her work the
words of which come in alphabetical rotation. It is probably some
ordinary book, which the writer of the cryptogram and the person
for whom it is written have agreed upon beforehand to make use of
as a key. I have no means of judging whether the book in question
is an English or a foreign one, but by it alone, whatever it may
be, can the cryptogram be read.

"Now, my dear Ducie, it would be wearisome for me to describe, and
equally wearisome for you to read, the processes of reasoning by
means of which the above deductions have been arrived at. But in
order to satisfy you that my assumptions are not entirely fanciful
or destitute of sober sense, I will describe to you, as briefly as
may be, the process by means of which I have come to the conclusion
that the book used as the basis of the cryptogram was not a
dictionary or other work in which the words come in alphabetical
rotation; and such a conclusion is very easy of proof.

"In a document so lengthy as the MS. of your friend the Scotch
laird there must of necessity be many repetitions of what may be
called 'indispensable words'--words one or more of which are used
in the composition of almost every long sentence. I allude to such
words as _a_, _an_, _and_, _as_, _of_, _by_, _the_, _their_,
_them_, _these_, _they_, _you_, _I_, _it_, etc. The first thing to
do was to analyse the MS. and classify the different groups of
figures for the purpose of ascertaining the number of repetitions
of any one group. My analysis showed me that these repetitions were
surprisingly few. Forty groups were repeated twice, fifteen three
times, and nine groups four times. Now, according to my
calculation, the MS. contains one thousand two hundred and
eighty-three words. Out of those one thousand two hundred and
eighty-three words there must have been more than the number of
repetitions shown by my analysis, and not of one only, but of
several of what I have called 'indispensable words.' Had a
dictionary been made use of by the writer of the MS. all such
repetitions would have been referred to one particular page, and to
one particular line of that page: that is to say, in every case
where a word repeated itself in the MS. the same group of numbers
would in every case have been its _valeur_. As the rep