presence of his master.

[43] See "Déc. en Chaldée," pl. 1 _bis_, Figs. 3-7.

[44] The fact that on seals of this later period the Moon-god is represented in the Sumerian mantle and headdress may well have been a result of the Sumerian reaction, which took place under the kings of Ur (see below, p. 283 f.).

[45] See below, p. 131, Fig. 46.

[46] See p. 49. In Fig. 14 the hair and beard of the god who leads the worshipper into the presence of the goddess is clearer on the original stone. In Fig. 15 the locks of hair and long beards of the seated gods are more sharply outlined; they form a striking contrast to the figures of Sumerians, who are represented as pouring out libations and bringing offerings to the shrine.

[47] See p. 50, Fig. 16.

[48] Cf. Langdon, "Babyloniaca," II., p. 142; this explanation is preferable to treating the crowns as a feathered form of headdress. The changes in the dress of the Sumerian gods, and in the treatment of their beards, appear to have taken place in the age of the later Semitic kings of Kish and the kings of Akkad, and may well have been due to their influence. The use of sandals was certainly introduced by the Semites of this period.

[49] See Clay, "The Amer. Journ. of Semit. Lang, and Lit.," XXIII., pp. 269 ff. In later periods the name was pronounced as Ellil.

[50] Cf. "Nachträge zur aegyptischen Chronologie," p. 44 f., and "Geschichte des Altertums," Bd. I., Hft. II., p. 407.

[51] See Thureau-Dangin, "Sum. und Akkad. Königsinschriften," p. 38, Col. I., l. 26; the word is _dam-kha-ra_, which he rightly takes as the equivalent of the Semitic _tamkhara_, "battle" (cf. also Ungnad, "Orient. Lit.-Zeit.," 1908, col. 63 f.).

[52] In this respect the early Sumerian texts are in striking contrast to those of the later periods; the evidence of strong Semitic influence in the latter formed the main argument on which M. Halévy and his followers relied to disprove the existence of the Sumerians.

[53] See Messerschmidt, "Orient. Lit.-Zeit.," 1905, col. 268 ff.; and cf. King, "Chronicles," I., p. 180, n. 3.

[54] See Ungnad, op. cit., 1908, col. 62 ff.

[55] See below, Chap. VII. f.

[56] See below, p. 141, Fig. 48.

[57] See above, p. 14, n. 1.

[58] See further, Appendix I.

[59] Cf. Langdon, "Babyloniaca," I., pp. 225 f., 230, 284 ff., II., p. 99 f. The grounds, upon which the suggestion has been put forward, consist of a comparison between the verb "to go" in Sumerian, Greek, and Latin, an apparent resemblance in a few other roots, the existence of compound verbs in Sumerian, and the like; but quite apart from questions of general probability, the "parallelisms" noted are scarcely numerous enough, or sufficiently close, to justify the inference drawn from them.