The inscriptions were in three different systems of assortment of the elemental signs, evidently representing three different languages, and as they were placed side by side, it was also evident that they were three versions of the same decree, or record of the same event. One of the versions, which always came first, was simpler than the others. This consisted of about forty signs, while the others were more complicated and numerous. Again, in this version the groups of signs, which evidently formed words, were separated, each from the other, by a slanting wedge which did not appear in the others. Grotefend also observed that each inscription usually began with a certain group of words. One of these words, on different inscriptions, varied, while the other words of this group remained the same. By a happy guess, he conceived these groups to be royal names and titles, the words which varied on the different inscriptions to be names of different kings, while the words which always continued the same in these groups were their titles. Upon this basis he began his work. It was known to scholars that certain Achæmenian princes—Darius and his successors—had erected some of the monuments from which copies of the inscriptions were taken. Turning then to the older Persian language, of the time of Darius, for the spelling of the name of this king, he gave alphabetic values to certain of these signs which he supposed might spell the name of Darius. Also, to the words which he supposed represented the titles of this king. These alphabetic values were based upon the spelling of the name and titles in the ancient Zend. In this way he obtained supposed values of six letters in the cuneiform. He then turned to another royal name which might be Xerxes. The name of Darius, in old Persian, or the Zend, is spelled: D-A-R-H-E-A-U-SCH. Again, the name of Xerxes, in Persian, is KH-SCH-H-E-R-E. Now, if the third sign in the spelling of the name of Darius was the same as the fifth sign in the spelling of the name Xerxes, in the Zend, this must have the phonetic value of _R_. The comparison proved the correctness of his conception. And again, further confirmation appeared in another royal name, Artaxerxes, where the latter part of the name was the same as the second royal name, and the sign for the second character again corresponded with the letter _R_. Thus he compared letter by letter, and sign by sign, until he had found agreement in signs and sound for the names of these kings and their titles.