nd remembering the words of Edgar Allan Poe--the strange
genius to whom we are indebted for the first workable system of
deciphering cryptograms: "It may well be doubted whether human
ingenuity can construct an enigma of the kind which human ingenuity
may not, by proper application, resolve."

The first conclusion to which I was borne was this: that the letters
comprising the word "Zagazig" were designed merely to confuse the
reader, and might be neglected; since, occurring as they did in regular
sequence, they could possess no significance. I became quite excited
upon making the discovery that the _punctuation marks_ varied in
almost every case!

I immediately assumed that these constituted the cipher; and, seeking
for my key-letter, _e_ (that which most frequently occurs in the
English language), I found the sign of a full-stop to appear more
frequently than any other in the first message, namely ten times,
although it only occurred thrice in the second. Nevertheless, I was
hopeful ... until I discovered that in two cases it appeared three
times _in succession!_

There is no word in English, nor, so far as I am aware, in any language,
where this occurs, either in regard to _e_ or any other letter!

That unfortunate discovery seemed so wholly to destroy the very theory
upon which I relied, that I almost abandoned my investigation there
and then. Indeed, I doubt if I ever should have proceeded were it not
that by a piece of pure guesswork I blundered on to a clue.

I observed that certain letters, at irregularly occurring intervals,
were set in capital, and I divided up the message into corresponding
sections, in the hope that th capitals might indicate the
commencements of words. This accomplished, I set out upon a series
of guesses, basing these upon Smith's assurance that the death of the
dacoit afforded a clue to the first message and the note which he
(Smith) had pinned upon the door a clue to the second.

Such being my system--if I can honor my random attempts with the
title--I take little credit to myself for the fortunate result. In
short, I determined (although _e_ twice occurred where _r_ should have
been!) that the first message from the thirteenth letter, onwards to
the twenty-seventh (_id est:_ _I;_g:-zagAz;i-;_g_;_-Z_,-a;-_g_azi;-)
read:--

_"Three Colt Street."_

Endeavoring, now, to eliminate the _e_ where _r_ should appear, I made
another discovery. The presence of a letter in _italics_ alte