not. He suspected that the word which changes would be the king's name—as different inscriptions would relate to different kings—and that the other words gave the king's titles. He felt convinced that a word which was constantly repeated signified "king," and conjectured that when two kings were mentioned they were probably father and son. Finding that the names of Cyrus and Cambyses would not suit, because no two names in the inscription he was dealing with commenced with the same letter, he tried others. Cyrus and Artaxerxes seemed equally inapplicable, because of their unequal length, the two names he was dealing with being of six letters each. The only names remaining were those of Darius and Xerxes; and these on further comparison appeared to agree so exactly with the characters that he did not hesitate at once to adopt them. Having thus found out more than twelve letters, among which were precisely those composing the royal title, the next business was to give these names their original Persian form, in order that by ascertaining the correct value of each character, the royal title might be deciphered. From the "Zeudavesta" of Anquetil du Perron, M. Grotefend found that the Greek form Hystaspes was originally represented in Persian by Gustasp, Kishtasp, or Wistasp. The first seven letters of this name were at once discovered, while a comparison of all the royal titles led him to the conclusion that the three last formed the inflection of the genitive singular, corresponding to the Latin Hystaspis. Thus did Grotefend proceed step by step, his ingenuity and perseverance being beyond all praise. Meantime Sir Henry Rawlinson, although stationed in Persia and cut off in a great degree from the results of European scholarship, was devoting himself with ardour to the study of the Behistun inscription, and making independent progress.
It turned out that of the three forms of arrow-headed
