We must examine other cases. This is to be noticed: If the reading is in horizontal lines from left to right, then the progress is from top to bottom in columns, as the case of 3035 and 3040 shows. This occurs at the end of a line, and the corresponding _chiffre_ required to make the pair is at the other end of the next line. I have marked this case with asterisks. If we must read in the lines from right to left we must necessarily read in columns from bottom to top. Thus the _lines_ are connected. A similar process with all the other tablets in STEPHENS leads to the conclusion that the reading is in lines horizontally and in columns vertically. The cases 1835-'45, 1885-'95, 1914-'24, and 1936-'46 should, however, be examined. We have now to decide at which end of the lines to begin. The reasons given by Mr. BANCROFT (_Native Races_, vol. ii, p. 782) appeared to me sufficient to decide the question before I was acquainted with his statement of them. Therefore, the sum total of our present data, examined by a rational method, leads to the conclusion, so far as we can know from these data, that the verbal sense proceeded in _lines_ from left to right, in _columns_ from top to bottom; just as the present page is written, in fact. For the present, the introduction of the method here indicated is the important step. It has, as yet, been applied only to the plates of STEPHENS' work. The definite conclusion should be made to rest on _all possible_ data, some of which is not at my disposition at present. Tablets exist in great numbers at other points besides Palenque, and for the final conclusion these must also be consulted. If each one is examined in the way I have indicated, it will yield a certain answer. The direction of reading for that plate can be thus determined. At Palenque the progress is in the order I have indicated. V. THE CARD-CATALOGUE OF HIEROGLYPHS. It has already been explained how a system of nomenclature was gradually formed. As I have said, this is not perfect, but it is sufficiently simple and full for the purpose. By it, every plate in STEPHENS' work receives a number and every hieroglyph in each plate is likewise numbered. This was first done in my private copy of the work. I then procured another copy and duplicated these numbers both for plates and single _chiffres_. The plates of this copy were then cut up into single hieroglyphs and each single hieroglyph was mounted on a library card, as follows: