the Semitic values being added to the Sumerian, the scribes soon found themselves in possession of a double set of syllables both simple and compound.  This multiplicity of sounds, this polyphonous character attached to their signs, became a cause of embarrassment even to them.  For instance, [symbol] when found in the body of a word, stood for the syllables hi or hat, mid, mit, til, ziz; as an ideogram it was used for a score of different concepts:  that of lord or master, inu, bilu; that of blood, damu; for a corpse, pagru, shalamtu; for the feeble or oppressed, kahtu, nagpu; as the hollow and the spring, nakbu; for the state of old age, labaru; of dying, matu; of killing, mitu; of opening, pitu; besides other meanings.  Several phonetic complements were added to it; it was preceded by ideograms which determined the sense in which it was to be read, but which, like the Egyptian determinatives, were not pronounced, and in this manner they succeeded in limiting the number of mistakes which it was possible to make.  With a final [symbol] it would always mean [symbol] bilu, the master, but with an initial [symbol] (thus [symbol]) it denoted the gods Bel or Ea; with [symbol]. which indicates a man [symbol], it would be the corpse, pagru and shalamtu; with [symbol] prefixed, it meant [symbol]—­mutanu, the plague or death and so on.  In spite of these restrictions and explanations, the obscurity of the meaning was so great, that in many cases the scribes ran the risk of being unable to make out certain words and understand certain passages; many of the values occurred but rarely, and remained unknown to those who did not take the trouble to make a careful study of the syllabary and its history.  It became necessary to draw up tables for their use, in which all the signs were classified and arranged, with their meanings and phonetic transcriptions.  These signs occupied one column, and in three or four corresponding columns would be found, first, the name assigned to it; secondly, the spelling, in syllables, of the phonetic values which the signs expressed, thirdly, the Sumerian and Assyrian words which they served to render, and sometimes glosses which completed the explanation.

[Illustration:  276.jpg Tables]

Even this is far from exhausting the matter.  Several of these dictionaries went back to a very early date, and tradition ascribes to Sargon of Agade the merit of having them drawn up or of having collected them in his palace.  The number of them naturally increased in the course of centuries; in the later times of the Assyrian empire they were so numerous as to form nearly one-fourth of the works in the library at Nineveh under Assurbanipal.  Other tablets contained dictionaries of archaic or obsolete terms, grammatical paradigms, extracts from laws or ancient hymns analyzed sentence by sentence and often word by word, interlinear glosses, collections of Sumerian formulas translated into Semitic speech—­a child’s guide, in fact, which the savants of those times consulted with as much advantage as those of our own day have done, and which must have saved them from many a blunder.