Nothing could at first be more bewildering than the immense number of signs. Grotefend counted only a hundred and thirty different characters in the third Persepolitan column. Mr. Fisher, in 1807, found that the East India House inscription contained two hundred and eighty-seven;[723] and Grotefend, in 1837, estimated that the whole of the Babylonian inscriptions known to him contained about three hundred different signs.[724] But Botta encountered no less than six hundred and forty-two at Khorsabad alone.[725] The unskilled eye will be disposed to agree with Löwenstern that at first sight the Ninevite character presented no analogy with the Persepolitan, or even with the characters on the Babylonian bricks. In his ‘Essai de Déchiffrement’ (1845) he was, however, the first to point out that a large number of them do really correspond to the third Persepolitan; and he based his attempted interpretation partly upon the analogy he had discovered.[726] In his first Essay, of 1846, Hincks also stated his belief that ‘the third Persepolitan agrees in character with the Babylonian and with the Assyrian writing in Schulz’s inscriptions.’[727] Löwenstern afterwards admitted that he would scarcely have recognised the similarity from Schulz’s plates; but it became clearly apparent in the more perfect drawing of Texier.[728] Indeed he was eventually so much struck by the resemblance that he hesitated to classify the Persepolitan with the Babylonian in preference to the Assyrian; in fact, he ultimately persuaded himself that it was nearer the Assyrian. Meanwhile Botta had begun the minute study of the Assyrian character to which later investigators owe more than they are always willing to acknowledge.[729] Those whose fortune it is to occupy the higher pinnacles of knowledge are only too prone to despise the humbler artificers who constructed the scaffolding that enabled them to achieve the ascent. Botta arranged the signs with great care into fifteen classes, according to the number of wedges they contain. The first class included all those with one wedge only; and so on up to the fifteenth class, where we find signs composed of fifteen wedges and upwards: though none appear to exceed eighteen. The result of his classification was to persuade him that the graphic system of Assyria was substantially the same as that found at Persepolis and Babylon. He accepted Rich and Westergaard as the most faithful copyists of the former, and he compared ninety-six of their signs with those at Khorsabad. He found that seventy-two were so similar that their identity could not fail to be recognised at first sight. Fourteen others exhibit a greater difference, but their identity is capable of demonstration. There are therefore eighty-six signs out of ninety-six concerning which no doubt can exist. He thought the difference was not so great as between Gothic and Latin characters. With respect to the writing at Van, he counted a hundred and twelve to a hundred and fifteen