(4) He also copied the large inscriptions in four tablets on the outside south wall of the platform: H, I, K, and L. They each fill twenty-four lines, and an idea of the size of the letters may be formed from the fact that the original covers a space twenty-six feet long by six feet high. Of these Le Bruyn had already copied the A inscription in the drawing he gives of the stairs, the six-line inscription of Darius (B, C, D), and a line of the inscription of Xerxes from the royal robe; while Kaempfer copied one tablet (L) of the inscription on the south terrace.[89] The others now appear for the first time. It has already been observed that Kaempfer was the first to draw the characters with a bold and steady hand; in this he has been followed by Niebuhr, so that his copies differ in no respect from those produced in the present day. The few remarks he has made upon the subject are of peculiar value and very materially assisted later scholars. He was the first to observe that the inscriptions are written in three different ‘alphabets,’ and that these always recurred together.[90] So slow is the progress of discovery, however, that he never seems to have advanced to what might appear to be the obvious conclusion that the three tablets are repetitions of the same text in different languages. He noted that the ‘alphabet’ in one of the tablets of the series was comparatively simple, and consisted of no more than forty-two different signs.[91] These he copied out, and they appear in Plate 23. He thus limited the first step in decipherment to the interpretation of a comparatively small number of signs. Till then the greatest confusion was produced by the appearance of detached portions of inscriptions, selected indifferently from all the three kinds of writing, a process that inspired the fear that the number of signs to be mastered was practically unlimited. Having clearly detected each separate letter from among the number of confused signs in a line of inscription, he farther assisted the student by marking off each separate letter by a full stop or colon in the copies he made of the inscriptions themselves.[92] From the division of the signs into letters he does not seem to have made the next step and apprehended the division of the lines into words by the diagonal wedges in the Persian column. He, however, directed attention to two different copies of the same inscription, where in one the letters that end the third line are in the other the first that occur in the fourth line; he pointed out that this practically settled the direction in which the writing should be read.[93] He also showed that the lines supposed by Chardin to be written from top to bottom are not in reality upright, but should be placed on their side, and when horizontal the letters correspond to those already known.[94]