tion of phonography we ask attention to one or
two facts regarding Landa's alphabet which do not appear to have been
previously noticed, yet have an important hearing on the subject.

The failure to reach any satisfactory results with this alphabet proves,
beyond a reasonable doubt, that this author was mistaken as to the
character of the Maya writing; yet the frequent occurrence in the
manuscripts of most, if not all, of the elements he presents renders it
certain that there is a basis of truth on which it rests. It is probable,
therefore, if we can find the key to his method, we may, after all,
obtain some satisfactory results by means of his alphabet.

I have already stated as my belief that--

He has undertaken to pick out of their compound or syllabic
characters the letter elements; hence it is that, while we find it
impossible to decipher the manuscripts by using them, yet we find
such frequent resemblances as to compel us to admit a fundamental
relationship.[367-1]

This opinion I still believe to be correct, but was, until very recently,
unable to get any positive evidence as to his method of obtaining these
elements.

While examining the Cortesian Codes I came across (on Plate 17) the
symbol for a turtle (the different varieties of which are shown in
marginal figure No. 4), which is nothing more or less than an attempt to
represent the head of the animal. In the more abbreviated form (_b_) I at
once recognized Landa's A (compare with _c_ and _d_, No. 4). As the Maya
name of the turtle is _Ac_ or _Aac_ it is apparent that in this instance
the old Spanish priest selected a symbol representing an object the name
of which contains a single syllable having, as its chief letter element,
A. As this symbol is simply a representation of the animal's head there
is no reason to infer that it is phonetic; on the contrary, it is more
reasonable to assume that it was used only as a conventional sign. It is
possible that after long usage it may have been adopted as a phonetic
character, though its exceedingly rare occurrence in the manuscripts
(being found only in the Cortesian Codex and with the turtle figure) and
the fact that it is seldom, if ever, used as part of a compound character
would seem to forbid this idea.

Precisely the same method was adopted in obtaining his B, which is given
in two forms, first as a foot print and second as a circle inclosing four
circular dots. The first, as al