100% found this document useful (3 votes)
341 views208 pages

Cryptography

This passage describes cryptography as the art of secret writing to hide messages from others. It provides examples of simple cryptography used by children and discusses more complex historical ciphers. It then recounts the author's experience deciphering a mysterious coded document from the 1800s. By analyzing letter frequencies and common letter combinations in French, the author was able to deduce that the most common symbols represented letters like e and u. Making hypotheses about possible words, he slowly uncovered that the coded message contained the French words "Ce que" and "heureux," revealing there was no conspiracy but rather an expression of feelings. The passage illustrates how statistical analysis of a language can reveal the underlying code or cipher used.

Uploaded by

Tapuwa Chizunza
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
100% found this document useful (3 votes)
341 views208 pages

Cryptography

This passage describes cryptography as the art of secret writing to hide messages from others. It provides examples of simple cryptography used by children and discusses more complex historical ciphers. It then recounts the author's experience deciphering a mysterious coded document from the 1800s. By analyzing letter frequencies and common letter combinations in French, the author was able to deduce that the most common symbols represented letters like e and u. Making hypotheses about possible words, he slowly uncovered that the coded message contained the French words "Ce que" and "heureux," revealing there was no conspiracy but rather an expression of feelings. The passage illustrates how statistical analysis of a language can reveal the underlying code or cipher used.

Uploaded by

Tapuwa Chizunza
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 208

MurmfflffimffMtmmffflHiB^

v<r<.if iir'wxj''

ifv

f>

-~*

*
i

CRYPTOGRAPHY

CRYPTOGRAPHY
BY

ANDRE LANGIE
TRANSLATED FROM THE FRENCH BY

MACBETH

C. H.
J.
AUTHOR OF "THE MARCONI CODE," "MARCONI DICTIONARY,"

CONSTABLE
LONDON

&7

ETC.

COMPANY LIMITED
BOMBAY
1922

SYDNEY

PRINTED IN GREAT BRITAIN

PREFACE
I

HAVE no intention

cryptography.
attainable

in

Finality,

the

even

domain

of

manuals are in existence on


of

some

of

them

will

complete manual of
very relative, is not

of writing a

this

art.

this subject,

Besides,

good

and the

titles

be found at the end of this volume.

A cryptographer of considerable experience, however,


can always add a few details even to the most complete
works of this kind.

My

object in writing this

book

is

simply to explain

what cryptography is, and to recall what it has been


from antiquity to the present day; in short, to relate my
experiences as a decipherer.

The

first

part of the volume contains a description of

the principal systems of cryptography, together with a


note on the rule played by cryptography in history.

In the second part I relate how I succeeded in decipherIn some


ing a dozen cryptograms of various kinds.
chapters of this section I give the texts just as they came
into my hands; but in the majority of cases, though
preserving the system of cryptography actually employed,

have, on grounds of expediency, substituted an approximate reading for the actual text, and have modified the
I

plan,

way

and

and even radical features


as to render abortive

localisation.

any

of the narrative, in

such

attempt at identification

PREFACE

vi

In the third part

give

some advice

in a general

way

which have proved profitable to me, and, further,


a certain number of tables and formulae; but while I

on

lines

recognise these to be very useful, too

much

reliance should

not be placed on them, under penalty of striking the


wrong track, as I shall have occasion to repeat farther on.
In point of fact, as I have found by experience, in
cryptography the exceptions are infinitely more frequent
than the rule.

TRANSLATOR'S PREFACE
I

have to acknowledge with grateful thanks the valuable


have received in preparing this work from

assistance I

the late Mr.


of

W.

Washington,

W. W. Smith

Jarvis, Lieut. -Commander

U.S.A., Mr. Albert

M.

Smoot

of the

Ledoux Laboratories, New York, and Miss A. Wishart


of the Radio Corporation of America.
It was not an
easy task substituting English text for the examples of
ciphers in French, and if there are any errors which have
inadvertently escaped detection I

humbly beg
J.

C.

H.

forgiveness.

MACBETH.

CONTENTS
PART

PAGE

DESCRIPTION OF THE PRINCIPAL SYSTEMS OF CRYPT-

OGRAPHY, WITH HISTORICAL NOTICE

PART

II

EXAMPLES OF DECIPHERING

PART
LISTS

47

122

III

AND TABLES

BIBLIOGRAPHY

158

PART IV
THE PLAYFAIR CIPHER SYSTEM,

Vll

ETC.

159

ACKNOWLEDGMENT
OUR thanks

are due to the following gentlemen

in connection with translating the


original French,

book from the

working out and substituting


for the French ones,
'

Examples

English

adding additional matter, and seeing the work

through the Press

H. Macbeth.

Mr.

J. C.

The

late Mr.

W.

J. Jarvis.

Mr. H. G. Telling.
Commander Smith, U.S.N.

Paymaster- Commander
Brown, C.B.E., Pt.N.

J.

E. A.

THE MARCONI INTERNATIONAL CODE

CO.,

MARCONI HOUSE, STRAND,


LONDON, W.C.
2~>th

July, 1922.

LTD.
2.

CRYPTOGRAPHY
PART

OF THE PRINCIPAL SYSTEMS OF


CRYPTOGRAPHY, WITH HISTORICAL NOTICE

DESCRIPTION

I.

EVERYONE
2

graphy,

Who

has,

at

some time

or

other,

used crypto-

or secret ciphers.

has not had occasion to

make some

note, or to

correspond with somebody, by dotting letters in a newspaper or book ? Even children amuse themselves in
this

way on

their school desks.

not necessarily required to

make

But a pen or pencil

is

use of a secret language.

More than one of us, in our young days, have been


embarrassed by a question from the schoolmaster. We
have been required to give a proper name in answer,
but

it is

precisely this proper

name which has

slipped our

memory. So we have glanced at a comrade with whom


we had previously come to an understanding. And the
latter passes a finger over his
ear,

We

ftair,

his ear, his Zips, his

and his nose, whereby we understand " Helen."


have thus corresponded by means of mimetic

cryptography.

What is cryptography, after all ? Cryptography is


the art of recording one's thoughts in such a way as to
1

This Part

I.

appeared in the Bibliotheque universette

and October, 1917.


From the Greek words Kpvn-Tos, secret; and ypdfaiv, to

et

revue

suisse in August, September,


2

write.
1

CRYPTOGBAPHY

make them unreadable


it

over,

to

Particularly, more-

others.

enables two persons to correspond under cover

at least in theory.
A man will
his
a
on
own
account,
invent,
perhaps
system of writing
can
write
and
means
of
which
he
preserve secrets
by
of

complete secrecy

which he prefers not to divulge, while ensuring the


possibility of reading them again at any time.
The great thinker, Alexandre Vinet, composed a system
of cryptography which was as simple as he was nobleminded. He used it to note in his diary his qualms and

The phrases he wrote in this way can be read


almost at a glance.
The celebrated Swiss physiognomist, Jean-Gaspard
Lavater, in his Diary pj a Self-Ob server, constructed

trials.

several systems of secret writing to preserve his private


reminiscences.
These passages, which are omitted from

the

new French

translation, are far

more

difficult

to

read than those of Vinet.


his

countrymen

had

Eight years after his death


not succeeded in deciphering

them all.
Some years ago I was asked by a friend, a professor at
a university in German Switzerland, to decipher a piece
of yellow paper, covered with strange characters, found
among the records of a Swiss politician, a contemporary
of

and which was supposed to have some


Here is a specimen, a part of the
and one word of the sixth

Napoleon

I.,

historical importance.
first line

PEINCIPAL SYSTEMS OF CKYPTOGKAPHY

friend had consulted his colleagues one had declared


was not Sanscrit, another that it was not Ethiopic,
and still others that it was neither Slavonic nor Kunic.
These professors spoke truly, for it was French
This system was one of the easiest to decipher.
There
were some 800 signs in the text. One of the signs, the
second in the above example, and the most frequent,

My

it

occurred something over ninety times, while another, the


fourth, occurred seventy times.

Now

it is

well

known

and most languages

that in English, French, German,


of Western Europe, the most fre-

quently occurring letter


in French,

is

e;

the letter which follows

is,

according to the writer; in German, n;


in Italian, i\ and in Spanish, a.
In Eussian

or

s,

in English, t;
the most frequently occurring letter is o, but i if the
language is written in Koman characters. In Polish the

most frequent consonant is z; not uncommonly three


may be found in the same word. In Arabic and Turkish
the letter
elif, corresponding to the French stopped
j,
In Chinese

or silent h, occurs oftenest.

newspapers
are

and

(chi,

oftenest

at least, in the

the characters found in order of frequency


of,

To

(kong, work).
in any language,

not, negative),
(pull,
ascertain which letters occur

genitive),

one

must

'

calculate

fre-

quencies."

The next thing

commonly

to do

is

to study
are es in

which letters most


French and en in

adjoin.
They
The most frequent groups of three are ent in
French, the in English, ein in German, die in Italian, etc.
Bulky works have been written on this subject containing
long lists, more or less complete, of the various articulations and disarticulations of words.
Of course, this
an
of
amount
enormous
labour, involving a
requires

German.

CRYPTOGRAPHY

4
statistical

study of texts containing

many thousands

of

letters.

To

revert to our example, I encountered a difficulty at


The sign which came second in order
the first onset.
1
frequency, and which I supposed should represent

of

either

At

or

caused me some embarrassment.


abandoned the books I had been using, and

s,

last I

began a new calculation of the frequency of letters in


and French newspapers. In the letters
of Voltaire I noted that the letter u occupied the second
certain authors

place in point of numbers, this being obviously due to the


"
"
and
fact that the words nous and vous (" we
you ")
are common in the epistolary and conversational style.

In the sixth line of the document, a group of signs


offered the peculiarity of conjoining twice in succession
the two most frequent characters, the supposed e and
the supposed u, separated by another sign and followed

by one occurring rather rarely. Accordingly a new trial


was made, which this time proved satisfactory. These
signs might imply the tail of the word valeureux or the
words peureux or heureux. This last proved to be correct.

Among

the

first

signs of our example, the supposed e

occurs preceded by the supposed u. In French, u followed by e occurs principally in the syllable que. It
could not be the word lequel here, the sixth sign not being
first.
The group must read: Ce que. A
on we meet again with the sign representing c, followed by the r of the word heureux and preceded
by e, a group of letters which might, for instance, form
the words ecran, decret, or, better still, ecrire.

similar to the

little

farther

The decipherment

is

based not only on

statistics,

but also on

"
hypotheses. In fact, the famous expression,
Suppose that
of
motto
the
is the
cryptographer

.,"

PKINCIPAL SYSTEMS OF CRYPTOGKAPHY


The

result of the deciphering showed that there was


of a conspiracy in this mystical writing, but

no question

of the enthusiastic sentiments inspired in the author


some charming person met at a fashionable party.

What

was

as follows:

am writing you
I am writing to

heart, since

here

is

merely to relieve

the dearest object in

to divert the frightful restlessness of


And so on for twenty-five lines.

my

days.

They produce heroes who mark

my

my
life

..."

Cryptography has provided an entertaining


novelists.

It

The first phrase,

was, perhaps, the rough draft of a letter.


translated,

by

field

in

for

secret

writing the route to be followed in order to recover a


fabulous treasure or to track the author of a crime, or
perhaps learned men who reveal some stupefying discovery.

We

have all read the story of the Gold Beetle, by


the American novelist, Edgar Allan Poe. It will be
remembered that, in order to recover the wealth buried
by Kidd, the pirate, it was necessary to let the gold
beetle fall from the left orbit of a skull attached to the
highest branch of a big tree, and to extend by fifty steps
a line leading from the foot of the tree and passing
through the point where the beetle fell. A hole was
dug at the spot reached/ and, of course, an incalculable
treasure unearthed.

Who

has

not

read,

also,

Jules

Verne's

Jangada

And who has failed to be interested in the researches


made by the Judge Jarriquez into a Portuguese document
in secret writing in order to save the
victim condemned to death ?

In

Voyage into

Jules Verne,

we

see a

life

of

an innocent

Interior of the Earth, also by


Danish scholar intent on piercing

the

CEYPTOGKAPHY

the mystery of a cryptographic parchment which

to

is

reveal the route to be followed in order to penetrate into


the depths of our terrestrial globe. But old Professor

Liden brock seeks too

far,

enough by reading the


It

and

young man, who

careless

it

is

attains

his

nephew Axel,

the

finals of the lines

object simply
backwards.

be pointed out that the system deciphered by

may

is
comparatively simple. He himself
1
acknowledges this, and claims to have deciphered keys
ten thousand times
more arduous. The system

Edgar Allan Poe


'

'

conceived by

Jules

Interior of the Earth

Verne in
is

his

Voyage into

the

As to that in
thanks to an incredible

also very easy.

Jangada, the problem is solved,


combination of favourable circumstances.

In one of the latest novels of the Arsene Lupin series,


The Hollow Needle, Maurice Leblanc has the idea of

uniformly substituting the consonants of a document by


dots.
Nothing can be said of this system, except that
it is

ultra-fantastical.

example, and this time historical: the poet


Philippe Desportes wrote in cipher the life of Henri III.,
King of France. If this work had come down to us and
final

been deciphered, probably not

many

would have been found therein.

But

edifying subjects

was burnt during


the troubles of the Holy League. 2
Some months ago I received a letter from a foreigner,
who informed me that he was very interested in cryptoit

graphy, and that he wanted to work on some official


texts. He begged me to lend him some diplomatic

documents, of which he would take copies for his use


The

'

'

'

'

in cryptography is the formula which enables a


text in cipher to be read.
Henri Martin, Histoire de France, vol. ix., p. 472, note.

key

PKINCIPAL SYSTEMS OF CKYPTOGKAPHY


and return me the
he wrote,

'

originals.

You do
me

but you can rely on

not

know me,"

entirely:

am

neutral."

Despite this excellent recommendation, I had nothing


However, touched by his candour, I
gave him some advice. Living in a large town, he had
to communicate.

at hand a mine of small cryptograms: he had only to


look in the windows of the curio dealers and antiquaries,
take note of a number of prices marked in secret

and try to decipher them. It is a cryptographic exercise as good as any other. There are certain
methods which enable one to guess which letter means 5,
which 0, which 9, etc. 1 I refrained from pointing them
characters,

out to him, since the value of these exercises lies precisely


I wonder
finding out these methods for oneself.

in

whether he followed
good.

my

advice, which I consider

was

We

were just now recalling some specimens of secret


writing where the key was in the hands of only one person.
Let us now consider another order of cryptography,
which enables two persons to correspond under shelter
of secrecy.

We

will leave aside the various

sympathetic
which affords no security,
whether used on paper, or, as has often been seen in
the course of the present war, on the skin that of the
arms or back since a simple chemical reaction exposes
inks,

the

employment

them immediately.

of

Conventional or shuffled alphabets


alone are of any use.
An example of a cryptography widely in vogue in the
1
NOTE BY TRANSLATOR. This is for decimal currency. In
s. d., the same methods would first disclose
ciphers representing
6, 1, and 0.

CRYPTOGRAPHY

Middle Ages is furnished by the so-called alphabet of the


Freemasons, of which the following is a specimen:

PKINCIPAL SYSTEMS OP CKYPTOGBAPHY


;

An

inquisitive intermediary
if

document

the

would see nothing


were

in

whereas

brief,

9
it,

the

especially
recipient will place the message received on to a graph
similar to that of the sender, and will have no difficulty
in deciphering:

RFC

'

I love

He
means

OQXDERKYJ LSZVMTAUGNBHW

you dearly."

(or she) will reply in the

of a thread.

beginning on the
letter required,

This

same way,

or perhaps

by

laid along the cipher alphabet,

is

and wherever the thread passes a


marked in ink. Arrived at the right

left,

it is

extreme of the alphabet, the thread


and a new start is made from the

is

moved a

section,

and so on indefinitely until all the letters required have been marked.
would require
I reciprocate
Thus, to write the words
'

left,

'

'

something over seven sections of the thread, each corresponding to the length of the alphabet.

This

method

1
It is needless to say that if the zigzag contains fifty angles and
the thread bears fifty marks, a decipherer could discover the key of
both.

10

CEYPTOGEAPHY

of corresponding is very ancient.


It is some 2,800 years
since .Eneas, the tactician, recommended a similar system

in his Poliorcetes. 1
is not the same thing as
cryptography, of
not a secret writing, seeing that hundreds of
thousands of persons can make use of it. Nevertheless,

Shorthand
it is

course;

an

artful

mind can combine shorthand and cryptography

in such a

writing.

way as to form a fairly complicated secret


In 1913 I was handed several dozen pieces of

paper which had been seized at a penal establishment in


French Switzerland. They were covered with shorthandlike characters

which had resisted the efforts of several


Here is a fragment:

professional shorthand writers.

t
was the correspondence of two bandits, authors of
robberies on a high scale, who were interned at the two
extremes of the prison. They had transmitted their
missives by means of a very well organised postal service.
Their letter-box was purely and simply the backs of the
volumes lent them by the prison library. They had
It

agreed in advance as to what books they would borrow,


and each found the letter of the other by opening the
book wide, which allowed the little piece of paper concealed in the hollow space in the back of the binding to
fall

out.

They had other hiding-places


Their
2

signs.
1

all

ready in case of alarm.

more than 200


alphabet
I will dwell a moment on the contents
consisted

of

different

of these

Chapter xxxi.

The method of deciphering applied here was to calculate the


frequency of the various lines and curves.
2

PEINCIPAL SYSTEMS OF CBYPTOGKAPHY

11

of
messages in order to show the usefulness of a safe form

cryptography.
These two bandits, M. and

schemed plans

of escape,

S., had drawn up welland were on the eve of carrying

them out when the unforeseen contingency of the deciphering occurred.


They had organised their future movements, and projected the burglary of a jeweller's shop
in an important Swiss town in order to get on their feet
before proceeding to effect a gigantic coup at a jeweller's
Kue de la Paix, Paris, or in Eegent Street, London.
"
"
but never
Take the small cases," wrote M. to S.,

in the

jewels displayed on velvet trays, for jewellers have a


trained eye, and can at a glance detect whether any piece
is missing from a
The large dealers always have
tray.

an assistant concealed in a corner, whose duty it is to


keep his eyes fixed on a mirror in the ceiling, enabling
him to watch most of the shop without the knowledge
of the customer.
When you go into the shop, find out,
without attracting attention, where this mirror is situated,
and operate outside its radius of reflection."

Some
'

further advice followed

should work in

first-class railway-carriages

operate

on solitary individuals, but never with a dagger, you


understand nor with revolver or chloroform. Hypnotism
at all times and everywhere.
So lose no time in taking
lessons in hypnotism as soon as you have left this en;

chanting resort."
M., who had a taste for mental pursuits and was well
read, mingled the practical advice which we have just
read with philosophical considerations on perfect friendship, on Schopenhauer and Nietzsche, on the destiny of
'

Ask
the soul, etc. Occasionally there is a postscript:
for
our worthy chaplain to borrow my fountain-pen
you.

CRYPTOGRAPHY

12
I will conceal a

watch-spring in

it,

toothed like a saw,

and you can begin on filing the bars of your cell, for we
shall have to be out by the end of the month."
He
Not one out of a hundred shorthand
further writes
experts in Berlin not one, I repeat would be capable
of reading my system Sto.
So it is still more likely to
remain a sealed book in French Switzerland."
The example given above in facsimile means
Be on
the alert
The pincers will be put behind the window;

'

sill

'

Its actual reading is:

this afternoon."

Zange wird nachmittags

am

Achtung

Fenstersims hinterlegen."

For

this correspondence took place in German.


I have
chosen this phrase from a sample which begins with a
succession of oaths of no particular interest for us.
/

II.

As we have

is of service to private
to certain private individuals
but
in
a
lies
of
corusefulness
means
furnishing

individuals
its

main

seen,

that

cryptography

is.

respondence between heads of States, Ministers, and


Generals.
In wartime, especially, by its aid plans of

and

action
relief

asked

diplomatic

can be communicated,
Cryptography, when employed for
military purposes, is termed
cipher,"

secret

information

for, etc.

or

be in the form of ciphers or figures,

whether

it

or

other signs.

any

prevail,

only

Government

is

letters,

conditions

and

departments

are in a position to utilise


of incalculable value to them.

authorities

which

Obviously, when war

military
1

cryptography,

The author should have said " legitimately."

It is a

matter

common knowledge

that numberless attempts were made by


spies to convey information to the enemy by means of more or less
ingenious ciphers. In most cases these attempts were foiled by the
of

ingenuity of the expert staff of cryptographers employed in the


various Cipher Departments.
TRANSLATOR.

PEINCIPAL SYSTEMS OF -CKYPTOGRAPHY

13

III.

The
quity.

origin of secret writing is lost in the mists of antiTo go back only to 500 years before the Christian

era, we find this record: "When Xerxes planned to


invade Greece, a Greek named Demaratus, a refugee at
the Court of the King at Susa, warned his countrymen

Lacedaemon by means of a message traced on wooden


covered with wax. At first nothing could be
seen on them, and it was Gorgo, the wife of Leonidas the
of

tablets

who discovered
ginians made use of a
King,

indicate

the

the

stratagem."

The Cartha-

similar process, which seems to


2
Heroemployment of sympathetic ink.

dotus 3 has recorded for us a not very practical system


which was once employed in the East. " Histia3us,
tyrant of Susa," he

tells us,

"

wishing to communicate

to Aristagoras, his lieutenant at Miletus, the order to


revolt, could find only one way, all the roads being

guarded. He had the head of his most trustworthy


servant shaved, made some incisions in the scalp, and

waited

till the hair grew


again."
(The era of the tele'As soon as this occurred,
graph had not yet arrived !)
he sent the man to Miletus without giving him any further

instruction than, on his arrival, to invite Aristagoras to


shave his head and scrutinise it. Now, the incisions

formed the word

'

Revolt

'

(ATroarao-^)."
This rather slow means of correspondence was not in
current use.
At the same period the Spartans had a far
better system of cryptography,

the scytales, of which

Herodotus, VII. 239.

Aulu-Gelle, Nuits atliques, XVII.


V. 35.

9.

CBYPTOGKAPHY

14
Plutarch,

among

scytale was a

others, has left us a description.

The

cylindrical rod round which the sender of

the secret message rolled a long band of papyrus in a


spiral, after the fashion of the emblems which cover reed-

On the wrapper thus formed he traced the


pipes.
words lengthwise along the rod, taking care to write only
one letter at a time on each fold of the ribbon of papyrus.
Once unrolled, this showed nothing but a meaningless
succession of separate letters.
The recipient rolled the
band round a rod of the same length and diameter as
that

of

the

sender.

The

diameter of the two rods

slightest

made

difference

in

the

the reading of the message

practically impossible.
To give an idea of the difficulty involved in deciphering
these scytales without having the proper rod, or with a

cylinder of a size dissimilar to that of the sender, it may


be stated that twenty letters can be combined in 2,500
billions of different ways.
A decodist who applied
himself to discovering the meaning of a document thus
transposed, and was so expeditious as not to devote more

than one second to the scrutiny of each combination,


would reach the trial of the final arrangement of these
If chance
characters at the end of 75,000,000,000 years.
favoured him, he might hit upon the solution at the
thousand and first or ten thousand and first trial, or it
might happen that he would have to persevere to nearly
still, he might encounter the solution
without knowing it and stopping there.
Nowadays, however, there is a process which enables
us to decipher these ribbons of papyrus comparatively

the end, or, worse

easily,

even without being in possession


1

Life of Lysander, ch. xix.

of the desired

PEINCIPAL SYSTEMS OF CEYPTOGKAPHY

Let us suppose that one of these messages has

cylinder.

and that

fallen into our hands,


of age

ness.
shall

15

have

left it

preserved in

its

twenty-five centuries

its original

state of fresh-

We

begin by making an exact copy, which we


manipulate in our own way, bearing in mind always

to leave the originals intact.


From one of the ends of
this copy let us cut off, say, three fragments, each containing ten or a dozen letters, or more or less if we like.

We place these segments one beside the other in the


order in which we have cut them.
This done, we slide
the second along the first, either up or down, and the
third along the second, endeavouring to form possible
syllables or fragments thereof.
(Assume, for convenience,
the document to be in English.) Let us suppose that

adjustments our attention


following combination
We observe that of the two groups

after various

fixed on the

is

of three letters,

WIL

forming a part

of

To

wild.

test

this,

is

the word

we

capable of
or

ivill

refer to

the

original scroll to count the intervals

between the three letters in the


group, and find that I is the eleventh
letter after W, and L the eleventh
after I.
It now becomes obvious
that

if

the eleventh letter after

another

or

D,

we

are

is

on the

The trial proves this to


be the case by yielding L. We now
make a new copy of the papyrus

right track.

into segments of eleven letters, which we


place one by one to the right, the result being that the
document becomes an open page to us, thus

and cut

it

CKYPTOGKAPHY

16

PEINCIPAL SYSTEMS OF CRYPTOGRAPHY


etc.

The Governments

Italian republics

made

17

of Venice, Florence, and other


use of secret writing from the

thirteenth century.
Since the Middle Ages numerous investigators have
pondered over an ideal system of cryptography. Among

them we may mention Francis Bacon, the

philosopher,

and Blaise de Vigenere, the French diplomatist, whose


ingenious table is still useful to-day, either for coding
or decoding.
Cardinal Eichelieu, the great statesman,
resorted
to cryptography.
Louis XIV. usedfrequently
so

complicated

Ministers

cipher

for

corresponding

when they were absent from

with

Versailles, or

his

when

he himself was with the army, that it was not until


175 years after his death that the key was discovered.
Let us here pause in this historical survey to examine

more closely the part played by ciphers. Nowadays all


the Great Powers have a Cipher Department.
There is one
in London, and others at Paris, Eome, Petrograd, Berlin,
Vienna, and elsewhere. When the head of a State and
his Minister of Foreign Affairs leave the country, they
are always accompanied by a staff of experts from the
Cipher Department.

M. Poincare, during

his last

journey

to Russia, a few days before the German aggression,


had with him the Director of the French Cipher Depart-

ment, with whose collaboration he was able to keep in


touch with Paris without running the risk of indiscreet
confidences.

Germany has

a department, the Chiffrierburd, staffed

by professional experts, whose mission is to find new


ciphers, both complicated and safe, and to decipher the
The newspapers insecret documents of the enemy.
1

G. Selenus, Cryptomenice, p. 282 (Alcuin).

CKYPTOGKAPHY

18

formed us that in February, 1916, the Department at


Vienna employed twenty-six cryptographers.
Cryptography," said one of the most genial of Swiss
'

Army commanders to me the other day,


science.
You must be a German, wear
and a

"

is

German

gold spectacles

bushy beard, before one can properly study

cryptography."

Not so long ago, however, when neither Berlin nor


Vienna were capable of deciphering difficult cryptograms,
they were glad, on occasion, and in secret, to have recourse to one of those little States which they so utterly
1

despise.

Each

step in the progress of cryptography is accompanied by a corresponding step in the art of deciphering.
History has preserved the names of some celebrated
decodists.
Thus, the geometrician Franois Viete succeeded in deciphering for Henry IV. a very complicated

system, formed of some five hundred signs, which was


used by the heads of the Holy League and the Spaniards. 2
The latter, angrily denounced Viete to the Holy See as a

wizard and a necromancer.

According to them, he could


have
into
entered
only
possession of the secret by calling

who had known the cipher during


But the Pope was a man of humour
he submitted the plaint to examination by a commission
"
of
with urgent recommendation." The
Cardinals,
Cardinals understood the hint, and the examination is
up the

spirits of

those

their earthly career.

still

unfinished.

1
See the Ziircher Post,, February 28, 1916, midday edition, and
the Bund, February 29, 1916, Sup. No. 100. The military Court
at Zurich, after seeming to hesitate subjectively over this point in
a paragraph of its judgment, admitted it objectively in another

paragraph.

Do Thou,

Histoire universelle,

Book

129, year 1603.

PEINCIPAL SYSTEMS OF CKYPTOGKAPHY

19

During the reign of Louis XIII. another decodist,


Antoine Rossignol, made himself known, to the discomfiture of the Huguenots.
"
1
It was in the year 1626," says Charles Perrault,
at the siege of Realmont, a city of Languedoc, then in
*

possession of the Huguenots, that he

first

gave proof of

was besieged by the army of the


Prince de Conde, and it opposed
King,
such a resistance that the Prince was on the point of
raising the siege, when a letter from the besieged was
intercepted, written in cipher, of which the most skilful
It was
in the art of deciphering could make nothing.
given to M. Rossignol, who deciphered it forthwith, and
his talent.

The

city

commanded by the

said that the besieged were sending to the Huguenots


Montauban to say that they were short of- powder,

of

and that if they were not supplied with some immediately


they would surrender to the enemy. The Prince de
Conde sent the besieged their letter deciphered, with the
result that they surrendered the same day.
Upon this
M. Rosinvited
to
he
Cardinal Richelieu,
being reported
signol to the Court, and the latter gave such astonishing
proofs of his skill that the great Cardinal, despite that

disposition which prevented him from


admiring many things, nevertheless could not forbear

extraordinary
expressing
usefully

his

surprise.

He

during the siege of

served

(Rossignol)

La

Rochelle,

very

discovering

the enemy's secrets by means of intercepted letters,


of which he deciphered with scarcely any trouble."

He

continued his activities under Louis XIV.,


held him in such high esteem that once, on the

all

who
way

back from Fontainebleau, he called on him at his country


1

Vol.

Les
i.

Hommes illustres qui ont paru pendant


Antoine Rossignol, Maistre. des Comptes.

ce

(11 th) siecje.

CRYPTOGRAPHY

20

house at Juvisy to which he had retired. The poet


Bois- Robert addressed many of his epistles in verse to
Rossignol, in one of which, in accordance with the wishes
of Cardinal Richelieu, he extols Rossignol's skill, regardhim as a redoubtable prodigy. The following is a

ing

rough translation of the passage


"

There is nought else beneath the skies


That may be hidden from thine eyes.
what a mighty art is thine
For by it provinces are won,
!

And
This

secret plans of kings undone.


is a right commodious art.

1 prithee

unto

me impart

Thy methods, and thus

justify

The years that be and those gone by.


The vanquished, fleeing from the fray,
Take oath a devil's in thy pay;
Hell's unseen

imps their packets steal,


Their secrets to thine eyes reveal."

There

is

extravagant
would wish
Staff,

amount

a certain

that

it.

both natural and

'

a good decipherer must have


acquired qualifications, the former

has stated that

necessarily

of truth underlying this

an Antoine Rossignol
eulogy,
Colonel Schaeck, of the Swiss General
not

playing a predominant part. The natural


are insight, the spirit of observation,
and perseverance. If a person be happily

qualifications

patience,
gifted in

any degree

for this kind of

work, and finds an


he inay

opportunity of developing his natural talents,

by study and practice a surprising degree of skill.


he will have to devote himself to a profound
of
the various systems of cryptography, have a
study
thorough knowledge of mathematics, and especially the
calculation of probabilities, and be acquainted with
languages and their literatures."
attain

For

this

PRINCIPAL SYSTEMS OF CRYPTOGRAPHY

21

Two remarks may

be added to this statement: First,


mathematics, we may be satisfied with
arithmetic; secondly, one thing is indispensable, which
Colonel Shaeck possessed, although he modestly refrained
from mentioning it common sense. I have heard of
default

in

of

a case where fifteen months of assiduous research failed


to produce
of a little

any

result, while, a little later,

common

by the

exercise

sense, the goal was reached in two

days.

In 1645 John Wallis, the English mathematician,


acting under the order of Cromwell, deciphered the secret
papers of King Charles I., which were seized after the
Battle of Naseby, and which proved that the King, in
negotiating with his adversaries, was playing a double
1

game.

On

1673, Louvois, the then French Minister


120 sterling, to
of War, paid 600 livres, equivalent to
one Vimbois for discovering the cipher of certain con-

July

2,

spirators; four days later he prescribed a similar fee to the


2
Sieur de la Tixere for a discovery of the same kind.
If these lines meet the eyes of any cryptographers, they
will regretfully

admit that the remuneration

for

their

3
arduous labours has dwindled terribly since that period.
In 1752 a German professor named Hermann, who had

defied

the

mathematicians

and

learned

societies

of

Europe to decipher a system of his invention, saw his


secret unveiled by a Swiss named Nicolas Beguelin or De
Beguelin,

son of the Mayor of

Encyclopedia Britannica,

Valerio,

De

art.

Courtelary,

village

Cryptography.

la cryptographie, vol.

ii.,

p. 11.

The amour-propre of the cryptographer does not always meet


with the respect due to it. For instance, a cryptogram which I was
3

charged
"

officially to

for the fool

decipher in May, 1917, resolved


reads these lines."

who

itself into

CEYPTOGKAPHY

22

situated in that part of the bishopric of Basle which was


then under the Bernese Protectorate. He had required

only eight days to discover the key. The story of this


incident is preserved in the History of the Royal Academy
1
of Science and Literature of Berlin.

was by methods used in cryptography that Miinter,


a Dane, and Grotefend, a German, succeeded in 1802
It

in deciphering a part

of

cuneiform inscriptions.
heads struck them by

the alphabet of the Persian


of angles or arrow-

One group
its

Miinter

frequent repetition.

'

'

to be equivalent to the word


king
pronounced
in
the harmonious language of the time),
(Kh-sliayathiya
it

and this supposition was eventually confirmed.


Mention may be made also of Bazeries, a French

officer,

who

not long ago succeeded in deciphering Louis XIV.'s


system of cryptography, comprising some 600 numbers,

some

of which represented letters and some syllables.


mine
could be written
Thus, for example, the word
'

in these four

ways
I.

46.

144.

230.
514.
535.

59.

125.

184.

374.
146.

IV.
still

i.e.,

II.

III.

and by

'

229.

other figure combinations.

As we have

seen, cryptography has at all times been


used
by conspirators, revolutionaries, and
extensively

secret societies.

On

this point I will confine myself to

the two following quotations


In May, 1603, a number of foreigners used to meet
in a house near Fontainebleau, which they had bought
:

Year 1758

(1765), pp. 369-389, with

two

plates.

PKINCIPAL SYSTEMS OF CKYPTOGKAPHY

23

the purpose of meeting secretly.


Their plottings,
however, were frustrated, as their house was raided, and
among other suspicious objects were found a quantity of

for

letters in cipher
'

Among
Lenotre,

'

which revealed the conspiracy." 1

the papers of the Chouannerie," says M. G.


are to be seen a number of sheets written

which formed the cipher used by


Georges Cadoudal and his associates at the time of the
Directory and the Consulate. The key of these is known."
The archives of the Foreign Offices in various countries
still contain cryptographic documents the keys of which
are lost and the deciphering of which the cryptographers,
accordafter interminable efforts, have had to abandon

in bizarre characters

A curious circumstance is that texts


ing to plan
written in cipher are encountered even among the hieroA certain inscription of Esneh contains a
glyphs.
!

profusion of crocodiles, in groups of as many as eighteen


The
at a time, the meaning of which is not apparent.

most hardened Egyptologists have not yet succeeded in


forcing the teeth

of

these redoubtable saurians apart


Certain mysteritheir secret.

and making them disgorge


ous languages

perhaps Etruscan, for instance

yield to cryptographic

methods

might

of decipherment.

*J

"

black cabinets," or postal espionage offices,


which were extensively used in France during the reigns
of Louis XIV., Louis XV., Louis XVI., and Louis XVIII.
If

the

unsealed letters to feed the police reports and to furnish


gossip to the Court camarillas, the black cabinets of the

German Empire
1

in the eighteenth century

were centres

Dulaure, Singularites historiques, p. 303.


"
See article on
Ciphers," in the Temps, September 29, 1917.

CRYPTOGRAPHY

24

cryptography. Count Briihl, Prime Minister of


Augustus III., Elector of Saxony, organised a completely
equipped establishment at Dresden. All the messages

of

received or sent

by the King

of Prussia's

Ambassador

in

that city were opened, copied, and deciphered during


a period of sixteen years, from 1736 to 1752. As soon
as the postal courier from Berlin arrived on Saxon territory, at Grossenhayn, his bag was picked during the

changing of horses, the

official letters

abstracted and sent

by a swift horse-rider to Dresden, where the black cabinet


unsealed, copied, and resealed them, and returned them
to the post, which delivered them at the same time as
the rest of the mail, which had arrived in the interval.
"
This black cabinet, known as the
Secret Dispatch,"

was directed by the Aulic Councillor Von Siepmann,


assisted by numerous experts.
Another dignitary, Baron von Scheel, officer of the
corps of cadets, excelled in forging handwriting, which

made

it possible to tear
open envelopes too troublesome
The Court locksmith was under orders to go
to unseal.

to the Legation and, with the connivance of the Prussian


Secretary, force the lock of the chest in which the Prussian

Minister kept the keys of the ciphers. 1


Thanks to their laudable activity, Saxony was aware of

the plans of Frederick II., and, when needful, communicated them to Austria and Bussia. Count Briihl, how-

game away at an official dinner, when he


mentioned
indiscreetly
something he had learnt through
his perverted laboratory.
Frederick II. changed his
of
and
thenceforth entrusted his
systems
cryptography,
correspondence solely to functionaries who were abso-

ever, gave the

Schldzers Staatsanzeigen, Part 62, p. 129

et seq.

PKINCIPAL SYSTEMS OF CKYPTOGKAPHY

25

1
But he did not complain, for
beyond suspicion.
for
some
time
carried on the same kind
he himself had
of espionage, which gave him a tangible advantage over
his opponents during the Seven Years' War.
a masterAustria, moreover, did not lag behind, and
her black cabinet was operated in a wing of the
stroke
Imperial Palace of the Stallburg, at Vienna. The staff,
who were Neapolitans and well versed in work of this kind,

lutely

directed their energies to the correspondence of all the


Ambassadors. On one occasion the deciphered copy

was placed

in the official cover addressed

to the Spanish

Ambassador, instead

from Madrid

of the original letter

The Spanish diplomat lodged a


Prime Minister, Prince
the
Austrian
with
complaint
Kaunitz. The matter was serious, and might have
extracted therefrom.

grave consequences, so
reprimanded the negligent official.

involved

The work done


accurately termed

in
"

the

black

Prince

the

severely

cannot be

cabinets
as

they merely
cryptography,"
deciphered cryptographic documents by means of the

*****

key, which they were


for themselves.

quite incapable

of

discovering

The literature on cryptography is very voluminous it


would be scarcely possible to mention in these pages the
titles of all the works which- have been published on
I need say no more than that, of all those
this subject.
I have read, the most substantial is the work of a Frenchman. I might mention, also, the name of Von Kasiski,
a German Major. Books, it is true, provide a great deal
;

of interesting material, but they do not help to decipher


1

The secretary changed

talents.

his

name and sought

other

fields for his

CEYPTOGEAPHY

26

documents which are


more than the best

in
of

any degree complicated, any


grammars can make a good

writer.

IV.

Let us

now examine some

of the principal systems of

cryptography or ciphers.

Broadly speaking,

all

the systems

may

be divided into

two categories: Substitutional, where the real letters of


a text are replaced by other letters, or by Arabic numerals,
or by any other signs; and Transpositional, which retain
the real letters, but shuffle

produce chaos.
1. In the Substitutional

them completely,
that

class

is

so as to

to say,

where

the letters are replaced by other letters, or by figures


or signs
are comprised the systems of which examples
have already been given: the first example, then those
of the Freemasons, of the zizgag
the two thieves.

and the thread, and

Here are some others: The Hebrew

cabalists

of

had

several cryptographic ciphers, which they used principally


to discover the hidden meaning of certain passages in

Thus the Athbash, the Hebrew spelling of


which forms the key A.Th.B.Sh consisted in writing

the Bible.

the last letter of the alphabet H (thaw) instead of the


first letter
X (alepli), and the last but one
(shin)

&

instead of the second


tion of the

Athbash

(beth),

resulted,

and

so on.

among

The applica-

other instances, in

1
identifying under the place-name Bheshak that of Babel,

or Babylon.

Another system, Albam, consisted in replacing the


letter of the first half of the
1

Hebrew alphabet X

Jer. xxv. 26.

first

(aleph)

PRINCIPAL SYSTEMS OF CEYPTOGEAPHY


by the

letter

first

of the

27

second half of that alphabet

7 (lamed), and the second letter of the


by the second letter of the second half

first

half

(belli)

etc.

(mem),
In a third system, the Aibakli, the interchange of the
But I shall
letters was based on their numerical value.
not dilate further on this, as that clever Hebraist, J. Buxtorf, has explained the whole thing far more clearly
in Latin than I can in a modern language.
Those
1
desiring further details are referred to his book.

Bacon thought he had found something wonderful

in

the following invention: He replaced each letter of the


plain text by a group of five letters, writing:

AAAAA

ABC.

AAAAB

AAABA

The method

of deciphering a document
obvious enough: the frequency of
the groups must be calculated instead of that of the
letters.
In the example given below, representing the
for

written in this

last

of

name,

ABAAA
we

is

a message, and, according to the most


supposition, the termination of a feminine

letters

plausible
Christian

way

BBBAB

ABAAA

are induced

by the frequency of the groups to read


ENE, and, accordingly, to presume such a name as
Irene, Magdalene, or Helene.
And, once we have arrived
at the probable value of two letters in a ciphered text,
success

is

only a question of time.

We

have already seen how the systems of Julius Caesar


and Augustus were written. They followed a parallel
progression: D for A, E for B, F for C, etc. But suppose
we break this symmetry, and say, for instance, that
1

De

Abbreviaturis Hebraicis, BasJe, 1613, pp. 24, 27, and 37.

CBYPTOGBAPHY

28

K=A, 0=B, V = C, P=D, H=E,

The

etc.

difficulty

then becomes apparent.


cipher square, or Vigeiiere's
is
table,
possible to write in cipher by means of
several secret alphabets, as many as four, five, six, or

By making

use of the

it

even ten or more, at a time, in periodical succession.


Here are the first few lines of Vigenere's table 1
:

ABCDEFGHIJKLMNOPQBSTUVWXYZ
abcdefghij k Imnopqrstuvwxyz
Bbcdef ghi k Imnopqrstuvwxyz a
Ccdefghij klmnopqrstuvwxyzab
D def ghi klmnopqr st uvwxy zabc
E efghij klmnopq rstuvwxyzabcd
j

etc.

Suppose we wish to conceal the word

'

'

hieroglyphics
in cipher by using three alphabets, in the first of which
B takes the place of A, C of B, etc. in the second E = A;
and in the third C=A. We accordingly adopt as the
;

key-word to our cipher the combination BEG, being the


letters

standing for A respectively in the three alphabets.


"
write the word
hieroglyphics," and under

We now

each letter add a letter of the word

BEG

in consecutive

order, thus:

rogly'phics
BECBECBECBECB

ie

The next thing

is

to look in the above table for the

in the horizontal line of capitals, and for B in


the column of capitals on the left ; at the point where the

letter

two

lines

the letter
1

commanded by these letters intersect we find


i, which we write as the first letter of our
The complete

table will be found on page 155.

PEINCIPAL SYSTEMS OF CKYPTOGEAPHY

29

The same operation for the letters I


m, which we take as our second letter, and

ciphered word.

and

so on.

yields
The finished result appears as follows

hieroglyphics
BECBECBECBECB
imgssimcrimet
Thus the word

"

hieroglyphics," written by the aid of


"
the key-word BEC,is transformed into
imgssimcrimet."
It will be noted that of the three i's in the cryptogram

only two stand for the same letter in the plain text;
the same is true of the three m's, while the two s's also
represent different letters.
Let us examine another cryptogram of the

tapvccigrqduprbhitvcc
a

same order

a c e o e a o

e c c.

this text has been ciphered


table and that the key -word is
reversing the process described

Given the knowledge that

by means

of Vigenere's

PIANO, we

operate by

above that is, we write the key-word PIANO repeatedly


under the letters of the cipher; then, looking for the
first letter of the key-word P in the vertical column to
the

left of

thereto the

the table,
first letter

we
of

find in the line corresponding


the cipher t, and at the head

column in which this occurs we note the capital


E, which will be the first letter of the deciphered
text.
Proceeding in the same way with the second
letter of the key-word and cipher, I and a respectively,
we obtain S, and so on until we have before us the whole
of the

letter

text deciphered as follows:


scraps of old iron."

Espionage compensation

As we have seen, deciphering by means of the keyword is quite easy when we know that word. When we

CKYPTOGKAPHY

30

do not know

it,

however, there are certain methods, a

too long to explain here, which permit of its discovery almost mechanically. All that can be said is

little

that in the above cryptogram, as well as in any other


document, the first thing to do is to find the

secret

vulnerable point in the armour and attack


at

weapons
Now, in

it

with the

just deciphered, the

weak

repeated three times, and it


After
this which will help us pierce the mystery.

spot
is

your command.
the text we have

the double letter

is

cc,

careful investigation, we find that they correspond in


each case with the letters on of the plain text Espionage,
compensation, iron, all three of which, in the ciphering,
fall by accident under the letters OP of the key-word
:

PIANO.

We

now examine

a somewhat different example.


are handed a document to decipher which reads as

Let us
follows

M ASE GX
I S

OM OX

AMOXEX
GK YYMNK
YKOSE K

The valuable information is afforded us that this


paper was confiscated from a traveller at Brigue, on the
that
Italo-Swiss frontier, and we therefore
presume
the text is in Italian. Noting that the first, third, and
fifth lines each contain six letters, while the second and
fourth have seven, we assume that the cryptogram is
more likely to be a list of words, or rather names, than a
'

'

phrase.

The

occurring most frequently is 0, which,


according to the rule of frequency, should represent e,
letter

PEINCIPAL SYSTEMS OF CEYPTOGEAPHY


but
in

it is

which

in vain that
it

thesis that

we

try to decipher the second line,

appears three times.

we have

before us a

are compelled at the

31

If
list

same time to

we adopt the hyponames, we


recognise how little

of proper

help may be expected from the manuals, otherwise the


grammars, of cryptography. For we encounter inter-

minable lists of family names in which the letter e is


not the most frequent: Bacon, Byron, Foch, Churchill,
Wilson, Dumont, Gounod, Marconi, Calvin, Loyola,
Cagliostro, Victor

Hugo, etc.
The axiom postulated by the books that the

letter e

the pivot in deciphering will not carry us very far,


so that another method must be adopted for deciphering

is

proper names not only those we have just enumerated,


but names in general. The frequency of their terminaIn
tions in each language must be taken as the basis.
French, for instance, 8 per cent, of proper names end
e.g., Mercier, Fournier, Gamier, Beranger,
Boulanger; 7 per cent, in on e.g., ond, ong, out : Masson,

in er or ier

Champion, Dupont, Leblond, Long; 6 per cent, in au


e.g., eau, and, aut, aux : Boileau, Rousseau, Moreau,
1
In Russian, ov and ev
Clemenceau, Nadaud, Caillaux.
terminate 35 per cent, of names; sky, 25 per cent.;
9 per cent.; itch, 6 per cent., etc.
Those who wish to take up cryptography

and

in,

to indulge

in these interesting calculations without undue mental


fatigue should confine their energies to Turkish family
names a by no means complicated task, for there are

none
even

In the Ottoman dominions

all that is necessary,


to say that one is called John
the son of James, or Ali the son of Mustapha. I once
1

It

mate.

for official records,

is

must be understood that these proportions are only approxi-

CKYPTOGKAPHY

32

who has long officiated as a


magistrate in those parts, how they managed to avoid
errors in a large city housing, say, 500 Alis sons of

asked a friendly Greek,

My

Mustaphas.

interlocutor

question, and answered


in identifying anybody."
:

'

seemed surprised at my
is no trouble at all

Oh, there

A little digression. In that happy country, not only


do fathers not transmit their family names to their
children, but, on the contrary, in certain cases, it is
rather the children

who

transmit their names to the

For instance, a certain Osman has a son named


who becomes famous. The father then changes
his name from Osman to Abu Taleb,
the father of
Taleb." An historical example is that of Abd el Caaba,
fathers.

Taleb,

'

who, having given his daughter in marriage to Mahomet,


was so proud of the event that he changed his name to
Abu Bekr, " the father of the Virgin." Later he became
Caliph and

first

successor to the Prophet.

Of German names, 25 per cent, end in er, and 6 per


cent, in the syllable mann : Troppmann, Bethmann,
Zimmermann, etc. Italian names end in i (40 per cent.),
o (30 per cent.), a (20 per cent.), etc.

This brings us back to our example. We will suppose


that the termination X, which is the most frequent,
At the end of the third name we find two
represents i.

supposed i's separated by a letter not yet identiNow, as our study of proper names has gone con-

of these
fied.

siderably beyond the rudiments set out above, we knowthat ini is the most likely ending: Bellini, Eossini, Mazzini,

We therefore assume that E =n.


problem now confronts us at the end of the
first word: n ? i.
Careful reflection leads us to suppose
that this word is a common noun in the plural, ending

Di Eudini,

similar

etc.

PEINCIPAL SYSTEMS OF CEYPTOGEAPHY


in nti (example: canti, conti, santi), and that
a heading or the title of the list, perhaps agenti.
this

make the

assumption, we

might be
Acting on

required substitutions

ready, of course, to try other suppositions


us

it

33

if

this fails

and our cryptogram assumes the following form

AGENTI
? E
G A

A moment's

I
?

reflection

A
N
? A
E N
?

? I

induces us to substitute the

which occurs three times in the second line,


once in the third, and once in the fifth. From ? e r r a r i
we automatically reach Ferrari. As our calculation

letter r for 0,

of frequencies

in Italian

name

terminations gives the

second place to o, we substitute that letter for the K's


in the last two lines.
The letter Y causes some hesitation,

but eventually we decide to replace


we have the following version:

by m, and

it

finally

AGENTI
F E E E A E
G A E I N I
T

MM A S

M OE EN

This method may seem empirical, even infantile, but it


often produces satisfactory results.
The difficulty becomes really serious in the system of
ciphering by means of Arabic numerals, in which a
letter, a syllable, or a word is represented by two or three
figures.

For example:

28. 71. 54. 75. 09. 62. 20. 65. 13. 79. 52. 32. 75. 88. 79. 43. 22.

stand
.

.").

for

"

'

Travaillez,

prenez
The numbers 54 and 09 each
.

("

Work,

mean

take

a; 13, 88,
3

CBYPTOGBAPHY

34

number 52 means nothing; the first 75?;,


The methods of deciphering here are
and
awkward to explain that I prefer
delicate, fragile,
leave them to the innate sagacity of the reader.

and

43, e;

the second 75 =r.


so
to

An undecipherable system is that which consists in


designating a letter by means of the number of the page
in a book, the number of the line, of the order of the
word

in that line, and, finally, the position in that


occupied by the letter in question, thus 1 27 6 4 2.
:

word
The

correspondent will decipher this if he has a copy of the


same book in the same edition as the sender.
Unfortunately, this system takes a long time to cipher,
and very long to decipher, without taking into account
the inevitable errors. Moreover, you may not find the

you may
you might use a c

you are using a French book, for


have to dispense with a k. True,
instead, but this would sometimes

lead to confusion.

Suppose you want to write:

letter

If

required.

instance,

'

is

besieged."
the same thing.

'

Kiel

'

del (heaven) is besieged


is
scarcely
Neither would your correspondent ever
:

His Majesty ill


cocher
phrase,
(coachman) summoned to general headquarters," the
cocher
was intended for the famous surgeon
word

guess that
'

the

in

'

Kocher.

*****

In Kussian books the letter /


in Italian publications

w and

is also infrequent, while


are
y
rarely seen.

Correspondence has sometimes been carried on in the


following manner: Most dictionaries are printed with

two columns on a page.

Instead, therefore, of writing

PKINCIPAL SYSTEMS OF CRYPTOGRAPHY

35

the required word, you adopt the word appearing on the


line in the parallel column, thus

same

WADE
THEEMOMETEE

instead of

BEAEEE
EQUAL

BELLICOSE

ESTUAEY
TOEEENT

TO
OCCASION

OMIT
The word

VISION

TEREIFY

appears here, but not

'terrify'

'terrified,"

which would not be found in a small dictionary.

And,

in fact, the disadvantage of using ordinary dictionaries


in this way is that the various grammatical distinctions

be shown.

Thus, with the aid of any dicSen<i letter," but not


Seni

cannot

all

tionary

you can say


which two phrases are diametrically opposed

'

letter,"

in

meaning.
Special

containing

have been compiled, each page


words in current use. Thus, for instance:

dictionaries
fifty

(page) 17

23
24
25
26
27

GEADUAL
GEANT
GEAVE
GREEK
GEEEN
'

Greek," you write


required to send the word
the number which precedes that word and the number
If it

is

of the page, 17.24, or the

Much more voluminous

whole in one number, 1724.


dictionaries have been utilised

or compiled, in which all the words are accompanied in


the margin by numbers ranging, say, from 1 to 100,000.
Let us endeavour to decipher the following crypto-

CEYPTOGEAPHY

36

gram, coded
words

from a dictionary

of

25,000

numbered

15128
21531
09043

24029
01643
24127
02432

24133
21682
05070
21531
01174

15311

The first thing to do and it is not easy is to determine


the exact meaning of two of the numbers, the same way
as when preparing a survey map of a country it is first
necessary to calculate with the utmost accuracy the
height and distance of two given points, to form a base
on which the triangulation of the whole region may be
effected, and the altitude of all heights therein calculated,
so in cryptography a secure base

a ciphered document.
Let us assume that
of the last

two numbers

we have discovered the meaning


in the

21531 =THE;
It will

must be sought decoding

above

09043=GENEEAL.

be noted that 21531 occurs twice, which would

favour the assumption that it represents a common word.


Success in deciphering this form of cryptogram, however,

depends mainly on a careful observation of the relative


values of the numbers and their comparison with the
approximate positions of the words in a dictionary. In
the above cipher, for instance, the three highest numbers
are all in the twenty-fourth thousand, and, as their values
are very close, we cannot go far wrong in assuming them

words beginning with W. This would place


the twenty-first thousand somewhere about T, so that
the probable initials of the first two words of the message
to stand for

are

W and T.

Leaving

this

on one side

for the

moment,

PEINCIPAL SYSTEMS OF CKYPTOGEAPHY

37

however, we will study the end, where the last two words
are assumed to have been definitely established as THE

GENEKAL.
numbers

Immediately preceding these, we note two


which occur numeri-

in the fifteenth thousand,

about

between those representing


and THE (21531). We accordingly
look in a dictionary, and find that the corresponding
Of words beginning with
position is among the O's.
to
THE
GENEKAL, we observe OF, ON,
likely
precede
and
and
OPPOSE,
OK,
provisionally select the first, OF,
to
15128.
We now have 15311, another
corresponding
cally

GENEKAL

half-way

(09043)

OF in the dicOPTION and ORDER, of which

presumed 0-word, occurring


There are

tionary.

than

later

the second seems the more likely. This doubtless follows


the word BY, which meaning we accordingly attach to
02432, the whole furnishing us with a useful tail-end:

BY ORDER OF THE GENERAL.


Great patience will be required to ferret out the whole
of the message, as there will be considerable fluctuation
in the position of the words, varying according to the

dictionary used to solve the cryptogram. We must


"
"
the most of the
landmarks
already more or less

make

The fourth word, 21531, is known to be


the word following is probably, though not
We note that the number reprenecessarily, a noun.
is
the
lowest of all, occurring doubtless
senting it, 01174,
identified.

THE, and

among the A's. The message being of a military nature,


we immediately think of ARMY and ARTILLERY, and
The next number, as well as
presumed to be a W-word, as we have seen.
The eighth number immediately precedes the phrase
By order of the General," and is therefore most likely

look for a further clue.


the eighth,

verb

is

expressing something

in

connection with

the

CRYPTOGRAPHY

38

supposed army or artillery. Consulting the dictionary


under W, we are attracted by the word WITHDRAW
or WITHDEAWN.
If the latter is correct, it should
"

some part of the verb


to be," and, in fact, the
seventh number, 01643, occupying numerically a position
follow

something over a third of the distance between AEMY


(or AETILLEEY) and BY, would seem to represent the

word BE itself. The sixth number, 24029, is a W-word,


and both from the context and its numerical position
a

WITHDEAWN

than
(24127) in the
excludes
other
dictionary,
reading but WILL.
any
We have now to tackle the first three words of the
little

earlier
it

cryptogram.

The

WITHIN
lines

or

WITHOUT.

offers

us

as

'probables'
After further study on the

we produce

described,

closely follows
(24127) in numerical

number, 24133,

first

WITHDEAWN

that representing
order, and the dictionary

WITHIN THEEE DAYS

words of the cipher. All that remains


to decide whether the fifth number, 01174, means

as the first three


is
'

artillery

"or

'

army."

The words occur

so closely in

the dictionary that this is no easy task, but after careful


"
"
and
calculation of the distances separating
be
by
'

'

from the beginning of the dictionary, we plump for


'artillery," and our complete message reads: WITHIN

THEEE DAYS THE AETILLEEY WILL BE WITHDEAWN BY OEDEE OF THE GENERAL.


It

should be added that in practice such documents


numbers written in this

are not often found with the

straightforward way.

and

all sorts of

Usually the figures are transposed

complications interspersed.
is to rearrange the order of the

A common method
figures in each

group upon a prearranged plan. Thus,


24133, 21682, etc., are transformed into 13432, 62182,

PEINCIPAL SYSTEMS OF CEYPTOGKAPHY

39

The great difficulty here is to discover the normal


order of the figures in each number, and to restore them
to their primitive form, before proceeding to the actual

etc.

deciphering.

It is a case of baling

The principal inconvenience

the ocean

of those

numbered

dic-

"
codes," is that when
diplomacy as
are
lost
or
in
most
others have to be
cases
stolen,
they
this
kind
be made in a
and
works
of
cannot
compiled,

tionaries,

known

in

day.

Even under the most favourable circumstances,

when

a fresh code

considerable

held in reserve for contingencies,


delav must ensue before instructions for
is

t/

use can reach those concerned, and the enemy


reaps the benefit. The following is one instance, among
others, of this disadvantage: During the Eusso-Turkish
their

War in 1877, the Ottoman Field-Marshal, Osman Pasha,


entrusted one of his Generals, Selim Pasha, with a confidential mission.
It so happened that Selim was the
officer responsible for ciphering,

and, being prudent, he

kept the code on his person. Unfortunately, he departed


promptly on his mission that he forgot to leave the

so

volume with

his chief.
The latter, during the whole
time of his Adjutant's absence, saw a pile of ciphered
telegrams from Constantinople accumulate on his table,

without being able either to read or reply to them.


V.
2.

The second category

of cipher

systems

is

the Trans-

which the actual letters are not changed,


but are mixed together or shuffled, and in effect really

positional, in

amount

to anagrams.

veil for live, are for ear,

for

Some anagrams
more

for

are very short:

Eome, wander

for

Vaussore 1 for Bousseau,

Galenus,
Angelus
1
The pseudonym adopted by Rousseau when giving
concert at Lausanne.

Andrew
etc.

his

In

famous

CEYPTOGEAPHY

40

we

however,

cryptography,

and 1,000

100, 200, 300, 500,

encounter
letters.

anagrams of
have seen one
be added that

comprising nearly 6,000 letters. It may


the longest are not the hardest to decipher; quite the

Among

numerous,

are included

'

we have already

daemonians, which

systems, which are very


"
the
of the Lacescytales

these

contrary.

considered.

easy enough to decipher is one which


"
the cryptologist Vesin de Eomanini called an
aerial

system

telegraph cipher." The first letter of the text is written


in the middle of the first line, the second letter at some
distance to the right in the same line, the third letter
similarly to the left, the fourth in the second line to the
left,

the fifth in the same line to the right, the sixth in

the middle, and so on, inverting the order of the letters


with each new line. Arrived at the foot of the page, a

new

start

in the

is

made

at the top, the letters being written


order as before, and immediately to the

same

A cryptoof those already put down.


grapher will have no difficulty in reading a text ciphered

right

or left

in this

way

EE

TO

HP

SO

EE

NT

GT

TT
00

UE

OH
TK

GW
HE

TA
which means:

THE STEONG OUGHT TO PEOTECT

THE WEAK.

It

Verne used

his

for

was a

similar

cryptogram

in

cipher

which Jules

Voyage into

the

Interior of the Earth.


"
"
Let us now pass to the
lattice," a wellgrille," or
known form of cipher. The grille is a square piece of
stiff

paper or cardboard in which a certain number of

PKINCIPAL SYSTEMS OF CKYPTOGKAPHY


holes

are

The square thus perforated

cut.

is

41

super-

imposed on a sheet of white paper, and a letter is written


in each hole.
This done, the grille is turned 90 degrees

what was the top left-hand corner


becomes the top right-hand corner. The further letters
of the message are now written in the holes, and the
to the right, so that

operation

continued until

is

four corners of the grille

all

have occupied the same position.

It

need scarcely be

cutting the holes in the grille, care must


be taken to arrange them so that overlapping of the
letters during the four turns will be avoided.
said that,

when

The following example can be read quite


means of the appropriate grille:
R

T M H
R

M
|W

SI
R

by

easily

t
G

IP
Y

L.

Deciphered, this reads:

EOME WAS COMPLETELY

DESTKOYED BY THE GAULS.


Grilles are usually larger than the above diagram,
which, however, will suffice as an illustration. As may
be seen, texts written in this code are very easy to read

CKYPTOGBAPHY

42

when one has the proper

Nevertheless, even
grille.
of
the
deciphering is not
grille
difficulty
of
this
volume I shall
and
in
the
second
very great,
part
mechanism
of
the
explain the
process by which it can

without the

be done.
officer

had

To complicate

cipher, a high Austrian


the idea of mingling a number of blank or
this

meaningless letters with the others; but this did not


increase the difficulty of translating, as is proved by the
fact that such a

preferred.

scarcely ever used.


employing dividers is much

system

The method

of

is

to

be

It consists of cutting vertical slices in a text

and mixing these columns of letters.


short example just three names

Here

is

a very

MAD K
V

P A E

first

proceeding then to
the letters of the second column, next those

of the last,

fourth,

first,

other order as agreed on.


2, 6, 4, 1, 5, 3:

How

I S

which we divide into vertical


write

E N N A

slices,

fifth,

and third

or in

some

In this case the key will be


I
V P I N S D E K,

ADAKN

the recipient of this brief message to deal with


it ?
He knows that the key agreed upon provides for
six letters in a line.
Since the text contains seventeen
letters,

is

he proceed^ to draw the following graph:

PEINCIPAL SYSTEMS OF CKYPTOGEAPHY

43

crossing through the last square, after which he writes


cryptograms in the order indicated

in the letters of the

by the key
etc.

the second column, the sixth, the fourth,


On no
three names will be restored.

when the

account must he forget to cancel the last square, for if


he absent-mindedly wrote a letter therein during the
operation of deciphering, the whole of the little text
would be thrown out of order. The name PAKIS, for
instance,

IAMDE.

would be changed to

This system

can be complicated indefinitely.


Let us examine another example :-

V H
U E
P

N E
N H
P L T N
D S S F

W
T

L E

MB

This contains thirty-six letters, and is therefore very


short; nevertheless, the number of possible changes of
position which one might effect in the letters, to ascertain
their meaning, is so enormous as to be practically in-

comparison, take wheatwe could isolate a particle, and under a micro-

calculable.
flour.

If

For the sake

of

scope count how many such scarcely perceptible molecules could be contained in a cubic millimetre, we should
us say, 100,000.
Now, to form a sphere as
as
our
terrestrial
large
globe, it would require a number
of these particles equal to that of the combinations which
find,

it

is

let

possible to

make with

the thirty-six letters of our

cipher, which would have to be represented by a

series

of thirty-seven figures.

The key

to this last text

preceding cryptogram.

is

the same as that of the

Accordingly, the plain text must

CKYPTOGEAPHY

44

be written six letters to the line, and the first letters of


our example will form the second vertical slice of the
graph. The phrase we are about to discover had for
its author a great Captain who lived a century ago, and
accomplished victories in Europe by the side of which
1

the present victories of our enemy are insignificant.


He succeeded in a much more magnificent enterprise:

he won the admiration of the enemy peoples.


ing by the graph produces the following:

KV

HE WOE

FK

The

first

line is

L P Z U

S I
S N

Decipher-

B L E

E N

of blank letters, intended

composed

simply to embarrass the decipherer. The text begins


"
from the second line:
The word impossible is not in

French." 2
'

'

This system of
dividers
-which distantly recalls
the Lacedaemonian scytales, and was dubbed by an early
nineteenth- century writer
the undecipherable cipher
'

'

par

excellence

-is

very

difficult to

decode when one has

more complicated than the elementary


3
It may be pointed
specimens we have just presented.

to deal with texts

1
NOTE BY TRANSLATOR. When M. Langie wrote this, the
Germans were inflated with their military successes.
2
The actual words are: " Le mot impossible n'est pas fra^ais."
One method of complication, calculated to exercise the patience
1

of the decipherer, consists in suppressing, without leaving any trace,


if I may say so, of a certain number of e's in a text in such a way

as to upset the calculation of frequencies.


But this proceeding is
dangerous, inasmuch as it does not offer absolute security, and one

runs the risk of entangling both oneself and one's correspondent.

PEINCIPAL SYSTEMS OF CEYPTOGEAPHY

45

out that the second and longer example, being regular,


difficult to decipher than the first, which, though

is less

shorter,

is

irregular.

There are systems in existence which are literally


undecipherable, the ciphers being sometimes composed by
means of ingenious machines resembling the cash registers of the shops.

But

there

is

"

but

'

-it is

probable,

and often certain, that systems absolutely undecipherable to an inquisitive outsider will also be so to their
recipients, however well provided the latter may be
with the desired keys. The reason is that important news
As soon as it is a question of
is nearly always urgent.
has
to
be
had to telegraphy 1 or radioresort
urgency,
Now, in a long alignment of letters which
telegraphy.
are meaningless to him, the most skilful of telegraphists
due to inattention?
will commit involuntary errors

And when a telegram


or reflex movements.
runs into a number of lines and has to be retransmitted

fatigue,

several times, the case

is

worse.

that not 10 per cent, of telegrams in cipher


In the first place,
are free from errors on their arrival.
It is stated

there

is

continual confusion between the letters u and n,

and a, e and c, e and I, m and n, even in plain texts.


Then it is so easy, by a false movement, to change the
one letter s (...) into the two letters i (..) and e (.),
into the single
or the two letters m ()
-) and t (
o

letter o

-).

letters that

do

all

It is precisely these extra or

missing

the mischief.

to make the whole


in the example
have
seen
text meaningless, as we
MADEID. Hence, if one must use systems very diffi-

One

is

error

is

sometimes

sufficient

Incidentally it may bs pointed out that the telegraphic alphabet


nothing else but a system of cryptography.

CEYPTOGEAPHY

4G
cult

none the

indispensable to
choose keys in which one error will not cause a repercussion throughout a document. Furthermore, it is not
to

decipher,

it

is

less

always convenient to carry about a dictionary or a code.

CONCLUSION.

When

one has a taste for cryptography, and oppor-

tunities arise to devote oneself to

it

seriously, the study

develops into a passion. At first the amateur is bewildered.


He must make persistent efforts, and not be
discouraged by reverses. At all costs he must continue,
assiduously persevering with trials not made haphazard,
but reasoned out and based on induction and hypotheses.
The slower the result is obtained, the more tardily success

crowns our
faction

efforts,

the more profound will be the satisthe goal, and, like

we experience when we reach

Archimedes, exclaim

"

Eureka

!"

PART

II

EXAMPLES OF DECIPHERING

A
ONE day
into

my

CONSULTATION.

a gentleman sent

up

and was shown

his card

office.

'

an unfortunate accident that


pleasure of making your acquaintance," he
It

is

to

What

affably.

from you
will

is

excuse

morning,

this: I

has driven

have

me from

owe the

said,

very

me

to seek enlightenment
been sent here on a mission; you

going into details.

had scarcely entered

my

brought in a letter addressed to me.


understood thing that those who sent

Arriving this
a postman

when
Now,

hotel

it

it was an
must write in

It is a
cipher all communications of any importance.
wise precaution, for you will see, if you examine the
it had been opened by steam, stuck down
dried.
and
By whom ? None of the
immediately
again,
hotel people could or would throw any light on the

envelope, that

subject.
"

The cipher agreed upon between us is a grille. I did


not bring the grille it might have gone astray but I
had noted on an old letter a method for quickly reconstructing the necessary grille, to be destroyed as soon
as

it

had served

its

purpose.

This method consisted

merely of a list of the numbers of the small squares to


be cut in a square sheet of paper, which would enable me
47

CEYPTOGEAPHY

48

to read the secret message transcribed on to another


square placed under the perforated sheet. Every night

which I no longer want, and I fear I


have
may
inadvertently burnt the letter containing the
in
I was able to get your address without
question.
key
difficulty, and am come to beg you to bring all your skill

1 destroy papers

into play, so that I

may know the

contents of the message

without delay."

While saying these words, he handed


which ran as follows 1

text,

me

the ciphered

aitegflytboeehreauwnanoarrdrteet
hoshfpetapotoyhlretihenemgaoarnt
a total of sixty-four letters, or the square of eight. Even
without the knowledge that I had to find a grille, that

would have been the


I

and

first

idea to occur to me.

visitor to call again at the end of an hour,


begged
immediately set to work. First I copied the text on

my

to a square divided into sixty-four sectional squares, like


I numbered the four corners in
that appearing below.

Eoman

numerals, and further added Arabic figures to the


sixty-four squares for the purpose of easy identification.
principle of the grille system has already been
explained on p. 41. I revert to it merely to point out

The

that the

grille,

grille

numbered

at the four corners in

Eoman

exactly over the text, the corner of the


numbered 1 covering corner I. of the text, corner

figures, should

fit

2 of the grille corner II. of the text, and so on. If, in


our example, the first hole in the grille exposes the letter

(square 1), when the position of the grille is changed so


that corner 1 covers corner II. of the text, and corner 2
1
NOTE BY TRANSLATOR. The text of the cipher, as well as
certain portions of the explanatory matter, have been modified to
meet the requirements of the English translation.

EXAMPLES OF DECIPHERING

49

covers corner III., etc., the same hole will expose letter
A further operation will reveal the letter T
(square 8).
(square 64), and a final turn the letter
(square 57).

similar result will be produced by all the other holes in


the grille, which, in each of its four positions, will enable
a quarter of the actual text to be read.

i
1

CKYPTOGKAPHY

50

shows through, and outside the four corners I write the


Arabic numerals 1, 2, 3, 4. I then examine the text and
endeavour to form a useful syllable.

On

the

first

line

my

attention

is

attracted by

AT

and 3) a common enough word in English,


and one which might easily form the beginning of a
message.
Accordingly, I mark the place of these two
letters on my tracing paper; after which I turn the latter,
(squares

not a quarter only, but a half round, so that it is now


reversed, and corners 1 and 2 cover corners III. and IV*

The marks made on the tracing paper now


K T (squares 62 and 64). This is a very
good word-ending, and it is evident that from the last
two lines we could easily extract the word H E A E T
of the text.

coincide with

squares 53, 54 (or 56), 59 (or 61), 62, 64. Marking these
and again reversing the tracing paper, we find in the
corresponding squares 1, 3, 4 (or 6), 9 (or 11), 12 the
combination A T E (or F) T (or 0) E. This not being
'

very satisfactory, I abandon the combination and try


another.

Having seen that the tracing paper


position, I turn

mark

my

is

in its original

attention to the second line, and

THE

This is con(squares 9, 14, 16).


ceded by investigators to be the commonest trigram in
English, and is almost certain to occur in a text of sixtydecide to

Keversing the tracing paper as before


to the corresponding squares 49, 51, 56
the
marks
brings
This is
in the sixth line, indicating the letters E T E.
four

letters.

quite a promising combination,

which should precede

it.

The

and

first

I look for the

vowel

that meets the eye

is

(square 45), while three squares farther back (42)


E T E, which
appears P. We now have the group P
seems to call for the final E, and sure enough this letter

EXAMPLES OF DECIPHERING

51

occurs in the last line at square 62, though, of course,


N T (squares 63, 64) are possible.

now

reverse the paper to ascertain what letters correspond to the new marks, and bring to light T (square 3) and
N (20, 23). We now have the group T T E N 0, the
I

being doubtless a final, and the last two the


a new word.
of
The next proceeding is tenbeginning
of the grille, which I
to
the
construction
begin
tatively
first letter

do by drawing squares round the letters

ER

(42, 45, 49, 51, 56, 62).

now

Let us

take crayons in four colours say blue,


With the blue crayon we make
red, green, and yellow.
a small mark in the text itself against each of the above

We now turn the tracing paper, but


a
quarter, and our six marked squares now
only
letters I
U S A G (2, 11, 18, 35, 41, 58). To
attach a brown mark. Incidentally, it may
six letters.

this time

cover the
these

we

be noted

that the results of the quarter turn, unlike those of the


complete half, are not necessarily to be read consecutively.

further turn brings us to the group T


20, 23), which we mark in red;

(3, 9, 14, 16,

turn produces

LAE

HEN

(7,

24, 30, 47,

HEN

and a final
54, 63), which

we

indicate in green.
have now neutralised 24 out of the 64 squares,
thereby narrowing considerably our field of research.

We

R T E R, we look
Coming back to our original group P
for a likely word to precede it, and are favourably inclined
There are two H's between
towards
(32, 36, 39).
the T and the E, and we adopt the second experimentally.
Marking these and reversing the tracing paper,
we find the three corresponding letters to be R T H

THE

(26,

29,

NORTH

THE

This enlarges our red group to T


a result which proves that we are on the right

33).

CEYPTOGEAPHY

52

track.
Accordingly, we mark in the four colours the
corresponding letters in the four positions, bringing the
total of neutralised squares to 36.

Progress onwards is by leaps and bounds. We have


simply to study one or other of the coloured groups,
ignoring meanwhile the now numerous ear-marked
squares.

For instance, on scrutinising our two red words,


and the unmarked letters following them,

THE NOKTH,
we quickly

discern a

P and an

0, and think of

'

Pole."

duly found in squares 38, 43, 48, and 50,


and, having marked these in red and the corresponding
letters in the other positions in their appropriate colours,

These

we

letters are

find

that

accounted

only

twelve

squares

remain

to

be

for.

The materials for the grille are now almost complete,


and we are able to cut out 13 holes from the 16,
which, in the four positions, will enable us to read the
whole of the text. Our four coloured groups, each
requiring three letters to be complete,
follows

now appear

as

TTHENOKTHPOLE
ELBEWA E FT HE AN
BANDTHEPOBTEB
Brown:
IF OUAR S AYINGA

Bed:

Green:
Blue:

glance at our text shows


letter between the

unmarked

(square 8) to be the only


in the brown line,

F and

which discovery enables us to fill in all the other blanks


automatically, and at the same time proves the brown line
to be the beginning of the' message.
We can now complete the cutting of the grille.

At

this juncture

my

visitor appears.

covered anything ?" he asks eagerly.

Have you

dis-

EXAMPLES OF DECIPHEEING
;

'Here is your
gether: IF YOU

grille," I reply.

Let us read

53
it

to-

ABE STAYING AT THE NOKTH


POLE HOTEL BEWAKE OF THE MANAGES AND
THE POETEK."

But our friend suddenly looks grave.


Good heavens !" he exclaims, I have some important
papers in my portmanteau." And with a bound he dis'

'

appears downstairs.
It is to be hoped he arrived in time.

WHERE
The Chief
invited

me

is

THE MONEY

French Secret Service Department had


upon him, and after giving me a cordial

of the

to call

welcome, said:
You are aware that, following on the robbery at the
Continental Bank, the notorious individual whose identity
has not been established, and who is known only under one
of his aliases, Pastoure, has just

been sentenced to

five

Had

he taken scrip payable to


years' imprisonment.
order or bearer, we should have been at ease, but he has
confined his attention to cash.

stamp would hide the stolen sums


them intact on regaining
his liberty, and he is too keen a psychologist to have
confided his secret to an accomplice. He has worked
single-handed, and we only managed to lay hold on him
thanks to an accident, a hole having been torn in one of
the fingers of the rubber gloves he wore in his operations.
By this circumstance we secured an imprint of his thumb,
and identified it with the anthropometric record already
made on the occasions of his previous collisions with

Now,

in such a

a fellow of his

way

as to be able to find

justice.

Unfortunately, six days elapsed between the robbery

54

and the arrest.


ments during the

CKYPTOGKAPHY
We have been able to

trace his movetwo days, but are still in the dark


as to how he employed his time during the first four.
Being determined to leave no stone unturned to recover
the money, I have continued a strict investigation.
last

Yesterday I visited the central establishment housing


our man, and learnt that Patoure had, almost immediately
on his entry, asked permission to write his will, which he
handed sealed to the prison registrar. I had the package
produced, and, having obtained a warrant from the Court,
took cognisance this morning of the last wishes of the

and
The text

prisoner,
;

in his presence.
of the will

was somewhat to the

effect that

author bequeathed his watch, a ring, the contents of


his purse, and his personal effects to his brother and sister,
its

who would make themselves known when

required if an
advertisement were inserted in a big daily asking for the
heirs of M. de Pastoure.'
'

'

I was struck by the aspect of the fourth page of this


document, which the prisoner had covered with figures.
I asked him what this meant, and he replied that they were
merely calculations of interest on the income from land

held in
"

common by

his family in their village


the paper itself. I do not know why the
contents of the fourth page perplex me. You see that
in the four columns into which the page is ruled off,

Here

Pastoure

is

has

multiplication,

written

and

sums

division,

in

addition,

Please take the document, examine

me know

subtraction,

with erasures everywhere.


it

at leisure,

and

let

your convenience what you think of it."


I put the will into my bag, and, having arrived home,
studied it with eager curiosity.
I

at

began by verifying the

results of the arithmetical

EXAMPLES OF DECIPHEEING

55

operations occurring therein. At first glance the whole


was a confused medley. The number 158, for instance,
multiplied by 86, was shown as 4,311 and similarly other
sums were quite wrong. After prolonged reflection, I took
a notebook and recorded the various observations sug;

I counted all the


gested by a preliminary examination.
on the page, and found there were 144. I then

figures

made

list

of

them, column by column, as follows

201010453787243667351876936177952984
847201224675294851852723645223612111
588643110225011813414375659521401593
579131544939454441021252448100642212
These 144 figures were distributed in the following order

of

frequency:

1245370689

Figure:

Times occurring: 24 21 20 17 12 11 10 10 10 9
I

became more and more convinced that there was a key


But the frequency of the ten

to be discovered here.

In French, even in a long phrase,


figures led to nothing.
the whole of the 25 or 26 letters of the alphabet are rarely
used.

On

the other hand, the shortest phrases require

at least 16, 17, or 18 different letters.


of Voltaire's,

where the

n's

The famous

line

'

Non, il n'est rien que Nanine n'honore,"


and e's form nearly half the total number of

letters, absorbs 12 different letters.

It is scarcely possible,
in
Kunic, to write phrases with a total
except perhaps
of ten different characters.
In French, when Arabic

numerals are resorted to for secret writing, groups of

two

figures, at least, are generally

employed.

Accordingly, I proceeded to divide the 144 figures into


sections of two: 20, 10, 10, 45,

and

so on.

There were

72 such groups, but on a calculation I found that there

CKYPTOGKAPHY

56

were never more than two groups alike: two 10's, two
etc.
Twenty-one groups were repeated, and 30

45's,

occurred only once.


Matters were not making

much headway. The letter e


on an average once in six letters,
no means of deciding which of the
I then
21 repeated groups could represent that letter.
48.
of
there
were
divided the 144 figures into groups
three;
But my disappointment was, if possible, still greater,

occurring, in French,
or 17 per cent., I had

one group being duplicated 201 all the rest


occurring only once. I went on to form groups of four,
then six, eight, and twelve figures, after which I stopped.
only

In each category I had, of course, arranged the groups in


numerical order, from the lowest to the highest. On

examining them in rotation,

my

attention was attracted

more

particularly by the three-figure groups, and for this


reason: I was struck by the very small difference between

Thus:
certain groups, which followed at intervals of 1.
010, Oil; 110, 111; 223, 224, 225; 453, 454; 642, 643,
then with a lacuna 645; 851, 852. It then occurred to
me to add together the three figures of the highest group
:

+ 4=21. The groups following this in value


or
9 + 5 + 2=16; 939, or 9 + 3 + 9=21.
gave: 952,
984.

'Hallo!" I said to myself,

seems to exceed 21 when

And

"none

add the three

of

these groups

figures

composing

thereupon reflected that in many French


phrases the letter z, 25th of the alphabet, does not occur;
nor y, the 24th; nor x, the 23rd; the last in common
it."

Perhaps, after all, each of the


three-figure groups represented the order of a letter in
the alphabet.

usage being

v,

the 22nd.

Accordingly, I made the trial, and added the figures


This resulted in the groups 010 and 100-

of each group.

EXAMPLES OF DECIPHEKING
sum

each producing a total

57

and 110, each 2;


The group which
I had at first sup-

of 1; Oil

021, 111, and the two 201 's, each


produced the largest sum was not,

3.

as

posed, one of the highest, but 787==22. According to


my new hypothesis, the plain text probably contained
the letters a to

v,

if

the language were French.

was

disconcerted, however, to find that the most frequent


total of the additions was not 5, corresponding to e, but 9,
'

equalling

i,

according to

my theory.

It

might be Latin,"

I thought.

Another serious irregularity which came to

light

was

that the most frequent totals after 9 were 15 (five times),


corresponding to the letter o, and 21 (five times), to u.

This

is

contrary to the rules of letter frequency, not only


e, the most frequent letters are
r, s, i (only the i of our three supposed letters has

in French, where, after


n,. a, i,

any place

here),

but also in Latin, where the letters

should occur in the following order: i,


etc.
Still we have here the i, u, and o

e,

s,

u, a, n, o, r,

and, in any case,


was desirable to put our supposition to the test.
The first three figures, 2013, meant c\ the next three,

it

010=a; then 453==12=1.


be:

to

The complete text proved

'

Calvisius,
3

'

Opus CJironologicum.

Bibliothegue

Municipale.
So he has deposited the sequel to his secret in a
And with a psychological forevolume," I thought.
'

'

by no means stupid, he has not trusted to his


memory, having had experience of the transformations
which memory can effect in a name after a certain number
sight

NOTE BY TRANSLATOR.

gram, the

There are two errors in the cryptobeing represented by the group


letter in Municipale by the group

final letter in Calvisius

936=18 = r, and

the third

CBYPTOGBAPHY

58

of years.
Furthermore, he has chosen, as the receptacle
of his confidences, a kind of work which is among the

Old books on
law or theology are sometimes referred to, but ancient
manuals of chronology are generally allowed to sleep in

least consulted in a collection of books.

peace."

The same day I repaired to the Bibliotheque Municiand asked for the volume by Calvisius. It was a
quarto tome bound in a thick leather cover. The cryp.
togram giving no indication of a page, I thought Pastoure must have made some secret entries on the first or
last pages.
There could be no question of sympathetic
pale,

ink, since the necessary manipulations to make it visible


I
were scarcely possible in a public reading-room.

expected to encounter some letters dotted in pencil,

which if joined would form words and phrases but my


hope was vain, although I did not stop till I had scanned
every page of the volume.
I was lost in conjecture, when the idea occurred to me to
examine the inside of the back of the book, pressing the
I perceived no note or anything
latter completely open.
;

else.

Still

reflecting,

looked at the inside cover.

noticed nothing at the beginning of the volume, but at the


end, in the top corner, the paper was somewhat creased
and seemed to have been moistened. Feeling the place

with

my

fingers, I

became aware

of the existence,

under

the paper, of a hard object, small and slender. It was


imperative that I should see what was hidden there,
so I had the book put on one side, and went out to
obtain a small sponge, a bottle of water, and a tube
of

gum.

Armed with

these objects, I returned for my Calvisius^


and, operating in the same way as Pastoure must have

EXAMPLES OF DECIPHERING
done, but vice versa, I slipped the

little

59

sponge, soaked

in water, into the suspicious place, and, while waiting


for the moisture to take effect, I turned to p. 215 and

plunged into the mysteries of the chronology of the Kings


Ezechias and Nabonassar. When the desired result was
obtained, I drew from

which bore the name

gummed

hiding-place a small safe key,


Bauche and a number. I re-

its

the paper, and, having returned the volume,'

went

in quest of the Chief of the Secret Service.


"
Have you any idea of the meaning of the figures ?"

he asked, shaking hands and indicating a seat.


"
"
They have enabled
Why, yes," I replied.

me

to

find this little metal object."

He
Picture the astonishment, then joy, of the Chief
made me describe point by point the development of
!

my

Then he

'

discovery.

started on the chase, accom-

Two days

on
hundred
opening my newspaper,
thousand francs which had been stolen had been recovered from the strong-room of a bank, where a compartment had been rented for fifteen years by a client
panied

by

his

sleuth-hounds.

later,

I learnt that the thirteen

about to start

for Australia

ARABIC NUMERALS.

now

give another instance of success in the discovery or key to a cryptogram. It was in Arabic numerals.
One day I received in the usual buff envelope
I will

the following text: 1

67534 34959 61496 54860 46495 14564 46496 25350


65646 04950 45664 45966 49664 56649 60494 96646
59665 06249 50536 65060 57496 85849
1

See footnote on p. 60.

CKYPTOGKAPHY

60

began by arranging the numbers in order


from the lowest to the highest:
I

of importance,

04950 06249 14564 25350 34959 45664 45966 46495


46496 49664 50536 54860 56649 57496 59665 60494
61496 65060 65646 67534 85849 96646
It will

be noticed that two-thirds of these groups range

from the 40 to the 60 thousands, and two 46495 and


46496 differ only by a unit. Have We here one of those
systems of combined codes, involving much complication
in their decipherment, but rarely used,

by reason

of their

extreme complexity ? On the other hand, these groups


of five figures may be purely arbitrary. 1
A striking peculiarity is the preponderance of 6's and
4's, which occur 30 and 24 times respectively, whereas

and

and 3

7 each occur only twice, 8 three times,


four times.
1, 2,

Perhaps

it

is

possible to group the figures differently.


into groups of four figures ?
No,

Can we divide the text

because there are 110 figures. Neither can we form


groups of three. We can, however, form 55 groups of

two

figures

67
64
59
50

53
46
66
62

43
49
49
49

49
62
66
50

59
53
45
53

61

50
66
66

49
65
49
50

65
64
60
60

48
60
49
57

60
49
49
49

46
50
66
68

49
45
46
58

51

66
59
49

45
44
66

Having made the division, we observe that the group


49 occurs twelve times, which is somewhat above the
normal frequency of the letter e in a total of 55 groups.
We will, therefore, assume provisionally that 49 =e.
1

NOTE BY TRANSLATOR.

At

first

sight this cryptogram

would

be hard to distinguish from the dictionary cipher described on


p. 35.

EXAMPLES OF DECIPHEEING

61

We next note the frequency of the numbers 49 (supposed


66 in the series 66, 49, 66, 45, 66, 49, 60, 49, 49, 66.
and
e)
As a general rule, only a consonant can follow the doubled
e in English, and the most likely consonants are n, t, d,
and I. The first three letters in the series might well be
d e d, but the sequence would be: ? d e ? e e d
which gives an unusual number of d's in a small group.
Suppose we try the much more likely n as the equivalent
of 66.
We now have the series nen^ne^een. The
}

letter

seems to be the obvious consonant to

fill

the

second lacuna, and by replacing the first by the vowel i,


we get the complete word nineteen," with ne as the
This promising result
tail-end of the preceding word.
'

yields us the following four equivalents

49=e; 45=i; 66=n; 60=*.

We proceed to make trials with these four letters, and


observe that, preceding the word
nineteen," there
occurs the group tetinttne (60, 49, 50, 45, 66, 44,
'

59, 66, 49).

The combination leads us to the idea that


question, in which case there can be no hesi-

a date is in
"
ted in June."
tation in filling in the three blanks thus:
Our theory is confirmed on examining the series following
"

nineteen

which

"viz.,

lundted
"

is

obviously
Before going any farther,

50),

(46,

59,

66,

50,

62, 49,

hundred."

we summarise the

results

so far obtained, to wit:

d=50; e=49; 7t=46; i=45; j=44; n=66; r=62; i=


and are immediately struck by the fact that the numbers
proceed in two regularly descending sequences, so that,
without
alphabet

further
:

trial,

we

are

able

to

construct

our

CKYPTOGKAPHY

62

a=53

EXAMPLES OF DECIPHEKING

63

stopped; we got out and proceeded on foot to a solitary


house, where a bright light gleamed from a window on

ground floor. Somebody knocked at the door.


While waiting for it to be opened, the leader of the expedition took me apart and informed me that a domi-

the

was about to be carried out in the rooms of


arrested that day, and who had
strongly protested, saying he had been entrusted with a
ciliary search

man who had been


mission.

In

examination

diplomatic
to

me

for

the

papers

was to search

to

be

submitted

for proofs of that

statement.

At

we were

and I was installed before the


on
which documents in various
drawing-room table,
languages were being piled. I buried myself in the
tedious and wearying task of selection, putting on one
last

side the screeds

examination.

inside,

which seemed to deserve a more minute


had no knowledge of the case or of the

man.
But something was taking place at a table at the other
end of the room, where a mysterious personage, who was
addressed as
Mr. Deputy," was occupied in taking
allegations against the arrested

'

notes.

It appeared that a considerably larger supply of


provisions had been found in the rooms than was authorised by the Food Order; and a detective, who had a

reputation for smartness, had brought in a tin box which


had aroused his suspicions, though it apparently contained

only

flour.

There was also a sack of flour in the larder,

and this pound or


had puzzled him.

so,

kept separately in a writing desk,

Contemplating the white powder with an air of absorp"


the detective murmured:
Old flour sometimes
contains worms; I wonder whether there are any here."
And while " Mr. Deputy " and other functionaries looked
tion,

CKYPTOGKAPHY

64

on with interest, he passed the flour through a sieve which


he had just procured, letting the fine powder fall on to an
Suddenly, a small cylindrical object
With obvious
a
of
case for steel pen nibs.
sort
appeared,
and
delight the detective examined this object, cleaned
therefrom
opened it, and, to our astonishment, produced

open newspaper.

two ribbons of pink paper, covered with characters in red


"
ink, which he deferentially submitted to the
Deputy."
The latter abandoned the air of indifference which he had
hitherto displayed, and eagerly seized the two documents,
which he began to study with deep interest.
Several of us formed a circle round him, and I was
able to read over his shoulder one of the texts:

YOUWOULDHAKDLYKNOWTHEBE
WASAWABPBOVISIONSABEPLEN
T

IFULANDQU ITE CHEAP.

("

You would

know there was a war.


and quite cheap.")

hardly

visions are plentiful

Pro-

The other text began with the letters USLAAVI,


followed by several more without any apparent significance, though the words IDOL and SHEBA stood out
the meaningless array.
at
Beturning to my seat, I glanced from time to time
"
the
Deputy." He was comparing the English text
with the strange medley on the second ribbon, and seemed
to be making great efforts of memory.
Finally, with the
careworn air of one who has not solved a problem, he

among

of paper into their case


carefully pressed the two bands
his
into
latter
the
and put
pocket.

hours had elapsed since our arrival. The exami"


nation was finished and the seals affixed. The Deputy
off at
disappeared, and we rejoined our motor. Going

Two

'

EXAMPLES OF DECIPHEEING
a good pace, under the light of the
about two in the morning.

moon,

65

home

reached

The word uslaavi kept running through my head.


Surely this was a Slavonic word meaning
glory."
And could one connect it with the words idol and
'

'

'

"

'

Sheba ? It might be some ritual. On the other


hand, the initial word of the mysterious writing might
be a variant spelling of the Kussian uslovie, meaning
'

terms." I imagined some semi-Oriental conspiracy,


and was frankly seized with a tormenting desire to know
the whole text of the document.

On the following morning I was immensely gratified


on hearing a policeman announced. He came to invite
me to call upon Mr. Deputy at an hotel near the
'

upon a matter of great urgency. I at once made


my way thither, and was immediately introduced into
the presence of that important man, who plunged
station

without preamble into the business.


You see," he said, these are the two documents
'

seized the other night.

One
it

Each contains

sixty-five letters.

evidently the transcription of the other, to which


has been attached in error. I have compared the
is

frequency of the letters in each, and here


Letter
Plain text
:

Cipher

We

is

the result

ABCDEFGHIJKLMNOPQRSTUVWXYZ =
70136103401404531533414020 65
71

57013402

03710

73310011 = 65

not hide from you


But it must be done,
for we have received similar writings from other sources,
on the same kind of paper and in the same red ink. Try
and get on the track of the method by which we can

have to discover the key.

the fact that

it

I will

will be difficult.

decipher them.
'

For instance," he continued,

'

both texts have an


5

CKYPTOGEAPHY

66

number

and T's. There are no


The cipher contains B, G, and Z,
which are absent from the plain text, and my theory is
that these and other redundant letters, such as seven

equal
J's,

of A's, H's, I's,

M's, or X's in either.

instead of six E's, are intended to play the part of letters


which are present in the plain text but absent from the
cipher."
I

went home and shut myself

with what eagerness


With the respect due to a relic, I drew
in

be imagined.
the precious paper from my pocket case and began to
study it. It read as follows:

may

USLAAVIPICASDHOIOTOEIDOLY
SHEBAHADADSTCESKENESONEZ
TUKUKDGOELOACSNB,
A

rapid glance led

ne to the conclusion that the three

words which had seemed so portentous were merely


accidental groups in the cryptogram, and I proceeded to
experiment on the lines indicated by the 'Deputy,"

comparing the text with the' supposed transcription.


I soon became convinced, however, of the absolute
impossibility of arriving at any result in this way, and
began trying other methods, putting aside the plain
text.

There being sixty-five

letters in the cryptogram, I


decided
against the theory of a grille, which
temporarily

usually requires a square number.

The

pairs

OE. AD,

repeated, gave me the idea


of loo ing for a key-word (see p. 70), but the intervals
between the repeated groups furnishing no satisfactory

ES, NE, and

UK, which were

indication, I passed

on to another hypothesis.

I noticed that the letter

sequence of five letters, thus

occurred three times in a


:

and that the same

EXAMPLES OF DECIPHERING

67

thing happened with A: A A A. This favoured the


idea that the cipher had been composed with the aid of
"

'

the system known as


that is,
dividers
(see p. 44)
the required phrases had been written in very short lines
and the letters separated into vertical sections, which,
placed end to end, had formed the text now before my
eyes.
Accordingly, I began to cut the text into groups
of

letters,

which

constructing the
the two groups

juxtaposed with the object

of

re-

As a nucleus I took
original
just mentioned, and arranged them

in vertical columns, thus:

text.

CKYPTOGKAPHY

68

to interpose a third group between the


upon STCES. This produced:

two, and hit

first

OCA
TEDA

|I

The second

W in the cryptogram, but

it

'

with," for there was no


"
might be I think." Adopt-

be

line could not

ing this idea, I succeeded quite easily in adding three more


groups to my word-skeleton, to wit: VIPIC, ENESO,

and

UKUKD,

and now had quite an imposing array:

OS

A V E U
H I N K
C A P E U
T E D I S K
D
S A C
T

But I could get no farther none of the remaining groups


would fit in.
I had, of course, marked each group of letters in the
cryptogram as I had used them, and now found that
several letters were isolated, and that there were two
groups with only four letters each, among some longer
;

I looked again at the partial


series, as yet untouched.
"
"
the
words
reconstruction.
think,"
save,"
Certainly

and
cape "seemed too good to abandon. I wondered
"
whether the last could be a part of the word
escape,"
and in order to test this, omitted my first column, OIOTO,
These
substituting the two groups OEIDO and USLAA.
could only be adjusted by moving them down one line.
"
think
was now preceded by OU instead
The word
'

'

EXAMPLES OF DECIPHERING
of I, so I

"

69

"

completed the word you by adding the group


read the following

YSHEB, and now

SAVED
NK
YOUTHI
SES CAPEU
HILEDISK
EDASACOD
BOA
"

.._.>

-i

The first line looked as though it might be save us


I had two spare S's, but neither of the groups containing
them would suit the rest of the context. I then extended
each of the last five columns by one letter downwards,
This made the last
following on from the ciphered text.
that
ANG was part
line read: BOAEDANG.
Assuming
"
"
of the name of a ship, the word
on
seemed the proper
"
word to precede
board." To introduce this, I prefixed the two groups DHOIO and OACSN.
I was gratified to note that the second line now read
"
do you think "; but the third line was not so flawless,
being HASESCAPEU. A glance at the last column
showed me a means of correcting this: it was the group
UKUKDG. By cutting off the first two letters and
was
think
sliding the column up two lines, the K of
j

'

'

"

has escaped."
preserved and the third line became
Success was now a foregone conclusion. It turned out
that the original text had been written in lines of eleven
and had then been divided into vertical sections,

letters,

of

which the

fifth

ciphered

text,

the

and

on.

The

so

had formed the


eighth
first

first

forming the
line,

SAVEU,

letters

second

had

of the
series,

to

be

abandoned, and the complete reconstructed text proved


to be:

CKYPTOGKAPHY

70

Y OUT HI N K Z
HAS ESCAPED!

DO

OCHILEDISGU
ISEDASACOOK
ONBOAEDANEU
TEALVESSEL
(" Do you think Z has escaped to Chile disguised as a
cook on board a neutral vessel ?")

The document

in plain

language which accompanied

the above was merely intended to throw investigators


off the scent.
Having found the key, I lost all further
interest in the cryptogram.
I was not at all curious to
learn for

the

whom

the message was intended, any more than


"
Z." I concerned
of the person referred to as

name

ciphered and

myself only with forwarding the whole


plain texts, key

and

my

rough working

to

my immediate

principal.

CIPHERING BY MEANS OF A KEY-WORD.


Let us suppose that I

am

requested to decipher the

following cryptogram:

pbvddzobgqwwnzscczafstx
xpt hktddds kbpf vpc
vpafsvkzfej tvybipqoaasyb
acr pwhsml net gkniysxf vyc
plddlahvwccvpzdqagtcwdj
i

vi ys ds

There are 120

letters in the text.

I note the following

repetitions: ip, dd, cc, fv, ds, sx, vp, vy, yp, afs, iys, cvp.
I calculate the intervals by making a pencil mark between
the i and p in the repeated ip's (there are three of them),

and count the


1

on

The reader
p. 48.

letters

is

between the marks.

again referred to the Preface

and

do the same
to the footnote

EXAMPLES OF DECIPHERING

71

with the other identical groups, and draw up the following


table

From

ip

dd
dd

dd 33
dd 62

cc

vp
vp
vy
yb
afs

,,

iys

cvp

11

,,

,,

vp
vp 60
vy 33
yb 9

,,

,,

31

,,2x3x3x5
,,2x2x2x2x3

fv 48
ds 11
sx 61

fv

ds
sx

,,

cc 90

,,

3 X 7
11
11

X
X
X
X

3
ip 63 letters, or
3
33

ip to

ip

61
3

,,2x2x3x5
X

11

,,3x3

afs 31

,,

,,

31

,,3x3x7

iys 63
cvp 60

,,2x2x3x5

be noted that the factor 3 occurs in eleven out of


it is fairly safe to assume that a keyword has been used in coding the text, and that this word
It will

the fifteen lines, so

The question

contains three letters.

is:

Can we discover

key -word and successfully decipher the text ? We


begin operations by copying the whole of our text into

this

three columns

that

is,

in lines of three letters,

numbering

each line to facilitate reference:


(1)
(2)
(3)

p b
v d d
z o b

(31)

s x
v y

(22)

v y b
i
p q

(23)

o a a

(24)

y b

p
(34) d d

bpf

(25)

cr

(35)

(16)vpc

(26)

pwh

(27)

(36)

(37)

v p

(28)

(29)

g k

d q a
(89) g t c
(40) w d j

(12)

x p t
f h k

(13)

(11)

d d
d s k

(5)

w
wnz

(6)

sec

(7)

z a

(8)

s t

(18)

(9)

(19)

v p a
f s v
k z v

(10)

(20)

(4)

g q

(14)
(15)

(17)

(21)

(30)

(32)
(88)

(38)

ahv
c c
z

CKYPTOGKAPHY

72

The first column begins with the letters i v z, and ends


with d g w
the second column begins with p d o, and
ends with q t d and the third column begins with b d b,
and ends with a c j.
;

The next thing

to calculate the frequencies in each

is

column, which gives us the following table:


First column:

each;

s, v,

7 each; d,

w, 3 each; a, /,

i,

n, o, p, x, 1 each.
p, 8; d, 5; c, y, 3 each; a, h,

g,

z,

/,

b, e, k,

Second column:

2 each; g,j, m,

w, x,

o,

Third column: c, 6;
2 each h, j, q, r,

b,
s,

z, 1

4; a, /,

w,

x, 1

i,

n, q,

s, t,

each.
/c,

v,

I,

3 each; d,

t,

y, z,

each.

E is the commonest
T
or
followed by
S; the commonest
bigrams are TH and HE, and the most frequent trigram
and three-letter word is THE. We may, therefore,
assume that p in col. 2 stands for E. In col. 1 we
According to the law of frequencies,

letter in English,

hesitate between

How

sent E.

s and v, either of which


can we arrive at a decision ?

Looking down our table


that

(col.

17, 37).

If,

2)

is

of

numbered

preceded by

may

we note

lines,

v three times

repre-

(lines

16,

we

get

therefore, v (col. 1) represents E,

the combination (lines 16 and 17) EE?EE, which seems


unlikely.
Eecalling that one of the commonest bigrams
is

HE,

col.

1.

c v p.

let

us substitute

In our
If

of

list

HE

we adopt

easily infer that c v

for

as the value of v in

repetitions we find the group


as the value of v p, we may

p equals THE, and

this

combination

and 36-37,
and 2 on the succeeding

does, as a matter of fact, occur in lines 16-17

the

in col. 3

and v p

in cols. 1

lines.

If

we

we have established one equivawe can immediately ascertain the

are satisfied that

lent in each column,

EXAMPLES OF DECIPHEKING

73

key -word used from one or other of the ciphering tables


and armed with this word decipher
the cryptogram automatically.
The process will be
at the end of this book,

explained in due course.


Meanwhile, it will be interesting to see whether it is
possible to effect the decipherment without knowing

and without reference to the ciphering


suppose that, for some reason or other,

the key-word,
tables.

We

will

we have not

at our disposal such useful adjuncts for


a
finding
key- word, and that we are without any clues
outside the cryptogram itself to help us in the deci1

pherment.
So far, then, we have established the following:
v (col.

1)=H; p

(col.

fc)=E;

c (col.

8)=T.

Our copy of the cryptogram, written in column form,


with numbered lines, should have sufficient margin to
attach the transcription of the letters as we ascertain
them. We now attach the letter
to all the v's in col. 1,

to the

in col. 2,

>'s

and T to the

c's

in col. 3.

Lines

once with the group


"
HETHE, which looks like a part of the word whether."
We therefore tentatively add as the equivalent of / in
16-17

attract

our

attention

at

col.

3,

and

same column, duly


the similar letters in the column.

as that of a in the
all

marking accordingly
The next thing we notice is the group WH?T in lines
31-32, and we decide to fill the blank with A, attaching
in the middle column.
?/'s occurring
For the moment we cannot go any farther in this
direction, so we fall back on the law of frequencies, which,

this value to the three

however, might easily prove a


1

NOTE BY TRANSLATOR.

edition, but is

pitfall if

we

did not recog-

This experiment is not in the French


added here to amplify the example.

CKYPTOGEAPHY

74

nise the possibility of

remembered that the


in col.

so that

1,

numerous exceptions.

letters s ancU;

we could not

It will

be

each occur seven times

at first decide

which was

However, having eliminated v by


likely to stand for E.
to
it
the
value
of H, and noted that the next
attaching

and v in order

letter after s

of frequency in the

column

we feel justified in assuming that


s=E, and accordingly mark in seven E's in col. 1.
We now find that one of these E's occurs in line 10,

occurs only three times,

and another
by the

and that in each case it is preceded


the second and third columns of the

in line 31,

letters i

(in

preceding line). As the most likely group of three letters


ending with E, and repeated in the same text, is THE,
we tentatively adopt T and
as the value of i (col. 2)

and y

We

(col. 3)

are

respectively.
able to resume the thread of our internal

now

clues with the group (lines 7-10)

which we
new equiva-

WE??HTHE,

construe as "weigh the," thus obtaining two


lents
i.e., t (col. 2)=I; x (col. 3)=G.

It will, perhaps, be as well to tabulate the results so far

obtained

Col. 1.

Col. 2.

Col. 3.

t=I
2/=A

/=W
x=G

y=K
A

glance at the above will show us

how we may

find a

possible short cut in our operations. It will be noted that


in the middle column ^ T and, vice versa, t=I.
can

We

soon ascertain whether this principle applies throughout.


The result of a trial, as far as we can go, confirms this
hypothesis,

and we quickly arrive at some gratifying

EXAMPLES OF DECIPHEEING

75

Our attention is first directed to the group


"
such as."
20-21) S?CHA?, which we identify as

results.
(lines

Isolated

groups

begin

to

join

E?T?YWEIGHTHE?B?EC?,

as,

"

ently weigh the object,"


adverb
an
yet to be discovered.

thing else but


part of

up,

for

instance,

which can scarcely be any-

ENTLY

being

Always sub-

we
stituting the new equivalents as we establish them,
20
line
continue to build up words and phrases. From
we can now read SUCH AS ?E ??Y EEAS?NA??Y
E??ECT F??? THE? WHAT, which is soon resolved into
"
such as we may reasonably expect from them what,"
etc.
In fact, we automatically decipher the rest of the
cryptogram as fast as we can note the equivalents, which
leap to the eye with ever-increasing rapidity.
Although we have solved the cryptogram (and the reader

should by now have the complete plain text before him


if he has
duly followed our reasoning with pencil and
still do not know the key -word by which the
we
paper),
cryptogram was ciphered and by which it could be deci-

phered without resorting to the long empirical process


just described.

Let us go back to our starting-point that is, to where we


had established only one equivalent in each column viz.

Col. 1,

=H;

col. 2,

p=E;

col. 3,

c=T.

These three letters are presumed to have been ciphered


from three separate cipher alphabets, each indicated
by a letter. The three indicating letters taken together
form the key-word, as agreed upon between the sender
and recipient of the message. Our object is to ascertain
this key-word.

Vigenere's ciphering table on p. 155,


look along the top line of capitals for the letter H,

Turning to

we

first

CEYPTOGEAPHY

76

from which we proceed directly downwards in the column


on
immediately below until we arrive at the letter v and
the left of the line in which this occurs we find the capital
letter 0, which should be the first letter of the key-word.
We proceed in the same way with the letters E and p,
producing L as the second letter of the key-word; and
;

Accordc, which gives us J as the third letter.


to
is
OLJ.
this, then, the key-word
ing
Thus armed, and with Vigenere's table before us, we

with T and

cryptogram, and proceed as de-

refer to the text of the

scribed on p. 28.

We

first

write the key-word repeatedly

under the text, thus:


i

pbvddzobgqw,

OL

OL

OL

OL

etc.

in the column to the left of


Starting from the capital
the table, we follow the horizontal line which it commands

and stop
text.

until

at the letter

i,

the

first letter

in the ciphered

From this i we ascend the column containing it


we reach the top line of capitals, where we find the
letter of the plain

letter

U.

This should be the

text.

We

continue in like manner with the second letter

cryptogram and
produces E, and so on.

of the

following:
But here
at

all.

first

p and L, which
thus decipher as far as the

of the key-word,

We

ueshsuldssfn.
stop, for this array of letters makes no sense
are evidently on the wrong track. What is

we

We

the next thing to be done ? Fortunately, Vigenere's


table is not the only ciphering instrument known to
that of
cryptographers. Possibly the table used was
Porta, which will be found on p. 153.
To use Porta's table, we take our first pair of equivaand we look in the top line for which*lents
i.e.,

v=H

EXAMPLES OF DECIPHEKING

77

ever of the two letters belongs to the first half of the


we then descend until we
alphabet in this case h
;

encounter

'in

the same column the second letter of the

At the

the double line containing the conjunction of the two letters will be found two capital
and Z. Either of these it is immaterial which
letters,
pair, v.

left of

key-word. We proceed
with
the
second
e
and p, which yields as
similarly
pair,
the second letter of the key-word E or F, while the third
will be the first letter of the

pair, c

We

and

t,

gives us S or

as the last letter.

say, therefore, that the key -word is YES.


As before, we write it repeatedly under the text of the
will

cryptogram, and, following the instructions accompanying Porta's table, proceed as follows
:

cz af,
pbvddzob gqwwnz
YESYESYESYESYESYESYES,
weshouldsuff ci entl yw,
s

In this

We

etc.

etc.
etc.

the complete text is deciphered easily


should sufficiently weigh the objects of our hope,

way

whether they be such as we

them what we propose


Our readers

may

reasonably expect from

in their fruition."

will doubtless recognise

Addison's obiter

this

as

one

oi

dicta.

A good cryptographer

would have detected at once that


more likely to have furnished the
the three initial pairs of equivalents which

Porta's table was the

key-word, for
gave the clue to the cipher consisted of letters belonging
to different halves of the alphabet, and Porta's table is
constructed that no letter can be represented by
another in the same half of the alphabet, whereas in
Vigenere's table there is no such restriction,
so

In order to decipher quickly by means of a table

it is

CKYPTOGEAPHY

78

as well to write out the whole of the text of the crypto-

gram, accompanied by the key-word repeated throughout,


then to proceed with the deciphering of all the letters
under the

of the key-word
as, for instance,
followed by those under the second letter, E,
and finally those under the last, S. In the case of Vigenere's table a set square is a useful aid.

in

first letter

YES

BILLET-DOUX.

gentleman called upon me and complained that the


behaviour of his son was not giving him entire satisfaction. It appeared that, while casually glancing through
the textbooks used by the young man, who was studying
for his B.A., he had found the missive which he now

produced.

Before mentioning the matter to his heir, he

was anxious to know the meaning of the three lines in


the document written in secret characters.
It was a sheet of blue paper, satined and perfumed,
signed with the initial J, and contained the following
(I have added numbers to the signs)
:

DBR

3L

10

Fib.
12

11

LQQ
RIORJL HJL
J

19

20

26 27 28 29 30

31

15

D_IL

16

17

18

32

33

"
36 37

38 39 40 41

Lb. 3LL_Z]
48 49

50 51

52 53

22

21

42

23

24 25

34 35

Id

43 44 45 46 47

RJRZIbJ
57 58
54
55

56

13

14

EXAMPLES OF DECIPHEEING

79
"

Very good," I said to the anxious father;


kindly call to-morrow about two o'clock ?"

will

Left to myself, I began to study the cryptogram.


The signs 1-42 are between quotation marks.

you

The

most frequently occurring sign is No. 7, which is repeated


nine times in all about the normal frequency of the
letter

in a total of fifty-eight letters.

One

peculiarity

struck me: the word starting from sign 43 begins with a


doubled letter. This furnishes us with a useful piece of
the text cannot be French, a language which
does not contain words beginning with doubled letters.

information

Examples
German, Aal

such words occur in English

of

Aar

(eel),

(eagle), Acts (carcase).

eel

in

Leaving

aside Gaelic, 1 a language not very extensively used, the


two principal languages which contain a considerable

number

of

words

of this sort are

In Eussian, a whole

series of

Kussian and Spanish.


words begin with w, the

commonest being vvedienie (introduction). A certain


number of other words begin with ss, among them
ssylka (exile), and ssora (quarrel).
Perhaps the word
formed by the signs 43-47 is this very Eussian word
ssora.
As if to confirm this, sigh 45 is the most frequent
in our text, and in Eussia o is the commoneust letter.
In this case, the word formed by signs 50-53 should be
odno (one) or okno (window). But n is one of the most
frequent letters in Eussian, whereas sign 52, supposed
it, occurs only twice in the whole text.

to represent

Furthermore, the word formed by signs 7 and 8, which,


according to our supposition, should be od or ok, is meaningless in Eussian.

We

must, therefore, abandon that

language.
1

NOTE BY TRANSLATOR.

with oog (eye), een

(a,

Mention might be made of Dutch,


and other similar words.

one), uur (hour),

CKYPTOGKAPHY

80

Let us now pass on to Spanish. Here the only letter


which can be represented by the double initial 43 and 44
is

of

I,

and, in fact,

II

common

very

forms the beginning of a large number


Spanish words. In this case, sign 45,

the most frequent, would be e. These two letters, I and e,


occur again at the end of the last word of our text, but in

reversed order, el. This is in a word of five letters, of


which the first is the same as the third, so that it can be

no other than papel (paper).

Knowing now the

and e, we observe that signs


and 10, 31, and 34 (a) are all
followed by the same final letter, which can only be s,
in which case 33-35 is las (the) and 50-53 undoubtedly
In our text we count nine e's, seven a's, and
este (this).
7

and

seven

letters a

41, representing

s's.

e,

According to the rules of Spanish crypto-

o occurs as frequently in that

graphy,
not more frequently.

language as

s,

if

Now

the sign occupying the fourth


rank in order of frequency in our text is No. 3, which
occurs five times. It is quite likely that this stands for
o.

less

We

should then have for the word 23-25 SO?, doubt-

son

(are).

and 21-22

With n tracked down, we

identify 5-6

as no (not).

In Spanish, the commonest group of three letters by


is que; the word 12-14 ends with e, and, its first two
letters being so far unknown to us, might well be que
far

(that, than).

tu (thou, you).

This seems probable, for then 48-49 will be


If 9-11 is

mas

(more),

86-42=MU?E?ES

must be mujeres (women).


Summing up the letters

so far obtained, we note that


the alphabetically consecutive letters m, n, o each consist
is
of a square, with this difference, that the square

n contains one

dot, and o two dots.


Comparing
these with the other angles and open squares, with and

blank,

EXAMPLES OF DECIPHERING

81

without dots, we are able to construct a symmetrical


graph containing the complete alphabet, from which we
can supply the letters still required to decipher the

cryptogram

CRYPTOGRAPHY

82
"

Tubwc

Efs ekftfo

cfs

Fsef hkfcu,

Fs xbs hftft'hofs woe hfinkfcu.


Fs ibssf efs Wotufscmkdilfku.
Ko tfkofs Obdiu tkdi pgu hfgsfwu
Ft gsfwf tkdi, xfs ekftft mkftu,
Ebtt Fs, hmfkdi kin, wotufscmkdi ktu."
!

We

begin by calculating the frequency of the letters

The letter occurring oftenest is /, of which


there are 33, whence we may deduce /==E. In actual
practice, e in German has a frequency of 18 per cent.,
in the text.

or

an average

156

of 1 in 5J letters.

we ought

letters,

18+10=28

E's.

The

supposed to be such,

As

have

to

this verse contains

here
of

proportion

proportionately
E's,

or

letters

therefore somewhat higher than

is

the normal average.


According to an Austrian authority, Colonel Fleissner
von Wostrowitz, the letters following E in order of fre-

N I E S T. We will suppose,
the letter in our text occupying the second
place in order of frequency (17 times), stands for N.
Next in order are k and t, each fifteen times. One of
quency in German are:

therefore, that

s,

these should signify

I,

the other R.
equalling S.

(eleven times), probably


choice between

we have the

e, i,

and

Then comes u
letter T

For the

o (each eight times).

Let us confine ourselves at first to the two leading


letters /=E s=N.
We have the more reason to believe
:

these equations correct from the circumstance that in


German n is the most frequent terminal letter. Now, out
of the thirty-three words comprised in the verse, ten do,
in fact, end with our supposed N.

ten end with EN, which

is

Indeed, nine out of the


with the

also in conformity

rule.

Now

that

we have

at

our disposal two practically

EXAMPLES OF DECIPHERING

83

certain letters, let us substitute the plain letters for the

ciphered ones standing for them. We shall then have


88 E's-f-17 N's, making 50 known letters out of a total
of 156.

The undeciphered

letters are replaced

' 4

by dots:

T71
-*Y"i

"- J

.1

(-*

\^

f-*n
JJ

(-*
\_J

.*,

I' vi
-L_^ J J
i

f*
'

c^
\^

En."

Here we are brought to a stop there is no such word as


en in German, whether with or without a capital letter.
We must have got on the wrong track through our too
;

docile adherence to the rules given us.


But riot much
is done, since w*e have
started.
Where
only just

harm
is

the fault ?

For the moment we will retain our confidence in e, and


assume that it is n which is out of place. A two-letter

w ord
r

or

es,

German beginning with e can only be eh, ei, er,


apart from such imported expressions as en bloc,

in

en gros, and en-tete.


Can the word in question, then, be Ei (egg) ? No, for
"
"
it occurs three times in the sextet, and
is not a
egg
term likely to be repeated so often in the poetic style.
True,

if

this

w ere the
r

case, the fifth

word

in the first line

might be Eile (haste), but the first word in the lirie,


formed of the same letters, would then be Lei, a term
non-existent.
Can our word be Eh (before) ? No, for
although this would enable us to read the fifth word as
Ehre (honour) and the first w^ord as Reh (roebuck), we
"

"

occurring twice in the same line,


which is incredible; besides, the second reh is not written
with a capital letter, and cannot, therefore, be a noun in

should have

German.

roebuck

Furthermore, the text would contain a

dis-

proportionate number of words ending with eh. Neither


can the word be Es, for while the fifth word would then

84

CKYPTOGKAPHY

be Espe (asp), the

first

word would be Pes, which

is

also

German language.
Er
is
and
we find this meets the case well.
left,
Only
The fifth word in the first line now becomes Er e, which

non-existent in the

can be no other than Erde (earth). The first word will be


Der (the, who).
Assuming, therefore, that s=~R and

e=D,

the

first line

'Der d.
Further

trial

be diesen

reads:
der Erde

e. e

..

e..."

favours the idea that the second word must

(this),

J=S

the value

first

word

in the

k=I

in the

word

fifth

diesen

line:
is

being arrived at from the


Improbably ES. That

confirmed by the second word

in the fourth line, which, with the letters so far ascer-

tained, gives us tfkofs=seiner (his),

word,

Ko=In

From

and by the preceding

(in).

deciphered we make the


that is, the ciphered letter
following discovery:
stands for the letter preceding it in the alphabet; s=B,
e=D, the same remark applying in each case. Perhaps
the letters already

/=E

be the same for the whole of the alphabet. We


accordingly make the trial, checking the result of the

it

will

equations from time to time:

a=Z

g=F

n=M
t=S
It will

b=A
h=G
o=N
u=T

c=B
i=H
p=0

w=U

d=C
k=I

q=P

x=W

be noted that the letters

e=D
1=K
r=Q

f=E

m=L

y=X

a, q, r, y,

and

s=K
z=Y
z of the

secret alphabet, corresponding to the plain letters Z, P,


Q, X, and Y, are absent from the sextet, which we now

read as follows:

EXAMPLES OF DECIPHERING
"

85

Der diesen Staub der Erde giebt,


Er war geseegner und geliebt.
Er harre der Unsterblichkeit.
In seiner Nacht sich oft gefreut.
Es freue sich, wer dieses liest,
Dass Er, gleich ihm, unsterblich

ist."

In English:
"

He who gives this dust to


Was blessed and beloved.
He waits for immortality.

earth,

In his night he has oft rejoiced.


Let him who reads these lines rejoice,
That he, like him, is immortal."

The same work contains a score of ciphered passages,


some of which are less easy to read than the above
example.
N.B. It

is

worth noting as a rare phenomenon that

this sextet contains only

German

terms.

It is far

more

German

texts to find a proportion of pure French


5 to 8 per cent., or more.
from
words varying

usual in

A SHORT

CUT.

have just received a picture postcard from a young


"
who signs himself M. J." It depicts a pretty

friend

rose-covered cottage near Penzance, in Cornwall.


On
the address side, in the part reserved for correspondence,
appears the following. I number the signs for reference:

1234 56789
+

101112131415161718192021

v-iv +x + nooTi-n>n<vLo
i

22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

+VooJLLVr+AcoD+<~inrA

42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

CKYPTOGKAPHY

86

We

begin by constructing a numerical table of

signs in order of frequency:


Sign.

Times.

all

the

EXAMPLES OF DECIPHEEING

87

already noted as a consonant often following E, and


These three triplets are
three begin with the sign 00.
20.21.22,

32.33.34,

of the triplets

is

and

53.54.55.

THE, we have

between the signs J,

"[">

and

Assuming that one

for the value

Q. The
PL occurs

H to choose

first,

J, occurs

once before and


twice before E, while the last,
once after E. The second sign, "f, appears only once
in the cryptogram.
We are therefore inclined to assume
the sign J to represent H.
Just as we are about to examine the possibilities of
the triplets 44.45.46 and 51.52.53, which open with E

and end with


clue.

T,

it

occurs to us to search for an external

Turning the card,

we observe the name

'

Pen-

Our young correzance," which suggests a short cut.


in his message.
the
name
mentioned
has
spondent
possibly

We

note that the word contains two E's, separated by


five other letters.
Examining the cryptogram, we find
that there really is such a group 21-28. The two N's
at 23 and 26.
The initial
are represented by the sign

P, however, proves to be the sign J, which we had assumed


to be H.
This letter must, therefore, be one of the two

or
signs
the text.

The

-that

is,

if

the trigram

THE

occurs in

results so far established are as follows:

1=A,

|=C,

+=E, V=N, J=P,

00 =T,

-=Z.

Having marked the equivalents in the cryptogram as


we have gone, we note that the group following
"
Penzance "i.e., 29-35 is ?EAT?E?, which it does not

far as

require

much

imagination to transform into

From A?EW

(11-14)=" a

few"

to

WEATHEE.

?A?N?F?CENT

in three or
(36-46)=" magnificent," we reach our goal

CKYPTOGEAPHY

88

four steps, thanks to our short cut, and finally read the

following
'

Am

spending a few days at Penzance.


Kindest regards."
magnificent.

DICTIONARY CODE.

The following cryptogram


5761 3922 7642 0001
8576 1378 2799
4049 3261 4176
2137 4049 2485
8576 8337 0702

8686
7670
8686
2475

Weather

9219
6018
6638
7948
9185

is

handed

6448
4212
4833
0300

to

6016
3940
4827
9712

In practice, this kind of cipher, which

me

4570
0644
0001
0300

is

4368
7262
3696
4212

7159
8686
6062
9576

very commonly

subject to arbitrary complications, and it may well


a long task to restore each number to its
quite
prove
original integrity, the sender having probably shuffled the

used,

is

four figures throughout in accordance with a formula


agreed upon with the recipient.

But

as it is always best to proceed from the simple


the
to
complex, we will act on the preliminary assumption
that the above numbers have not been changed; and are

We

to be read just as we see them.


a list of the forty-five numbers, of

begin by making
which the lowest is

0001 and the highest 9712, arranging them in numerical


order
2485
0001
4212
6062
8576
4212
0001
2799
6448
8576
3261
0300
8686
4368
6638
0300
3696
8686
4570
7159
3922
0644
8686
4827
7262
3940
4833
9185
7642
0702
4049
9219
5761
1378
7670
4049
6016
9576
7948
2137
6018
2475
4176
8337
9712
:

EXAMPLES OF DECIPHEKING
We

89

apparently have to deal with a dictionary code,


Faced with a system like
1 to 10,000.

numbered from

simple and regular, one has to be on the alert


should conceal a trap. On one occasion, in an
example which seemed quite as clear, I produced the
this, so
lest it

"Either X or Y warmly recommended."


But subsequently I ascertained that the numbers had
"
"
in the cipher, and that the true reading
been
cooked
"
of the phrase was
Both X and Y absolutely unknown."

reading:

Assuming in the present case, however, that the


numbers are unaltered, we make the following observations: The number 0001 occurs twice, as do 0300, 4049,
The
4212, and 8576, while 8686 appears three times.
following pairs occur with very short intervals:

2475 and 2485, 3922 and 3940, 4827 and 4833,


6016 and 6018, 7642 and 7670, 9185 and 9219.
All this should

be borne in mind, as

it

will

probably prove

useful.

We

that the number 0001 represents


next take a small English dictionary
and begin on the real work, making use also of the table
at the end of this volume giving the proportion of words
will

now suppose

the letter A.

in

We

Webster's Dictionary,

initials.

classified

according to their

From

this table

we note that the middle

of Webster's

Dictionary occurs numerically about half-way through L.


But as this bulky tome is rather difficult to handle, we
will use in preference a small dictionary suitable for rapid

reference,

though there

is

the inevitable drawback that

the proportions of the letters vary to some extent with


every dictionary, particularly in the middle of the
alphabet.
1

See p. 138.

CBYPTOGRAPHY

90

We

have begun by supposing that 0001 =A. The


is to look for certain words which one would

next thing

expect to find in most texts, as, for instance, the prepo"


"
sitions "of
and
to," the conjunction
and," the
article
Now, we learn from the table that
the," etc.
'

in Webster's

Dictionary,

divided

theoretically

into

are
hundred equal sections, words beginning with
58
If
the
61
cent,
and
marks.
between
per
comprised
the dictionary were divided into 10,000 parts instead of
a hundred, the
section would be found between 5,800
and 6,100. In the list of numbers in our ciphered text

we observe

three occurring in this section: 6016, 6018,


and 6062. Can one of these be OF ? From its position
we tentatively give the first this reading, and, on looking
"

up

'

of

in the dictionary, our attention

the words closely following


a

'

it:

is

drawn to
"

'

off,"

offend,"

offen-

common

is
military term
6018.
the equivalent of our second presumed
number,
At any rate, the close proximity of the two numbers is a

sive."

Surely this last

promising indication that our surmise is correct.


"
"
and
the
It will be useful now to seek such words as
"

The dictionary table shows T's in Webster to


between 8715 and 9298 (substituting the 10,000 diviAo already noted, the number
sion for the percentages).
8576 appears twice in the text and 8686 three times.
These numbers are outside the T limits, and fall in the
to."

fall

S section.

Nevertheless, allowance has to be

made

for

variations in the proportion of letters according to the

dictionary used, and our cryptogram was probably not

coded from Webster.

We

may, accordingly, venture


to suppose that either 8576 or 8686 represents THE.
[Referring to the text of the cryptogram, we find that
these

two numbers occur consecutively


u

8686, 8576
*

EXAMPLES OF DECIPHEKING
which favours the assumption that the

first

91

equals

TO

and the second THE.


Another number occurring twice is 0300. The dictionary shows A to extend to 6'43 per cent, of Webster,
"
"
and is about half-way through
or 643 per 10,000, and as
the A section, this word may well be the reading of 0300.
There are two other pairs of duplicate numbers 4049
and 4212. These fall somewhere about H, but there are
so many likely words with this initial, such as HAVE,

HAS, HE, HIM,


isolated

etc.,

that

it

favour any

difficult to

is

word without the assistance

of the context.

It will be as well at this juncture to endeavour to


construct a part of the text by using the words so far
Can we fill in the lacunae in
obtained as a skeleton.

the

group

TO THE

OFFENSIVE,

The two missing words


the numbers 1378 and 2799. This

instance ?

the E's.
"

We

have, then,

latter

falls

OFFENSIVE

for

are represented by

doubt-

other
happens
falls
which
under
is
almost
C,
number, 1378,
half-way
between 0000 (A) and 2799 (ENEMY), and the only

enemy

less

in

It

the

word

in the dictionary occurring in this position


may, therefore, not be far wrong
coming."

likely
"
is

offensive."

that

among

We

TO THE COMING ENEMY

reading this group:

OFFENSIVE.
Another group that attracts our attention is AND ....
This is followed by the number 4212, which occurs

AND.

again after the word

'

offensive."

Numerically,

the

nearly half-way to 8576, to which we have


attached the reading THE. Allowing for a small margin,

number

is

we

did in the case of the T's, No. 4212 should coincide


with the beginning of I's rather than the H's. Tenta-

as

tively adopting the

"

pronoun

'

for this

number, we

CRYPTOGRAPHY

92

next note that the number occurring between the two

AND's

is 9712, the highest in the


cryptogram. As this
near the end of the alphabet, the pronoun
you
seems to be indicated, and we have: AND YOU AND I.
'

'

is

The number following

'

'

in the above group is


and
therefore probably a
9576,
highest,
WERE
or
It is followed by
WILL.
word, perhaps
an
D
undoubted
2475,
word, and the next is THE.
What can this D word be? Alphabetically it occurs
"
"
somewhere between coming
(1378) and
enemy
The interval between these two is 1421, and the
(2799).
difference between 1378 and 2475 is 1097, or roughly

the

second

'

three-fourths of the interval.

DI's or DO's.

There

is

This brings us

'

the

among

another number in the text

occupying about the same dictionary position i.e., 2485.


We have, in fact, 2475 and 2485, one of which might

Suppose we give this reading to the second


and look for a word closely preceding it which
The dictionary shows us
diwill suit our context.
"
"
and
divide."
The
we
are
group
studying
may
vulge
therefore be: AND YOU AND I WILL DIVIDE THE."
We must proceed patiently in this way, calculating
intervals and working out the position of each letter.
be DO.

pro

tern.,

'

We

make a false step occasionally, but


established
strengthens our foothold, and the
every word
context guides us more and more surely as we fill in the
shall, of course,

gaps.

In this way, the three numbers 8337, 0702, and 9185,


"
which follow the group
and you and I will divide the,"
are quickly resolved into SUM BETWEEN US, the suggestion in the context, coupled with the approximate
dictionary positions of the numbers, effectively narrowing
our choice. After going on to establish some G and

EXAMPLES OF DECIPHEKING
words, such as
slight

as

HIM, HAVE, and GOOD,

imagination

A GOOD

is

93

only a very

required to convert such a group


TO into " a good opportunity to,"

...
we produce

and eventually
the complete reading of the
cryptogram as follows:
'Mi ... has secured a valuable piece of information
in regard to the coming enemv offensive.
I have been
hundred
to
send
him
five
requested
pounds. It is a good
to
denounce
him.
Do
so, and you and I will
opportunity
divide the sum between us."
Thus, all the words are deciphered with the exception
of the first.
The number of this, 5761, occupies a position
relative to 4833 (IT) and 6016 (OF), its nearest neighbours
numerically, which brings it among the ME's or Mi's.
It is apparently the name of an individual.
We might,
a
of
minute
so
much
the name
by
investigation, identify
i)

as to reveal the nationality of its owner, but it does not


matter much to us. The person who gave me the docu-

ment

to decipher will probably be in a position to throw


light on the individual I am not competent to do so.
In ciphers of this sort a ready reckoner is a useful
;

adjunct to facilitate the calculation of letters, proportions,

and

intervals.

THE SLIDING EULES.

A copy of the Berliner Tageblatt has been put into my


hands with the object of verifying a suspicion that some
hidden message has been concealed therein, the copy
having been intercepted on its way to a quarter believed
to be harbouring

enemy

agents.

Opening the journal, I observe an article with big


headlines announcing an enemy victory. The article is
heavily

marked with red crayon.

Concluding that this

CRYPTOGRAPHY

94

mere

is

column,

blind, I scrutinize every page,

until,

quotations,
of figures

on the

my

attention

marked with

after

is

dots in ink.

Taking a sheet of plain paper,

make

the figures marked in order as follows

all

column

among the Stock Exchange


attracted by a certain number

last page,

a careful copy of
:

1856295937693674187422742555
3754286943673562216626703567
3758396932443268297936634163
3174255928683277811966236328
7629653276316025613680227617
2276227242742563316135591858
4256426718792369387237623663
24681866532473267639612166
Altogether

there

dictionary code ?

are

222

figures.

Have we

here

No, because 222 cannot be divided

however, divisible by 6 or 3. With


dictionary of a million words (including
be
can
constructed, but this would be too many.
000,000)
With three figures one might compose a dictionary of a

by

5 or 4.

six

It

is,

figures a

thousand words (including 000), but


too few for practical purposes.

this

is

obviously

dictionary code being apparently out of the queswe entertain the theory of a system of ciphering
by groups of three figures, each group representing a
letter.
We accordingly make a trial, dividing the figures
tion,

into groups of three, which


lowest to the highest.

we arrange

in order

from the

Of the seventy-four groups thus obtained, we note


that six are duplicated
viz., 166, 267, 276, 425, 532, and
742.
If

we admit

that each of these seventy-four groups of


we shall require a pro-

three figures represents a letter,

EXAMPLES OF DECIPHEEING

95

portion of at least nine E's, and the total number of


repeated groups does not reach that. Then how are
we to get over the great difficulty of identifying the
alphabetical value of the sixty-two groups not re-

peated ?
Let us put aside for the

moment

our notes on the

three-figure groups, after adding thereto the observation


list shows a certain number of groups which

that the
differ

from each other only by single units, to wit 135-136,


:

255-256, 267-268, 296-297, 316-317, 366-367, 591-592-593,


622-623, 762-763, 868-869.

While being almost certain that

much

use to us,

we

forlorn hope.

It

figure groups

may

is

repetitions should

hold

this will not

be of

in reserve as a possible
just possible, too, that these threewill

it

stand for syllables, but even so the


be more frequent.

still

then occurs to us to add the figures of each group to


whether the totals will correspond to the numerical
rank of the letters in the alphabet. Putting the larger
It

see

numbers

to the test,

=20, corresponding
So
ever,

far, so

On trying the small numbers, howa check, there being no A, B, C, D,


in fact, no number produces a lower total
good.

E;
than 7 (160) or 8
without a single E

(035).

Now,

will scarcely

a text of such a length


be found in any language

Western Europe.
It is evident that

different direction.

We
of

(9+3+8)

we meet with

or even

of

we get the equivalents 938


to T; 879=24=X, etc.

we must pursue our


The number 222 is

researches in a
divisible

by

2.

will, therefore, divide our cryptogram into sections


two figures, classifying them in numerical order.

This enables us to produce the following table of


quencies

fre-

96
17 once

CRYPTOGRAPHY

EXAMPLES OF DECIPHERING

97

Out of the above intervals, three are divisible by 4


and three by 3, which would suggest a possible key-word
of three or four letters.
On splitting up the cryptogram
into segments of three (supposed) letters, and arranging
them in columns, we find that col. 1 alone has no fewer
than twenty-seven different numbers, which cannot, theremany different letters. Furthermore,
no number has a higher frequency than three in any of the
columns, and, with the exception of 59, which is the only

fore, represent as

number occurring three times

in col. l,.and might,


the
for
stand
therefore,
E,
frequencies of 3, 2, and 1 are
too evenly dispersed to furnish any clue as to their
significance.

The

solid features to

which we must revert are the

the groups 74.25, 21.66, 36.63, 63.31,


These doubtless represent such frequently
occurring bigrams as TH, ER (or E with another letter),
IN, etc.
An examination of the table of frequencies set out above
repetitions

and

of

22.76.

reveals a peculiarity which may help to put us on the


It will be observed that there is no number
right track.

lower than 17 and none higher than 81.

may,

therefore,

The cryptogram

have been ciphered by means

of the groove

or slide system.

The numerical slide system is constructed as follows:


Take a piece of cardboard, oblong in shape, and at each
end cut a certain number of slits. Into these slits insert
long strips of stiff paper or parchment, some of which
are inscribed with the alphabet and others with a series
of numbers.
Calendars are sometimes made on the same
principle.
By sliding backwards or forwards a slip
bearing the alphabet, the letters thereon are
coincide with different figures on the numerical

made
slips,
7

to

and

CEYPTOGKAPHY

98

by this means a great variety of secret alphabets represented by numbers can be formed.
Where the respective positions of the strips as adopted
at the beginning

remain unchanged to the end

cryptogram, the system

When

is

of

the

that of numerical fixed slides.

the respective positions of the strips are changed

once or several times during the process of ciphering, we


are faced with the system of numerical movable slides.

Let us examine the simpler system, that of fixed

slides

and, since strips of paper or

we

and

easily torn,
of plain wood,
all the rules by

parchment are very fragile


will replace the whole by small rules

two long and two short. We graduate


means of equidistant strokes, and in the
divisions thus made we inscribe, on one of the two longer
strips, the numbers 1 to 50, and on the other 51 to 100.

On one

the

of

short

rules

we

inscribe the alphabet

and on the other the alphabet


The diagram will
in reversed order: Z, Y, X, etc.
better illustrate the part which the four rules can play
the usual

in

order,

in ciphering.

nA|B|c|D|E|F|Q|H|l|j|K|L|M|N|o|p|Q|R|s|T|u|v|w|x|Y|zp
|

51

4 5 6
|

8 9
|

|l

O|n|l

7|1 8|1 9|2 0|2

12

2|2 3|2 4J2 5|2 6|2 7|2 8|2 1

EXAMPLES OF DECIPHERING
As regards the system
farther

in

on,

proximate
numbers.
In what

of

movable

99

rules,

we

give

'

the

chapter

coded

instance,

way must we

entitled

by

adjust our

Ink,"
instead of

Spilt

letters

wooden

rules or slides

cryptogram which we are now


have already noted that the double

in order to decipher the

studying

We

group 74.25 occurs three times in the cryptogram.


Let us now endeavour to ascertain whether it corresponds

bigram TH. For this puradjust the upper alphabet in such a way that T
above 74 in the lower numerical strip. We then move

to the frequently occurring

pose we
is

the upper numerical strip until the 25 thereon


in the upper alphabet.

falls

below

Beginning from the

first pair,

74.25,

which occurs in the

the cryptogram, we decipher AT as the two


The next number, 54, falls outside
letters following TH.
the range of the alphabet. Ignoring this for the moment,
first line of

we, proceed:
wrong track.

KOZMRHDLIP.

Plainly

we

are

on the

Suppose we try another of the repeated groups, 21.66.


This pair occurs at the end of the cryptogram. Adjusting
the rule so that T and H in the upper alphabet correspond
to 21 and 66 in the upper and lower numerical rules, we
proceed to work backward, but are brought to an abrupt
pause by the number 39, which is far beyond the range.

On jumping

over this, we produce nothing more promisthan


WOYE9CTH.
ing
There are other duplicated pairs open to investigation,
but the fact that our first essay, though a failure, produced initially the combination THAT induces us to

TH. This time


restore the rules to the position 74.25
we take the second pair, which occurs in the fourth line

CKYPTOGEAPHY
of the cryptogram.
We get as far as THEKNOW7BLFI,
100

and are again baffled. Yet there are the initial letters
that seem so promising.
Suppose we revert to the combination THAT (presumed) in the

We
"

first line

and

ETATSOTELBA.

get

'try the letters preceding

Keversing

this,

we

it.

recognise

At last we are making definite proBut we have now reached the beginning of the
gress.
text and when we attempt to go in the other direction,
we get a mixture of comprehensible and incomprehensible groups, with occasional numbers which have no
Able to state."

corresponding

and

letter.

Such numbers are

17> 44, 53, 54,

81.

These perplexing numbers must be either punctuation


marks, blank letters, or as we are beginning to suspect
'

changes of alphabet."

In two cases, certainly, such

from non-intelligible groups.


Perhaps Nos. 54 and 81, occurring, as they do, on the
lower numerical rule, are intended to indicate that the

numbers separate

intelligible

groups following are to be read from the lower alphabetic


rule, in which case 17 and 44 will refer to a change to
the upper alphabet.

On

putting this theory to the test,

No. 54, we

group

are

agreeably surprised

PLANISWORKING.

By

to

working from
encounter the

continuing to follow

numbers, we are
in
of
the
text complete, as
soon
plain
possession
very

the

indications

follows

given

by

the

key

Able to state that plan is working well. Only six


Your people must have everything ready
in the know.
Signal three two.
the
further
What
history of this interesting plot was
I am unable to state.
We may at least suppose that

by May

fourth.

EXAMPLES OF DECIPHERING

101

the interception and disclosure of the message went far


to bring it to an untimely end.

CONTRIBUTION TO HISTORY.

The post brings me a letter; I recognise in the address


the handwriting of a well-known historian, with whom,
however, I have not yet been in correspondence. On
opening the envelope, I find therein nothing but a sheet
cryptogram in the same writing.
would have expected a communication in cipher
from such a man ?
Decidedly, everybody is taking
up cryptography nowadays. Let us see what he has

of paper containing a

Who

to say:

ouszehnsobonl
cous eevr
macpks
g

h
i

ox

102
ou-ou

CBYPTOGKAPHY

EXAMPLES OF DECIPHERING

103

(1)

(17)

(33)

(18)

y
u

(2)

(84) c

(8)

(19) s

(85)

(4)

(36)

(5)

n
h

(37)

(6)

g
b

(38)

k
u

u
u

(20)

(21)
(22)

(7)

(23) a

(89)

(8)

fl

(24) a

(40)

(9) o

(10) e

(25) d
(26) p

a
a

(11) r

(27) r

(12)

(28)

u
u

(29)

(30)

u
u
h

age

(13)

(16) s

(14) e
(15)

(31) r
(82) r

(42) a

n
k
n

(43) g

(41)

(44)

bag
v

(45) o
(46) a

Our text is thus arranged in three columns, the


beginning o z n, the second u e s, and the third s
Col. 1 is presumed to have been ciphered by the
letter of the

and

cols.

first
li

o.

first

key-word, which remains to be discovered,


and 3 by the second and third letters

same key-word.
The best way to find the key-word is to ascertain, if
possible, which letter represents E in each column, or,

of the

failing that, to establish at least

Now, although E
in English,

it

is

one letter in each column.

the -most frequently occurring letter


followed so closely by T and A that
is

allowance has to be

made

for

one or other of these preit should

In looking for E,

in a short text.

dominating
be borne in mind that this letter very commonly follows
also that

H,
is

TH is

a very frequent bigram and that

the commonest trigram.


Now it happens that the

repeated in line

9.

first

The word

beginning, and the fact that

"

three letters, o

s,

are

not an unlikely
one of the two letters

the

s is

'

THE

is

CEYPTOGEAPHY

104

highest frequency in col. 3 favours the


that
it stands for E.
In lines 1, 9, and 30
supposition
the letter follows u, whence we may draw the legitimate
inference, subject to correction, that ws=HE and ous=

having

the

THE.
Armed

with these three

letters,

we

will

now

consult

Vigenere's table on p. 155, and endeavour to reconstruct the key-word.


From the capital letter T in the top
line of the table we descend the column which it heads

we reach

till

o,

and

to the left of the line in

which

it

occurs

we

find the capital letter V, which should be the first


letter of the key-word.
As will be seen, the top horizontal

of capitals represents the letters of the plain text, the small


body of the table are the ciphered letters,

letters in the

and the column

of capitals to the left are intended to

the key-word.

This relationship must always be borne

in

mind when ciphering

or deciphering

form

from Vigenere's

table.

Proceeding in the same

way with

and

u, the

second

supposed plain text and the ciphered word


as the second letter of the keyrespectively, we obtain
word, and, continuing, from E and s we obtain 0. Accordletters in the

ing to this, then, our key-word is VNO.


We must next write out a portion of the text of the

underneath, the key-word repeated


of the table we proceed to the
continuously.
with
the
decipherment,
following result
and,

cryptogram,

By means

ouszehnsobonlhhi cma
theertsfagbzqut

cpksc
VNOVNOVNOVNOVNOVNOVNOVNO
Plainly,

a false

it

useless to go any farther; we have struck


and must patiently go over our ground anew.

is

trail,

EXAMPLES OF DECIPHEEING
We

105

cannot resort to Porta's table for enlightenment,

it

being so constructed that the plain letter cannot be in the


same half of the alphabet as its ciphered equivalent, and
this condition

is

not met in cms

= THE.

There

is

another

duplicated trigram in the cryptogram which might represent THE


This occurs in lines 23-24 and 41-42,
i.e., ika.
i being in the middle column, k in the third, and a in the
line.
This group, ika=
does not prejudice, nor is it prejudiced by, ous=
the difference being due to the fact that the letters

column on the following

first

THE,
THE,

did not

to be ciphered under the

fall

There could, of course, be


three alphabets were used.

On

still

same alphabet.

another form for

THE

if

putting ifca==THE to the test by means of .Vigewe produce the key -word PDW, but this

nere's table,

merely proves another failure. There can be no doubt


that three alphabets were used, and, as we are unable to
get any assistance from a key-word, the obvious conclusion

is

means

that

we

are faced with a cryptogram ciphered by


This makes our task

of three irregular alphabets.

more complicated, and we shall have to discover


the meaning of the letters one by one.
We make a beginning by assuming that ous in line 1

rather

columnar table represents THE, and, in addition


marking this word in the margin, we mark T opposite
every o in col. 1,
opposite every u in col. 2, and

in the
to

We must always proceed


the
and marking the
columns,
going
way,
through
appropriate transcription throughout, every time we
opposite every

s in col. 3.

in this

establish the value of a letter.

our clues

and build up the

By

this

means we obtain

solid fabric

of

the

plain

text.

In the present instance this marking, besides bringing

106

out the word


in

CEYPTOGKAPHY
THE repeated in line 9,

two consecutive

lines, 13-14,

the ciphered letter u.

we have already given

shows

to occur

followed in each case by

This cannot be meant for E, since


this value to s; neither can it be

T, which would produce the combination HT?HT.


is limited to A and I, either of which is far

choice

The
more

we assume A, we have in view (T)HA(T)


HA(S) or (W)HA(T) HA(S), but these are both ruled out
by the fact that the supposed T is represented by g,
whereas T has already been adopted as the value of o in

likely

than 0.

If

the same column.

On

the other hand, if we assume the letter following


I, we have the possible group (W)HI(C)H I(S),

to be

and

as neither of the parenthetical letters usurps the


position of T, we will boldly adopt this reading, which

gives us the equivalents:

u=I

We

W, e=C, w=S

(all col.

1),

(col. 3).

have already noted another trigram in the crypto-

gram which appears

likely to represent

THE

i.e., ilia.

but enables us to produce


in lines 23-24 and TH(A)T IS in lines 37-39.

It occurs isolated in lines 41-42,

THE(E)E

The word

"

"

would naturally be expected in a text


the length of our cryptogram, perhaps more than once,
"
the."
The group THE in
and probably preceding
of

preceded by nl, and as this bigram occurs


twice in the text, we may not be far out in ascribing to
It occurs isolated in lines 4-5, but in
it the value OF.

lines 41-42 is

it gives us the
very substantial result THAT
IS (EA)ST OF THE. The parenthetical letters cannot
be WE, because E (col. 3) has been established as the

lines 37-42

equivalent of s, whereas the ciphered letter here is o.


Let us pause here a moment to summarise our
coveries

dis-

EXAMPLES OF DECIPHEKING
Col

o=T,

1.

3 letters

Col

u=H,

1=F, 2
i=A, 4

s=E, 4
o=E, 4

a=E,

i=T,

e=C,

2
2
6

g=W,
u=S,

2.

6 letters

Col

107
3.

s=E, 5
o=A, 4

letters

ii^O, 4

u=I,

k=H,

JJ

Total: 67 letters out of 136, which indicates good proIt is as well to summarise results in this way
gress.

from time to time, as it shows how far the realm of hypothesis is being narrowed down by the extension of that of
certainty.

To show how the summary

will elucidate

such a group

as that in lines 25-28 (the capitals represent the plain


text so far as discovered, and the small letters are ciphers
still

under investigation), apEarTjEge, it will be observed


a occurs in col. 3.

that the repeated ciphered letter

Therefore, it cannot represent any of the letters E, A,


0, 1, or H, any more than p, the second letter in the group,

and occurring in col. 1, can be intended for T, F, A, E,


W, or S. The commonest bigram ending E is HE, and
the commonest trigram THE, so that, as T has not yet
come to light in col. 3, nor H in col. 1, we attach these
values to a and p in the group, which now appears as
THE TrTjEge. The only letter that fits the r sandwiched between two T's, and not yet accounted for, is I.
C,

This enables us to submit the group to the following transTHE TIT(L)E (OF), the parenthetical letters

formation:

requiring confirmation.
The solution is now in sight. The letters remaining
unknown are merely isolated rocks in an ocean of under-

standing.

Thus the group extending from

line

to

108

now

line 17

hEsdly.

CKYPTOGEAPHY
appears as: THE CevITAx OF WHICH

IS

The second word can be no other than CAPI-

TAL, while

as to the last, the

name

of a capital having

which the second is E and the fifth I, there


need not be much hesitation in pronouncing it BERLIN.
The cryptogram holds no further terrors for us.
six letters, of

BERLIN makes us think of (P)E(U)S(S)IA (lines 35-37),


and eventually we have this table of all three alphabets:
Plaintext:
Cipher Col.
,,

,,

,,

,,

ABCDEFGHIKLMN PRST UVWY


alnprvxzymwsuokb -g
ua-d -eg -stiv
o
3:op-mseckurjbfnvhgai---

l:ide
2: eh

The plain text proves to be as follows


The Margrave of Brandenburg, the capital of which
is Berlin, has no right to assume
thereby the title of king.
He is king only in Prussia that is, east of the Lower
:

Vistula."

NOTE BY AUTHOR.
not a

This statement by

my correspondent, who

is

man

to assert anything lightly as a rule, aroused my curiosity.


Upon investigation, I find he is right, as is borne out by the admis"
"
sion of German jurists who are regarded as authorities in Prussian

public law: Hermann Schulze, Ludwig von Roenne, and Ludwig


Bornhak, who, with considerable reticence, acknowledge that the
Margrave of Brandenburg is only, and has never been other than,
'

it

king in Prussia." The immense kingdom of Prussia, as we know


is only a fiction; its existence has no serious historical or

to-day,

juridical basis.

SPILT INK.
This morning's post brought me a letter from a wellprofessional man who has been utilising his spare

known

time in constructing a safe cipher. He sends


specimen, and warns me that he has submitted
several amateurs,

who

failed to decipher

it,

me
it

a
to

the last re-

EXAMPLES OF DECIPHEEING

109

turning it intact with the remark that it was so much


"
Perhaps you will have better luck," adds
spilt ink."
'

correspondent, no doubt smiling up his sleeve as he


wrote.

my

Let us glance at this cryptogram.


in deciphering
false

it,

I will frankly

If I

admit

do not succeed
it

without any

shame:

sseguhckxzdgzggzszdsj nfj
wphfxqugogghznyslpaysfim
owl w n n o d
hvmpxkqbz h h
nrexbemf
xf
z

iznhkd

yyt

Total: 93 letters.
find that the

and

z,

cl

I prepare a list of frequencies,

most numerous

which each

and

letters in the text are h,

figure eight times.

Can any

s,

of these

represent E ? Normally, there should be about a dozen


E's in a text of this length.

The next letters in order of frequency are g (seven times),


These figures are
/, n (six times), and d, x (five times).
too close for a simple alphabet cipher.
Besides, what
double letter could the initial ss stand for ? If we assume
a name beginning LL, they would have to be followed
or E, and as e, the cipher equivalent, appears only
by
three times in the whole of the cryptogram,
to go any farther in this direction.

it is

useless

It would appear that more than one alphabet has


been used, and that we may have to seek a key-word.
The procedure to ascertain this has already been described in the chapter on
Ciphering by Means of a Key'

'

Word."
It

will

be noted that there are several duplicated

pairs of letters in the text, zd, for instance, occurring three

times.

Accordingly,

we

insert a stroke

between the two

CEYPTOGKAPHY

110
letters in

each pair and count the intervals, making the

following tabulation:
3

3 letters, or
8

gz-gz
zd-zd

2x2x2

3x13
,,2 X 2x5

39

gg-gg20

ys-ys 5
zn-zn 37

37

Thus, there are two intervals having the factor 2, two


with the factor 3, and two with 5. This absence of a

predominant factor does not augur well for the key-word


theory, and the experiment of dividing the text into three
columns on the ground that, of the three equal factors,
3 is the most likely to indicate a key-word, if any, leads
to no result.

The cryptogram contains


gg (twice), nn, dd, Jill, and

several double letters

ss,

yy.

the fact that the gg in one case


betic sequence,

and in

like

is

viz.,

A close scrutiny reveals


followed by h in alphais followed by o.

manner nn

This detail gives me a clue to the right track. The system


"
of ciphering used appears to be that known as the
Saint

Cyr Slides."

By means

of the Saint

six different alphabets.

slides we can obtain twentyAnyone can make these slides.

Cyr

is
necessary is to obtain two rules or strips of
one long and one short. On the short rule
wood,
plain
mark equidistant divisions, and in them inscribe the

All that

twenty-six letters of the alphabet. Proceed in the same


way with the long rule, but with the difference that two
consecutive alphabets
i.e., fifty-two letters, A-Z and A-Z
must be marked here, in order that, when the smaller
rule

is

moved up

or

down

in juxtaposition with the

EXAMPLES OF DECIPHEKING
will

always be in contact with twenty-six letters

longer,

it

on the

latter, as

111

shown:

CEYPTOGKAPHY

112
ciphered ?

If

we knew the

secret

the

of

first

letter,

everything could be unfolded mechanically; but


and this is the mystery of the " spilt ink."

we do

not,

However, there are ways and means. When the groups


ggh and nno attracted my attention just now, I consulted
a notebook in which I record rules which
appear deducible from a long series of observations on
ciphers, and
found an entry entitled: " Cipher established by means
of Saint

Cyr slides, with automatic change of key at


every letter." This is what I read:
A. When two like letters occur together, the second
represents the plain letter

B.

When two

A (mm=?A).

letter in alphabetic sequence, the

in the trigram represent

by the next
second and third letters

are followed

like letters

AB (mmn=?AB).

When an

a occurs in the ciphered text, the letter


which follows is identical with the corresponding plainC.

text letter (ae=?E).


I

have noted some further

rules

on the

these three will be ample for our purpose.


them to our cryptogram:

Kule A.

,,

nn

,,

,,

,,

ad

,,

,,

,,

,,

riri

,,

,,

li

,,

A
A
A
A
A

. .

A.

In

ss

the second

99

5 ,

firi

yy

equals

r,

fi

subject, but

Let us apply

A
]

Rule B.

In ggh the last two letters equal

Kule

In ay the letter y equals

nno
C.

,,

AB
AB

By these rules we would appear to have accounted for


twelve letters (gg occurring twice).
Using these as a

EXAMPLES OF DECIPHEKING
check,

one

it

now remains

113

to decipher the remaining eighty-

letters.

We

do not know the value of the

first letter, S,

but that

does not matter; in the dozen letters presumed to have


been established we have a plentiful choice of startingWith the Saint Cyr slides to our hand, we select
points.

beginning the

for a
letters

i.e.,

ggh.

group containing two known

first

From what

that the ciphered letter h

is

the

we infer
new alpha-

has gone before,


first letter of

The cipher being represented by the long rule and


the plain text by the short one, we slide the latter until
A thereon is above H in the lower rule. We now have
to see which letter on the upper rule corresponds to
the letter following h in the cryptogram, of which a
bet.

section

is

reproduced for convenience:

gghznyslpay
The

letter in question is

to S.

Adding

this to the

which we find to correspond


two letters already known, we

z,

obtain three consecutive plain-text letters, ABS.


The letter z now becomes the ciphered equivalent of
A in a new alphabet. Proceeding as before to adjust the

we identify n as the equivalent of 0. Again changthe


ing
alphabet by giving the value of A to n, and continuing similarly with each letter, we decipher the above

rules,

group as the word ABSOLUTELY.


The whole of the text is thus deciphered quite easily,
with the exception of the first letter. The cryptogram
begins with ss, and we know that the second s stands for
A.
Further, we know that this is by virtue of the fact
that the value of the first s was altered to A, but we nave

no means of knowing what the original value of this


initial s was.
However, we have the context to guide us
8

CKYPTOGEAPHY

114

where our formulae are impotent, and effectively the series


from the second prove to be, AMCONVINCED,

of letters

we may

so that

'

am

safely conclude that the mysterious first


is as follows

The complete transcription

letter is I.

Accordingly, I

undecipherable.

vented

convinced that the present system

am

is

absolutely

proud to have

in-

disappointment of

my

it."

With the view

of softening the

whom
am able

communicated the
I
to
inform
him in a covering
deciphered text,
note that this system offers safeguards by no means
am appears
negligible, since, for example, the word
in two different disguises, se and ht; the 'in' of the
words
-invented
convinced,"
accordingly," and
correspondent,

to

have

'

'

'

is

'

'

'

dissimulated under three separate forms: kx, xk, and/s;


"
in
ent
invented
is ciphered
and
present

while

as iwp

'

'

'

and rex

'

'

respectively.

AN UNDECIPHERABLE

SYSTEM.

In the spring of 1917 the post brought me a cryptogram


to which was attached a visiting card with the words,
in the handwriting of a friend:

was

You

are fond of solving

I wish you
Here you are, then
The cipher text, which contained fifty-two letters,

difficult

joy."

'

problems.

as follows:

ylirxqjzmp

tcmovzngrqlfkve

wnodsds ckutiuhpfywbhegbxja
began by calculating the letter frequencies, and, to
in fact,
stupefaction, found two a's, two &'s, two c's
two of each letter of the alphabet, neither more nor less.
I

my

Only one group, ds, was repeated, and, that being the
it was useless to seek a
key-word.

case,

EXAMPLES OF DECIPHERING

115
'

diThere could be no question either of a grille or


In any case, these two letters of every kind
were a strange coincidence, though instances almost as

viders."

I reflected
curious are encountered from time to time.
on the possibility of a dictionary code. There are conventional codes composed of three-letter groups: aab

wkf, etc.

the letters of the alphabet in threes, a


of groups can be obtained, sufficient to

By combining
large

number

replace the w ords of a considerable-sized dictionary.


Thus, the letter A, followed by one other letter, gives
twenty-six different groups, and each of the other letters
r

the alphabet, followed by another letter, similarly


In this way, 26x26=
yields twenty-six combinations.
of

676 different groups of two letters can be formed, and

676x26=17,576 groups of
I might have made some
but for two obstacles:

(1)

three letters.
investigation in this direction
fifty-two letters of the text

The

were not divisible by three. One of them might be a


blank letter, but which ? (2) In whatever way the text
was divided into three-letter groups, these were all
different,

and

needed at

least

one repeated group to

serve as a base or starting-point.


I thought of a code composed of four-letter groups;
52 is divisible by 4, but the sectioning of the text into
four-letter groups failed likewise to furnish any guide.
Only two groups began with b bheg and bxja. But all

my

efforts

proved

futile.

could not identify these

groups approximately with such frequently occurring


"
"
"
"
words as
and,"
be,"
but," and others with
at,"
initials in

the early part of the alphabet.

Not being able to obtain the


method of ciphering employed,

faintest clue as to the


I

called

on

my

friend,

CRYPTOGRAPHY

116

informed him

of

my

lack of success, and begged

him

to

me

with the key.


acquaint
Quite simple," he said.
'

"

This

is

how

wrote that

cryptogram: I cut fifty-two slips of paper, on each of


which I inscribed a letter of the alphabet. After using up
twenty-six, I repeated the alphabet on the other twentysix.
I dropped the whole fifty-two slips into my hat, and,
after shaking them up, took them out one by one haphazard, and noted them down just as they came to hand.

The text thus formed I sent


Then it has no meaning

to you."
!" I

exclaimed.

"

'

Of course," he replied, and I must ask you to forgive


little trick; but you are so clever at cryptography
that if I had submitted you a text with any meaning at

my
all,

you would probably have deciphered

quickly

it

far

too

!"

THE ANTIQUE DEALER'S.

had asked me to meet him at the tramway


was there to time, with a minute to spare,
and was first. While I walked to and fro, with an eye
on the various approaches, the long hand of the clock
tripped jauntily on its way, marking off the minutes in
silence.
At the end of a quarter of an hour I had decidedly
I almost
lost all right to repeat the famous remark,
had to wait," attributed gratuitously to Louis XIV.,
who appears to have said quite the opposite. 1
How should I pass the time ? There was no newsfriend

terminus.

vendor.

news

Besides,

when one

is

at certain regular hours,

in the habit of reading the


it is just the same as with

meals
1

kept

"

one has no appetite between.


Why scold him ? Don't you think he

me

is sorry enough to have


waiting ?" (CEuvres de J. Racine, Hachette's edition, 1865,

vol. v., p. 125).

EXAMPLES OF DECIPHEEING

117

But I descried an antique dealer's shop across the road,


and as from that spot I should easily see my dilatory
friend or be seen by him, I went over to examine the
articles displayed.
My eye quickly fell upon the prices
inscribed on the labels, and, succumbing to the fascina-

innocent hobby, I set myself the task of


the
values of the letters which took the place
deciphering
of Arabic numerals.
Drawing out my notebook, I took
of
a
of
the
articles exposed for sale in the
note
number
tions of

my

four or five windows, as well as their mysterious prices


viz.
1.

2.
3.

Bronze statue (Psyche emerging from bath)


Incense burner
Coloured engraving (national costumes) -

z.r.p

i.mp.p

m.mp.p

much

4.

Large double mirror, old frame,


rubbed

5.

Inlaid card table

e.p.p

6.

Iron dagger
Small picture (glacier), white wood frame Small engraving (Marie Antoinette), black

mp.p

frame
Small picture (The Flirt), worn gilt frame 10. Black and gold metal tray, flowers in centre
11. Old barometer
12. Old picture (rustic scene)

mi.z
mi.z
mf.z

7.
8.

gilt

mi.mp.p

9.

mr.p

m.r.p
f.z

13.

Four old engravings: the four

m.mf.z

14.

Grandfather clock Card table, with inlaid draughtboard

mr.p.p

15.
16.

Glass cheese dish

b.r.p
f.z

Small Flemish painting, copy (? of a copy)


18. Large trunk, much patched
19. Devotional picture, cloisonne worked on
17.

20.

wood
Head of

21.

Concave

22.

Large oval metal tray,


in centre

lion in bronze (door-knocker)


shield, bas-reliefs, in gilt frame

i.mp.p
z.r.p

mf.z
mi.z

....

Watteau subject

z.r.p

f. z

CEYPTOGKAPHY

118

23. Small mirror, large black wooden


24. Mirror, brown carved frame
26.

Tin candlestick
Oval silver tray, tarnished

27.

Bronze bowl on three

25.

feet

frame

r.p

m.mp.p

Large bronze hand lamp


29. Small china vase, coloured and

z.p

28.

30.
31.

(the Signing of

Engraving
Engraving (Friends
;

Library of
49 plates

32.

Small

silk

i.mp.p
gilt

Magna

Charta)

m.mp.p
-

r.p

Death)

mp.p

Famous Men," volume with


m.p.p

33. Silver bell


34.

till

b.p.p

m.mr.p

mat,

m.i.z
silver fringes

o.p

35. Silver strainer


36.
37.
38.

z.p
m.i.z

Old decanter, silver stand Locket, with cat's-eye and amethysts


Chased silver egg-cup

Copper seal
Old silver chafing dish
41. Old decanter
42. Liqueur stand, with two flagons
43. Old beer mug, coloured stoneware
44. Bronze medal set in ring of chased silver

mf.z
mf.z

39.

m.m.p

40.

o.mp.p
mi.z
b.b.p

mp.p
-

o.p

thing I noticed was the large number of p's,


a
of all the letters I had noted.
It occurs
third
reaching
but
the
and
never
shillings,
among
pence
among the

The

first

pounds.

We

know

the role played by zero in arithmetic


numbers
come into play. Zero, the value of
when high
which is defined as nil, then assumes an extreme imthat is,
portance, provided it appears on the right side
all

the figure that best gives the notion

to the right.

This

of infinity,

repeated to a sufficient extent.

if

is

If I

were

a mathematician and were commissioned by the Board


of Kesearch, I would willingly write a book on The Value of
Nauylit.

EXAMPLES OF DECIPHERING
From

the above

it

follows that, our letter

119

p never

occurring to the left of any of the prices, but being often


repeated to the right, we may boldly conclude p=0.
Next we observe that the only other letter in the pence

column besides p

these two sharing the column in the


proportion of two-thirds and one-third respectively. The
conclusion is fairly obvious that z stands for 6.
is z,

Of the bigrams occurring in the shilling column, the


always m, and since numbers in this column
do not go beyond the teens, m can mean no other than 1

first letter is

may be noted that there are only


Parenthetically
nine different letters in the price list, so that one figure
it

unrepresented. This is most likely to be


Nine
rarely seen in prices of antiques.
or 19 shillings or pounds is very unusual.
Prices
hovering in the region of the 9's are wandering asteroids

out of the ten


9,

is

a figure that

is

which usually succumb to the minor attraction


smaller planets 8 or

of the

7, or the increasing attraction of the

larger planet 10.

We

have presumptively disposed of the four figures 0,


and 9, and almost certainly know that the price of
the coloured engraving (3), tarnished tray (26), and china
1, 6,

vase (29)
(32),

is

10s. each; of the

1; of the

copper seal

"

(39),

'

Library of Famous Men


1 Is.; of the dagger (6),

engraving (31), and beer mug (43), 10s. each; and of the
bronze bowl (27) and silver strainer (35), 6s. each.

The incense burner (2), small Flemish painting (17),


and bronze hand lamp (28) are all the same price that
The
is, 10s., plus a number of pounds indicated by i.
double mirror (4) is the same price, augmented by 10
i.e.,

Further, the engraving (8), picture (9),


(20), and decanter (41) are each priced
These articles seem to me quite dear enough

mi.mp.p.
door-knocker
at mi.z.

CEYPTOGEAPHY

120

at 12s. 6d., so that there is

value than 2 to

no need to ascribe a greater

i.

This brings the price of the incense burner,


and that of the double mirror to 12 10s.

2 10s.,
are also

two items

stand (3G)

The

silver bell (33)

priced at

letter

1 2s.

etc.,

and decanter with

to

There
silver

6d.

shares with r the third place in order of

frequency among these prices. The bronze Psyche (1),


trunk (18), and shield (21) are each priced at 6 plus r
shillings; the picture of a glacier (7) is 10 plus r shillings;

the barometer (11), 1 plus r shillings; grandfather clock


(14), 10 plus r pounds; brown-framed mirror (24), 30 plus

and the candlestick (25) and engraving


Charta
So ubiquitous
Magna
(30), each r shillings.
a letter can scarcely be intended for anything but 5.

shillings;

'

'

Certainly the two last-named objects would not fetch


5s. each, while such quotations for the other
articles as 15s.,
1 5s.,
1 15s.,
6 5s., and
10 5s. are

more than

commonly
5 to

seen.

We

therefore

now

attracted to the letter

attach

the

value

of

r.

Our attention

is

b.

There

a mirror (23) at b pounds, a liqueur stand (42) at b


guineas, and a card table (15) at b pounds 5 shillings.

is

The first named is a very woebegone-looking object, and


must be regarded as dear at 3. The card table is more
presentable at 3 5s., but however good value this may
be, the liqueur stand at b guineas is an obstacle to the
placing of b at any higher value than

we appraise

3.

Accordingly

b at 3.

Summarising, w e have established six out of the nine


Those remaining to be discovered are 4, 7, and 8.
digits.
r

The

/ occurs among the shillings, and is a] ways


accompanied by z (=6) pence. This latter factor induces
letter

EXAMPLES OF DECIPHERING
us to ascribe the value of

On

7,

121

rather than 4 or

8,

to /.

assumption, prices are as follows: Black and


gold tray (10), devotional picture (19), Watteau tray (22),
locket (37), and chased silver egg-cup (38), 17s. 6d. each;
this

rustic scene

(12},

cheese dish (16), 7s. 6d. each; four

engravings, the four (13), 1 17s. 6d.


The letter o occurs three times. The silk

mat (34) is
marked
b.z., or 3s. 6d.,
o.p.
originally
which price has been crossed out. The value is scarcely
marked

It

was

have jumped suddenly to 8s., so that the only


alternative is 4s.
Assuming, therefore, that o equals 4,
the price of the bronze medal (44) is also 4s., and that of
likely to

the chafing dish (40) 4 10s.


The prices of all the items have thus been established,
with the exception of the inlaid card table (5). This is

marked e pounds, which must mean either 8 or


The letter e occurs nowhere else, so we have no means
drawing

any

reliable

inference.

9.

of

Compared with the

other card table, which appeared fairly good value at


3 5s., the present article is relatively not cheap at 8.

While I was debating within myself whether to invite


confirmation from the dealer, who had come to the door

and was regarding me with an inquisitive air, a commotion


took place behind me, and my friend, a good hour and a
quarter late, greeted me in breathless tones:
So sorry, old fellow; but, you know*

'

me

If you had kept


Yes, yes; I know," I interrupted.
waiting ten minutes, I should have been annoyed;

but people who are more than an hour late are assumed
to have been victims of an accident, and they are always

excused in

wasted

my

advance.
time."

But don't worrv.


w

have not

PART
LISTS

III

AND TABLES

NOTE. --This third part consists of a series of calculations


and combinations in English and

of letter frequencies

certain foreign languages.

Of the practical value of these

lists,

compiled as a

numerous experiments, there can be no doubt,


but the fact must not be lost sight of that they constitute
only one of the factors which the decipherer must take into
account if he would push his investigations to a successful
issue.
Cryptograms are often encountered in which the
normal frequency of letters has been deliberately upset.
The second factor is untiring effort, supported by

result of

persevering study.
is flair, or insight.
This need not be
or
in
the
nature of a lucky
instinctive
as
purely
regarded

The

third factor

reasoned and discerning ingenuity plays a large


part here, as well as the exercise of that gumption or
common sense which enabled Christopher Columbus to

gift.

stand an egg in a position contrary to the laws of physics.

ENGLISH.

I.

Order of Letter Frequency.

According to Edgar Allan Poe

EA

DHNB
-

U Y,

etc.

According to Vesin de Bomanini:

LC

W U M,

etc.

122

ETAONIBSHD

AND TABLES

LISTS

123

:ETOANIBSHDL
UCMPFYWGBVKJXZQ.
Telegraphic frequency (Hitt) EOANIKSTDLHU
CMPYFGWBVKXJQZ.

Normal frequency table

(Hitt)

Order of Frequency of Final Letters.

According to Valeric
also English.

ESDNTBYOFA,

etc.

(See

III.)

The Commonest Bigrams

(Valerio).

TH, HE, AN, EB, ON, BE, IN, ED, ND, AT, OF, OB,
HA, EN, NT, EA, etc.
Frequency of Double

EE, 00, FF, LL,

Letters.

SS, etc.

According to Valerio: SS, EE, TT, LL,

MM, 00, FF,

The Most Frequent Two-Letter Words

OF, TO, IN, IT,

IS,

etc.

(in Order).

BE, HE, BY, OB, AS, AT, AN, SO,

etc.

ENGLISH.

II.

The Commonest Trigrams

(Valerio).

THE, AND, THA, HAT, EDT (triED To, carriED The),


ENT, FOB, ION, TIO, NDE, HAS, MEN, NCE,
OFT, STH.
The Commonest Three- Letter Words.

THE, AND, then FOB, ABE, BUT, ALL, NOT,

etc.

The Commonest Four-Letter Words.

THAT, WITH, FBOM, HAVE, THIS, THEY,

etc,

CEYPTOGBAPPIY

124

Words
A,

I,

of

One

Letter.

0.

Proportion of

(Valeric): 13 per cent.

Proportion of vowels (Valerio): 40 per cent.

ENGLISH.

III.

1
(COMPILED BY TRANSLATOR.)

Order of Letter Frequency in Relation


of Letter in Word.
Initial letters

Second

Third letters

TAOMHWCIPBES,

letters

EIAUNB

ABN

letters

Antepenultimate
Penultimate letters:
Final

letters:

lish.

to

Position

etc.

T, etc.

I, etc.

T E

AHN

0, etc.

ENAEHILCO, etc.
ETSDNBYG, etc. (See

also

Eng-

I.)

Consonant Bigrams

at the

Ends

of

Words (Order

of Frequency).

NG, ND, NT, DS, KS, ST, TS, TH, HT, ET, SS, CT, LL,
LT, GH, SH, CH, DD, LD, LS, NS, EN,. ES, WN,
FF, LP, MS, ED, EL.

ENGLISH.

IV.

Final Bigrams.

An

English text of 1,000 letters contains, on an average


(excluding two-letter words)
:

words ending HE.


10 words ending ED.
11

This and the following sections up to page 138 have been


piled specially for the English Edition.

com-

LISTS
8
7

AND TABLES

125

words each ending EE, NG.


words each ending OK, KE.

each ending AT, ND.

ending NT.

words
words
4 words
3 words

ending LY.
each ending AN, DS, EN, ES, LE, ON,

BY, SE,

TY.
2 words each ending
1

AD, AS, CE, HT, ID, IS, KE, KS,


ME, NE, OT, OW, BT, SS, ST, TS, TH, VE.
word each ending AL, AP, AE, AY, CH, CT, DE, EE,
EM, ET, EW, EY, GE, GH, HY, IG, IL, IN, IK, LD,
LL, LS, LT, NS, NY, OM, OU, KN, KS, SH, TE, UE,
UL, UK, US, UT, WN, WO, YS.
Final Trigrams.

An

English text of 1,000 letters contains, on an average

(excluding three-letter words):

words ending ING.


words each ending ENT, HAT.
2 words each ending AVE, ERE, GHT, ION,
5

IED

NDS, PLE, KTY, VER.


1

word each ending ACT, AID, AND, ANT, ART, ATS,


EEN, END, ERY, ESS, EST, HED, HEN, HER,
HIS, ICE, IES, ISE, ISH, ITH, LLY, LOR, NCE,
NED, NTS, OKS, ORE, RED, TED, TER, UND.
Initial

Consonant Bigrams (Order of Frequency).

TH, PR, WH, CH, FR, SH, TR, CL, SP, CR, PH, PL,
BE, GL, SO, SM, ST, WR.

CKYPTOGRAPHY

126

ENGLISH.

V.

Like Letters at Equal Intervals (separated by two


Letters).

A
Ar A
A
A
A
A
p
sA

b b
b i
b 1
b o

d i
d 1
d r
d u
d v
d w

Acy
A
Ative
Ard
A
-

heA
q u
gr

A
A
heA
A
A
p A
A
mA
A
A
a v A

Air

Aid

-A

Ant

A
A

c r

Age

Ament

g e
g g
g n
g r

h
i

And
Angle

A1
A

Ay
A
A
A

Ad

Able

m Ant

-A

complA

Ant

clA

breAkfAst
-A

dAmnAble
cAmpAign

A
A
terrA
mA
A
A
J A
c A
c h A
A

evAcuAte
r

AmbAssador
A mi Able
-A m m A

A-

t r

w Ays

Au A,

c
f

g
i

An
A
Act
Agonist
Ar y
As
A n
A
i

A-

A
A

Age

A
A

A
Ake
-

stArvAtion
A
A
A

^_

Able

A
A

q u

-A
-A

p 1
p p
p t

remArkAble
ArmAment
cArnAtion

signA m An
1

n d
n e
n i
n s
n t
n u
n v

whA r
A r
AT

pA
-A
vA

mwAy

devA
c

A
A

A
Amed
A
A

Ate

u Al

As

LISTS
-

A
A

u d

A
A

AND TABLES
E

restAurAnt
A

v
-A v

A-

drAwbAck
A
1 A

x p
y

Ay
m An

Bar
B
B

o
r

er

B-

mB
i

suBurB
coBweB

Calculate
(J

127
a

chEapEr
E

a r

E
E
E
E
E

a t
a v

E-

rE
E

d e

EasE
b

c r

c t

cl

E
E
E
E

cl

Em
E

expEdiEnt

neEdlE
WEdnEsday

128

CEYPTOGKAPHY

gEomEtry

AND TABLES

LISTS

coNveNe
-NviNce

s m Iss
-TssI

d I

-I s t I

Britain
K

h a

Ki

booKmaKer
singuLarLy

Leu

Late

pOisOn
f
-

coMmeMorate
n

Mon

MumMy
N c a Nd
NceN

N c N
N c o NN daNN d e N
N d N s
N d o N
N f a Nt
N geN

N
N

N
N

-0

t t

N
N

u d

bO

teN
t

00-

tOpmOst
-0 p p 0fOrbOre
fOrgOt
OriOus
fO r
Orn
enOrmOus
-OrpO
r r 0fOrsOok
-OscOexplOsiOn

o o
o u

reNowN
N
N
N
N

-0 p h 0-

caNnoN
aNoiNt

cOnvO-

NjuNction
N kiNg
N a Nd
NmeNt.
1

-0 n

eNgiNe
i

cOmpO-OndOerrOneOus
cO

mm

e s c e

cOmfOrt

Ong

OctOOdiOus
lOgwOod

-ceLluLar
M a d Man
M a M

ObsOlete
Mor c c
atrOciOus

KicK

ca

129

-pO

OtiOn

nOxiOus
x

w Oo

d
9

130

bO

y b

CBYPTOGKAPHY
diS
Ood

PalPable
PamPer
Pan Per
P
P

e o
e r

P r e
P r o
P u 1
P urn

P
P
P
P

PurP-

fEatEicide
extEaoEdinary

-EdeE
EeaE
caEeeE
E

waE

peE

a
o

laE

-E

g e
i

diS

g u St
S h e S
h u S i aS
S
Sh
i S s
-S

ma
m

-S n e Ss

conS
d

puE
-E
hoE

p e
p o
r

cuE

-E

o
e

depaE

E
Et
E
Eace

b e

T
sT

a c
a n
a r

T
T
T

TasTe
TauT

T
-T
T

e c
e

n T

e s

TexT
T
ouT
ThaT
f

paTheTic
wiThsTand
i

Te

Th

E-

-T

Ee

ouT

-T
-T

Ey

FebEuaEy
foEwaEd
aS

S u b S

E
E

-EpaEt

S p o S
s

Em

coEneE
E o a E
-EogE-E

E
Ee

E
E

poSseSs
a S
S t
chaStiSe

treSpaSs
deSpiSe

pEimEose
-E o E
-E k
-E m

u Ss

asSetS

Pie
P-

PorPoise

S e a Son
S e n S

St os

n T

o r

T
T
T

TouT

ouT

p u T

LISTS

AND TABLES

Te

r a

-TreTTion
-T r
beTroTh

U m

U n

TruTh

rUptUre

tU

aT

Tudo

pU

T
T

f
i

sU

b d

Ue

c t

U-

gU

Ural

Zag

ValVe
VelVet
VolVe
W a y Ward

sUccUmb

sUffUse

-U

Ulent
Ue

mUscUlar
MU e Urn
brUsqUe

siTuaTion
T u
T u

Ucky

Unsuitable

ouT

r a

Ur

UnhUrt

131

o Us
U t Uh U m b U g
1

ENGLISH.

VI.

Like Letters at Equal Intervals (separated by


Three Letters).

AbeyAnce
hAbitAble
lAborAt
AbreAst
A

Ad

AbstAin

-A b ul A-

coAchmAn
blAckmAil
bAckwArd
ActuAl
erAdicAte
AdorAble
heAdquArters
AeriAl
-AginAcoAgulAte

dAhliA
r

w Ay

AcquAintAnce
stAircAse
chAirmAn
mAlefActor
-A

-AlisAtion

-A

AlphAbet
AlreAcly
steAmboAt
nAmesAke
AminAt
finAnciAl

lAndmArk
chAngeAble

CBYPTOGKAPHY

132

lAnguAge
g Ang w Ay
mechA n c A
An mA
orgAnisAtion
A n n u Al
AnomA y
mA n
Aught er
trAnspArent
substAntiAl
mAnufActure
dilApidAted
i

teC

h n i Cal
C h u r C h
C i e n C
C i f i C

neCklaCe
C

s 1

A
A
A

p i
p p
p p

e
1

A
A
A

A
A
A
A

r o

r r e
r r

r t

C u

p h

s s

A
A
A

g e

A
1A
A

e r

coA

of

Arms

A
A
nA

v
v

e r

A
Age
Ate

s t r

c
r

Al

E
E
E
mE
IE
cl E
f E
E

volCaniC

Cy

De

e s

Deration

a
a
a
a
a
a
a
a

D
D

E
c h E
E
g
g r E
g u E
n s E
r
Es
r n E
b

rehEarsE
E
E

C h a n C e
C h a r Coa

WeDnesDay

AvocAtion
A v o w Al
AwkwAr
a B s o r B

r a

e c

reAsonAble
A
A

D
D
D
D

Dr e
Dw

Al"

CrutCh
r
C
CubiC

eleCt

Able

Ar s
Age

CounCil

u d

mA

Ch

C o n

pArliAment

chArcoAl
chArgeAble

chA

ClutCh
C o a Cellar
C o m Cal

a
a

-E b

dE

E
E

s t
t

Er

h E
t

E
E

mEn

corrE'ctnEss
s
1

E
E

u
u

r
t

E
E

AND TABLES

LISTS

E d i a
-^-E d u c
schE d u 1

dE

-E
-IE
b
d

E
E

a c
f o r
g a t
h a v
f

E
E
E
E
E
E
E
E

va

q u

E
E
f E
b E
r

1 i

Er

-E

o p
o v

bE

Ed

u d

nv

E
E
E

E
Em
Ent
E

m n
-E m s
rEmunErat
parEnthEsis
sevEntiEth
i

En
cE n

nevErthEless

EEn

mE

t i
t

o k

En

v
v
v
v

Ent

s E
E
v E
E vo kE
i

benEvolEnt
d
b

E v
Ew

o t
a r

E
E

ExchEquer
E x
E x
F e
G o

rE
E

cxpErimEnt
dE

E
E

t r

E n u mE rat
p E o p 1 E
r E p 1 i E d
-E p o s E
s c E p t r E
-E p u t E
E r a g E
E r a s E
-E r a t E
o v E r d u E

t c

EvanEscent

in b

E
E
E
E

invEstmEnt
r E s u m E

-E

Er
E
n Eate

dE
e

rE

dE
-

forEignEr
E
E
d E

133

Gr

H
H
arcH

t r

a r
i

11

u d

a r s
a t c

bi

Ess
E me
Fu1
G

Ge

H
H
Hop
H

Heat
H e g Ht
H
c H
tHo ugH
i

I e

t I

-I g h

com pi

Ing
Ion

pilgrim

a 1

11

Ew

wholEsalE

Implicit
I

mp r

n q u

E
E

v
c

CKYPTOGEAPHY

134

Inspire
Instinct
Intuition
diminution
Invoice
circuit
mischief
d

I s c r I

biscuit
d I

N
N
e N
a N
N

-r
c o

capitalist
-Itatlon

-N

stocKbroKer
KnacK

coN

Kno c K
K

L
L
c L
L

n u

a b e
a n d
e a r
e g a

L
L
Ly
L

symboL

c a

L
L
L
L
L

i
i

k
v

e
e

o c a
o v e
o y a

L
L
Ly
Ly
L
Ly
L

LuckLess
absoLuteLy
MadaM
comMandMent
M a x Mum
aMazeMent
i

aM

n d

Ment

soMetiMes
M n Mum
MonuMent
MoveMent
i

meN

N
c h a N t
e N t
c
d e m N
d
g N
d me N t
e m e N t
f r o Nt
c e r

o
o

m N
i

w Ng
i

deliNqueNt

siLentLy
L

EN g a Nd
saNguiNe
cogNisaNce
pheNomeNon

compLeteLy

M ent
Ment
o M

p
p
p

aMuseMent

c o

t r

- M
coM
s y M

s i

-N
-N
-N

iN
N

h
i

s t
t

t e
t

Ne
N
o N
a Nt
N
r N
o N
i

b v

c t

u
n

c
r

g a

tOils Ome

whOlesOme
sOlilOquy
sOmebOdy

meh w
cOmprOmise
s

c
c

n g
n t

meration
-

anOnymOus
p

LISTS

AND TABLES

fOregO
hOrizOn
pO

r t f

r t

Olio

Odox

-OrtiOn
tOrtuOus
pOstpOne
O tatOthrOughOut
neighbOurhOod
OutdO
OutgOing
ParaPet
PersPP h

Phate

PlumP
PostPone
PromPt
ProsPer
Q u

n Quennial

-E a t o E
tEaveRse
-

suEchaEge
aEdouE
EecoEd
EeveEse

buE
bE

g
i

b e

Ey

auEifeEous
wE
e E
coEkscEew
i

aEmouE
bEokeE
pEopeE
uE h e E
AE thuE
paEtneE
moEtuaEy
t

diS

ur Se

diScloSe

135

SeleSs
SeneSs
diS g u

beS

Se

d e S

buSineSs
suSpenSe
diSperSe
reSponS-SticS
ChriStmaS
StreSs
abStruSe

TainT

-TeenTh
Temp T
-T

e r a

TheaTre
T h

auThenTic
hiTherTo

ThirTy

aThle Tic
wiThouT

ameThysT

-TiciTy

-T

T
T

i
i
i

n T*
T e
gh T
e

g a

esTimaTe
-T

culTivaTe

pTiviTy
-TmenT
uTmosT
T
T

o a s
a c

T
T

-TraiT

conTrasT
T r e a T
sTreeT

CRYPTOGRAPHY

136

sTricT
p a T

sUmptUous
c
U
Undoubted

p U n

sTrucT
TrusT

congraTulaTe
-orTuna'Te

TwenTy
trusTworThy
qU a dr UchaUffeUr
rUinoUs
frU

Ul

sUlphUr
scUlptUre
ENGLISH.

conUndrUm
UniqUe
UrioUs
UsefUl
U

beaU

Ul
Ul

caUtibUs
VatiVe
t

Ve

Westward
WindWard
VII.

Three Like Letters with Intervals of One.

pA

Able

MAlAyA
CAnAdA

cArAvAn
cAtArAct
extrAvAgAnt
E

E dEnt

piEcEmEal
prEdEcEssor
rEfErEnce
EgEnErate

vEhEmEnt
ElEmEnt

ElEvEn
cEmEtEry
whEnEvEr
whErEvEr

Iblllty
hlbltlon
rigidity
diminish
n

I a

I s I

crltl c Ism
clvl llan

division

IGcOmOtive

chrOnOlOgy

mOnOpOly
mOnOt On
chlOrOfOrm
sTiTuTUnUsUal

LISTS

AND TABLES

137

Bigrams Eepeated.

dE
I

BEE
I

ININg
bAG g AGe
bAE b AEous
B A r BArous
CA
CA
CA

CAreous
C Ade

u C

TH

As-

THer

mUE m UE
U

UT

T p

ABle
hAB
CHur CH
piCKpoCKet
DE c DE
DE u DE
i

IN
IN

INe
INge

INstlNct
DA r DAnelles
perlTonlTis
pEA c EAble
OE a OE
EN d ENt
PHosPHate
EN
EN
POstPOne
revEEbEEate
EE g E E
QU n QUennial
remiSSneSS
pEE v EEse
diSTr u ST
IN g INg
forTHwiTH
IN k INg
sENtimENt
MA d MAn
MAtheMAtics
MU r MUr
seNTimeNT
NG NG
coNTineNT
NT e N T
cOUrteOUs
ON
ON
PAble
PA
plENipot ENtiary
PO r POise
intEEpretEE
coNTraveNTion
EE d EEss
EE p EEsent
cOUrageOUs
t

SEnSE

aSS

SSin

S S

S S

inTErpreTEr
uNDerstaND
etc., etc.

CRYPTOGRAPHY

138

of Ten Different Letters, which may be used in


stitution of the Figures 0-9 or 1-0, and thereby

Words

Sub-

form

Numeral Key-Words.

AUTHORISED

BACKGEOUND
BANKRUPTCY
BUCKINGHAM
CHIVALROUS
COMPATIBLE
COMPLAINTS
DESOLATING
DESTROYING
EXHAUSTION
FLOURISHED
FORMIDABLE
GELATINOUS
HYDRAULICS

HYPNOTISED
HYSTERICAL
ILFRACOMBE
IMPERSONAL
IMPORTANCE
JOURNALIST

PATRONYMIC
PLAYWRIGHT
PRESUMABLY

LACHRYMOSE
MACKINTOSH
MENDACIOUS

PROMULGATE

METAPHYSIC

PURCHASING
REGULATION
REPUBLICAN

MINERALOGY
MISFORTUNE
MODERATING

SUBJECTION
SYMPATHISE
UNSOCIABLE

PATRONISED

WORKINGDAY

Surnames such as Tichbourne,


as Fair Custom,

PREVIOUSLY
PROCLAIMED
PROFLIGATE

may

or short sentences such

also be used.

ENGLISH.

VIII.

Proportion of Words in Webster's Dictionary


classified according to their Initials.

A
B
C

D
E
F
G

H
I

K
L

LISTS

AND TABLES

139

O'Ol per cent, in the total is clue to the


The above proof the calculations.
nature
approximate
from
another.
to
one
portions vary
dictionary

The extra

FRENCH.

I.

Order of Letter Frequency.

According to Valeric:

ENAIBSTUOLDCMPVP,

etc.

According to Langie

(in

the works of Bossuet, Voltaire,

Maupassant, and France)


According to Kasiski

AT

N,

etc.

:ESEIANTOUL, etc.

Order of Frequency of Final Letters (Valerio).

ESTEANLIUD,
The Commonest Bigrams

etc.

(Valerio).

ES, EN, LE, DE, ON, OU, NT, EE, NE, ED, TE,
SE, EE, AE, ME, AN, IT, ET, IE, Tl, EL, NS,
Frequency of Double

Letters.

According to Valerio: SS, LL, EE, NN, TT, FF, CO,

MM, PP.
According to Kasiski: SS, EE,
Double Letters

at the

EE.

EM,
UE.

NN, TT, FF,

End

CC,

of Words.

EE.

EE,

CKYPTOGBAPHY

140

FRENCH.

II.

The Commonest Trigrams.

ENT, EDE, LES, LLE, QUE, AIT,


EME, ION, EUK, ELL, SSE, EST, DAN, DEL,
MEN, DES, TIO, ESE, ANS.
According to KasiskifENT, QUE, ION, QUI, TIO, ONT,
AIT, ANT, OUK, ANS, LES, AIS, OUS.

According to Valerio:

The Commonest Two- Letter Words.

AN, AU, CE, CI, DE, DU, EN, ET, IL, JE, LA, LE, MA,
ME, NE, NI, NIL ON, OU, SA, SE, SI, TA, TE, TU, UN.
Words

of

One

A,

0,Y.

Letter.

Four-Letter Groups repeated in Succession.

NOUS NOUS, VOUS VOUS.


Proportion of E (Valerio): 17 per cent.
Proportion of vowels (Valerio) 44'5 per cent.
:

FRENCH.

DI.

Order of Letter Frequency in Relation


of Letter in
Initial letters (Valerio)

Second

letters {Langie)

Third letters

to

Position

Word.

-.DLEPACSMEIF,
:

(Langie):

etc.

A U N E I T, etc.
S E U N T I K, etc. (order

in-

different).

Antepenultimate
I

letters

(no order), etc.

(Langie):

E,

followed

:EUNILOES,

Penultimate letters (Langie)


Final letters (Valerio) E S T
:

by A,

EANLIUD

etc.

C X,

etc

LISTS

AND TABLES

141

Consonant Bigrams (Valerio).

Initial

BL, BE, PL, PE, FL, FE, VE, CL, CE, GL, GE, TE,
DE, CH, PH, TH, SO, SP, ST.
Final Consonant Bigrams (Valerio).

NT, NS, ET, NO, CT, EC, SC, ND, ED, NG, EG,
MP, NQ, ST, GT (doigt, vingt), SS (express).

FEENCH.

IV.

Final Bigrams (Langie).

French text of

,000 letters contains, on

an average

25 words ending ES.


23 words ending NT.
10 words ending

EE.

words each ending NE, ON.


7 words each ending UN, LA.
6 words each ending ME, SE, UE, UI, LE, NS, EE.
5 words each ending TE, UE, DE.
4 words each ending IE, EC, ET, EN, OU, UX.
9

Final Trigrams (Langie).

French text
9

of 1,000 letters contains,

words each ending LES, ENT.

7 words ending
6 words ending

ONT.
EES.

5 words each ending

4 words ending

GES, INE.

TEE.

on an average

CBYPTOGKAPHY

142

FBENCH.

V.

Five- Letter Group repeated in Succession.


o

FAIBE FAIBE.
Eepeated Groups separated by a Single

VIS A VIS,

Letter.

PEU A PEU, PETIT A PETIT, DOS A DOS.

Consecutive

Words ending with Like

Letters.

LES BELLES ACTIONS.


Three- Letter

Words ending with E.

UNE, QUE,

etc.

Final S preceded by Three Like

Letters.

CBEEES.
Q

is

is

always followed by U in the body of a word.


preceded by U, except in the words six, dix,
prolixe, mixture, etc., axe, sexe, boxe,

is

preceded by
C, as in chemin, cheval, cher,

fixe,

etc.

etc.

P, as in photographic, etc.
T, as in theatre, etc.

Word

of Twelve Different Letters, which may be used in


Substitution of the Figures 1-12, and thereby form

a Numerical Key-Word.

IMPBEVOYANTS.
1

In English "considerably" might be used. -TRANSLATOR.

LISTS

AND TABLES

FKENCH.

143

VI.

Like Letters at Equal Intervals (separated by

Two

Letters).

AffAire
cAlcAire

cAmpAgne
A n 9 A
scAndAle

ebr A n

p p

A
A

A rrA-

pArtAge
AA
AtlAs
A ttABarBare
BomBe
s

CalCaire

CirC-

C o n CC r o C-

speC

Cle

-DarD-

DinDe
Er
E

m Er
i

EmpEreur

bArbAre
s

evidEmroEnt
f E m m E

EncE
-E n d E

gE

-EnnE
dEnrEe
gEnrE
def E n E
Ens Signer
EnsEmble
pEnsEr
s

E
E
u E
b E

-E n
nE p

-E

hE
-E
-E

EE-

casErnE.

gouvErnEment
-E
-E

siE

rE
E
rE
rE
sE
ail E

u Eillir

rE

Ect

-E

E-

-E

g
g
g

Et
Echir
E

-E

E-

fl-

pE

sEcrEt

m Ent
r

Esse

chancEliEr
E

r
r

v E-

dEscEndre

-E t
-E u

-EurE
E

CBYPTOGBAPHY

144

-E u v Er

E
E
F
G

v u
x p
o r
o n
a c

dl

Icile

a I n

ml

Ion
Ion

p I

mill

a I n e
a Ire

u L

c a

IN f
u N
a N n

N
N
N
N
p

fO

Nterie

o
o

N
N cer
N
N
N
N

Ours
Onne

Pie

PP

P
P
P
e

B
E

t r

empeE
"

o
a o
r

E
E

-E

moE

u
u

foE

-E

-S
S
S
S
a S
d e S
S

s
s

-T
T

U
U

Eesse
E

i
Sn S
e n S
o u S

bealJ

Besentant

E
E
Ee
E
Ee

a
a

-T

-T

Edinaire

t e r

n T
n T

o r
o u
t

u S

c o

T
T
Tud

Up
Ure

0-

poUr

Uo
Uite

Oge

Ue

hO

u
u

cO

OppOse
-0

cOmmOtion
n

0-

Ion

mmI

crl p

On

e.

si g n Ifier
al g u Hie
recuel 1 1 Ir

ml

v I c
f

-0

Edition
Fait

He
I m

explO

p o
r

etc., etc.

LISTS

AND TABLES

145

Like Letters at Equal Intervals (separated by


Three Letters).

nAissA
cApitAine

ChanCelier

prEcisEr
sEcouEr
EcutEr
E

iNst

iN

Er

EtalEr

E
mE t t r E
ExtrE
i

o u

deSsaiSir

me

Tr
eTr

ai T
o

etc., etc.

T
Instinct
,

Net

pEendE e
laEgeuE
pEopoEtion
maEqueE

mEsurEr
EtagE
t

p a

prEsquE
pE

EntrEr
-E

s t

hOrizOn
propOrtiOn
PourPre
aEceuE

ElogE

-EmblEr
-EndrE
-E nt

LegaL
LocaL
siMpleMent
eNchaNter
ciNquaNte

Bigrams Repeated.

cANcAN
b

AE

BA
CA
CA
CA

EN
EN

b
r
1

n
s

AEe
B Ar e
CAire

CAn
CAde
ENt

-ENsENt
-BE
-EE

EN
EE
EE

g
fEEmEE

gouvEEnEE
vEEsEE
EU

MU
NS

NT
NT
-ON

a
e
t

tOU
PE

SE

E U

MUre
NS
NT
NT
ON
OUrs
PEe
SE

poSSeSSion
10

146

TE
fUK

n
e

CEYPTOGEAPHY
TEr
o I N

a I

QU

UE

mUE m UEe

-EEndEE
T E a TEe
qUE q UE

CH er CHer
chEE chE E
cHE r c HEr
FEENCH.

Q Ue

etc., etc.

VII.

Proportion of Words in Littre's Dictionary classified


according to their Initials.
Per Cent.

A
B

..

..

E
F
G

..

..

..

The shortage of 0*48 per cent, in the total is due to


blanks and the approximate nature of the calculations.
The above proportions vary from one dictionary

to

another.

ITALIAN.

I.

Order of Letter Frequency.

-.EIAOELNTSCDP,

According to Valeric
According to Vesin de Eomanini

LNE

S, etc.

etc.

0, followed by

AND TABLES

LISTS

147

Order of Frequency of Final Letters (Langie).

IAEONLKDU.
The same
five

letter frequently

ends two, three, four, or

consecutive words.

The Commonest Bigrams

(Valerio).

EK, ES, ON, KE, EL, EN, DE, DI, TI, SI, LA, AL, AN,
KA, NT, TA, CO, IN, LE, TO, 10, AK, NE, OE.
Frequency of Double Letters (Valerio).

LL, SS, TT, EE, PP, NN, BB, GG, CC.


All the consonants

may

Words

be doubled except H,
of

One

A, E,

I,

J,

and Q.

Letter.

0.

ITALIAN.

II.

The Commonest Trigrams

(Valerio).

CHE, EKE, ZIO, DEL, ECO, QUE, AEI, ATO, EDI,


IDE, ESI, IDI, EKO, PAK, NTE, STA.
The
The

J and

letter

is

HAI, HA,
The

are always followed by a vowel.


used only in the groups CH and
and in four forms of the verb avere (to have):
letters

letter

is

Proportion of

HANNO.
always followed by U.

(Valerio): 12-6 per cent.

Proportion of vowels (Valerio): 46 per cent.

SPANISH.

I.

Order of Letter Frequency (Valerio).

EAOSIKNLDTCUP, etc.

GH,
HO,

CEYPTOGEAPHY

148

Order of Frequency of Final Letters (Valerio).

AESONLEYIDZU,
The Commonest Bigrams

etc.

(Valerio).

ES, EN, EL, DE, LA, OS, AE, UE, EA, EE, EE, AS, ON,
QU, ST, AD, AL, OE, SE, TA, CO, CI, 10, NO.
Frequency of Double
CC, LL,

EE,

infrequently

Letters.

AA, EE, 00, NN.

1
According to Valerio: EE, LL, EE, SS, DD.

Words

of

One

Letter.

A, E, 0, U, Y.
Single-letter
or
A.

words that

may

occur in succession are

SPANISH.

II.

The Commonest Trigrams

(Valerio).

QUE, EST, AEA, ADO, AQU, DEL, CIO, NTE, OSA,


EDE, PEE, 1ST, NEI, EES, SDE.
Doubled Letter beginning a Word*. +

LL.

The

letters Z, J,

H, and

are always followed by a vowel.

always followed by U.
Proportion of E (Valerio): 14 per cent.
Proportion of vowels (Valerio): 48 per cent.

is

1 Neither S nor D can be doubled in the same word, but


they
occur consecutively as the final of one word and the initial of th e

next.

TRANSLATOR.

LISTS

AND TABLES

GERMAN.

149

I.

Order of Letter Frequency.

ENIESTUDAH, etc.
:ENEITSDUAH, etc.

According to Kasiski
According to Valerio
According to Vesin de Eomanini: E, then
Y J C.
etc., the rarest being Q
:

NIE

U.

Order of Frequency of Final Letters (Valerio).

NEETSDHUZF,
The Commonest Bigrams

etc.

(Valerio).

EN, EE, CH, ND, DE, IE, TE, EE, El, UN, GE, DI,
ES, BE, IN, IT, HE, etc.
The Commonest Final Bigrams.

EN, EE, then

the letters S, T, and E.

Frequency of Double

EE, TT, LL,


Double Letters

NN,

at the

SS,

Letters.

DD.

End

of Words.

SS, less frequently LL,

GEEMAN.

EE.

II.

The Commonest Trigrams (Kasiski).

EIN, ICH, DEN, DEE, TEN, CHT, SCH, CHE, DIE,

UNG, GEN, UND, NEN, DES, BEN, ECH.

CKYPTOGKAPHY

150

The Commonest Two-Letter Words.

AB, AM,

AN

DA, DU, ER, ES, IM, IN, OB, SO, UM,


then JA, NU, etc.

WO, ZU,
The bigram

is

is

UN

frequently
always followed by U.
always followed by

commences a word;
or K, except in

"

"
foreign

words.

Proportion of

(Valerio): 18 per cent.

Proportion of vowels (Valerio): 35 per cent.; (Kaeding):


42-12 per cent. 1

GERMAN.

III.

Proportion of Words in Sachs's Dictionary classified


according to their Initials.
Per Cent.

A
B

Per Cent.

Total.

Total.

11-50

D
E
F
G

H
I

K
L

M
1

The German

authority, F. P.

W. Kaeding,

based his calcu-

he established the
presence of 9,260,044 E's, 6,363,537 N's, etc. It may be noted
that one volume of the large dictionary of Larousse contains about

lations

on a total of 60,558,018

20,000,000 characters.

letters (!);

AND TABLES

LISTS

151

The shortage of 0-61 per cent, in the total is due to


blanks and the approximate nature of the calculations.
These proportions vary from one dictionary to
In this case, the middle of the dictionary occurs

NOTE.
another.
at

in Feller's pocket dictionary it occurs at


numbered dictionary at L, etc.

K;

M;

in

Niethe's

RUSSIAN.

I.

Order of Letter Frequency. L

According to

The
Texts

in

AN

Langie:

letter I

Russian characters

DTMVEUKP,

etc.

English transliterations:

R Z K P,

OYAIELNHDTSMUV

etc.

French transliterations:

K P,

E T V E L K M, etc:
French transcription.
A I L E N, hard sign,

I S

predominates in

OIAELNTDCHMUVE

etc.

Final

Letters.

According to Langie: Hard sign, then 0, E, (I)A,


(ignoring final hard sign), A, I, etc.

Russian texts: Hard

sign,

U,

I, soft sign,

I,

(Y)A, E, Y, A,

(ignoring final hard sign), etc.


transliterations
: A,
U, E, 0, I, soft sign, Y,
English

M,

(K)H

M, V,

etc.

The Commonest Bigrams (Langie).


ST, NO, EN, GO, TO, KA, KO, NA, EE, RA, LI, SK,
OS, M', RO, PO, ZA.
1

Including results of supplementary investigations by trans-

lator.

CEYPTOGEAPHY

152

The Commonest Trigrams (Langie).

AGO, STV, ENI, OST, YKH (bigram in Eussian


characters), TOE, STA, IKH (bigram), ENN, NOV,
OEO, STO, EGO, LIS, NI(I)A, SKA, AL', OM', NNO,
EEE, ISK, NY(K)H, etc.
NOTE.

The apostrophe represents the

final

hard sign.

Double Letters (Langie).

NN,

(I)A(I)A,

Words

of
I,

EE, (I)U(I)U,
One

SS,

00, ZZ.

Letter (Langie}.

(I)A, 0, U, A.

NOTE BY TRANSLATOR. To these may


hard sign is ignored, V, K', S'.

be added,

if

final

EUSSIAN.

II.

The Commonest Tetragrams (Langie).

NY(K)H', PEAV, TSTV, VENN, UET:, VSTV.


The colon in the above represents the soft sign.

NOTE.

The Commonest Pentagrams (Langie).

SKAGO, STVIE, L:STV.


The Commonest Hexagrams (Langie).

STVENN, NNOSTI.
Like Letters separated by One Letter (Langie).
ILI,

KAK, OBO, OVO, OGO, ODO, OKO, OLO, ONO,


OSO, POP, TOT, TUT,

Proportion of

etc.

(Langie) 10-7 per cent.


of
vowels
Proportion
(Langie): 43-5 per cent.
:

LISTS

AND TABLES

153

POETA'S TABLE.
This table was composed by Giovanni Battista da
Porta, a Neapolitan physician, author of a work on

cryptography entitled De furtivis litterarum


de ziferis, Naples, 1563.

notis, vulgo

CKYPTOGKAPHY

154

below

accordingly s is taken as the cipher equivalent of i.


Again, the plain letter n, ciphered by means of the same
G, will be represented by d, which occurs immediately
i ;

above.

The

rule, therefore, is to

take the letter which

occurs either below or above that of the plain text in the


double line corresponding to the key-letter. For instance,
to cipher the

CAE, we

'

word

red

'

by means

of the

key-word

look for r in the double line to the right


of C,,and find immediately above it the letter/.
Proceeding in like manner with the second letter e (key-letter A),
first

and the third

letter

d (key E), we obtain the result:

red

CAE

=f

For deciphering, we adopt exactly the same method,


"
the cipher word
vtu," with the key-word NOT, for
example, resulting in:

vtu

NOT

=b

a d

VIGENEEE'S TABLE.
This

translator
entitled

was established by Blaise de Vigenere,


and French diplomat, author of a work

table

Traite des

cliiffres

ou

secretes

manieres

d'ecrire,

Paris, 1586.

The upper horizontal

line of capitals represents the


the column of capitals to the left is

plain-text alphabet
used to form the key-word.
;

Supposing the

first letter of

the key-word

is

E, and the

of the plain text i, we descend from I in the


line
of
capitals until we reach the line of small letters
top

first letter

LISTS
beginning from

K in the

of intersection

we

find

AND TABLES

column to the left. At the point


which becomes the first letter

z,

in the ciphered text (see p. 28).


<;
"

To decipher the word

BEX, we

first

kik

by the aid

of the

key-word

look for k in the horizontal line beginning


\
Letters of plain text

155

CKYPTOGBAPHY

156

NUMBEE OF POSSIBLE COMBINATIONS.

....
Number

With
3 letters 1

of Combinations.

-24

4
5
6
7

8
9

10
11

12
13
14
15
16
17
18
19
20

120
720

5,040
40,320
362,880
3,628,800
39,916,800
479,001,600
6,227,020,800
87,178,291,200
1,307.674,368,000
20,922,789,888,000
355,687,428,096,000
6,402,373,705,728,000
121,645,100,408,832,000
2,432,902,008,176,640,000

BKITISH SURNAMES.
Frequency of Terminations (compiled by Translator).

In a

list

of over a

thousand

different surnames, numerical

position was occupied by the following terminations,

in

order of frequency:

SON, TON, EK, ING(S), LEY, FOBD, STON(E), MAN,


OCK, BY, HAM, LAND, ICK, ETT, WELL, FIELD,
KIN(S), LOW(S), WOOD, MOBE, BUBN, HUBST,

WOBTH, DALE, SHAW, BOBOUGH, STOWE,


BIGHT, WAY, STEAD.
1

I.e.,

ABC, ACB, BAG, BCA, CAB, CBA.

LISTS

AND TABLES

157

FEENCH SURNAMES.
Frequency of Terminations (Langie).

Out

of 1,000

French surnames (approximately):

50 end in IEK.

38 end in
21 end in

19 end in

AKD.
EAU.
AUD.

13 each end in

LET, LLE.
AND, NET.

12 each end in

AUX, ERE, EET.

15 each end in

end in LOT, EON, SON.


each end in OUX, TTE, ULT.
each end in CHE, GEE.
end in LIN.
each end in EIN, UET.

11 each

10
9
8

CEYPTOGEAPHIC MATEEIAL.
One

or

two manuals

of

cryptography

(see Bibliography).

Dictionaries in several languages.

English, French, German, etc., rhyming dictionaries.


Two Saint Cyr slides (see p. 110).

Two graduated

rules,

one numbered from

to 50, the

other from 51 to 100.

Paper ruled in squares.


Slates ruled in squares.

Tracing paper.
Coloured pencils.

A
A

T-square (useful for consulting Vigenere's Table).


ready reckoner for rapidly calculating proportions.

CKYPTOGKAPHY

158

few hundred counters on which the letters of the


alphabet are inscribed. For instance, in 100 counters,
one would have 18 E's, 9 S's, 8 B's, 7 A's, 7 I's, 7 N's,

The use of counters from time to time rests the


and enables one to try a number of combinations
more rapidly than could be done with pen or pencil.

etc.

eyes,

BIBLIOGKAPHY
WORKS RECOMMENDED.
La cryptographic devoilee, by C.

F. Vesin de Romanini.

Die Geheimschriften und die Dechiffrir-Kunst, by F.

Paris, 1857.

W.

Kasiski.

Berlin, 1863.

der Kryptographie, by Ed. B. Fleissner von Wostrowitz.


Vienna, 1881.

Handbuch

La

cryptographic militaire ou des chiffres usites en temps de guerre,


by Aug. Kerckhoffs. Paris, 1883.

La

cryptographic
Paris, 1885.

Essai sur

La

les

et

ses

applications a Vart militaire,

by H.

Josse.

methodes de dechiffrement, by P. Valeric. Paris, 1893.


by A. de Grandpre. Paris, 1905.

cryptographic pratique,

PART

IV

THE PLAYFAIR CIPHER SYSTEM,

ETC.

BY TRANSLATOR
IT

is

surprising that there.

is

no work in cryptography

in English, although M. Langie points out that there is


an extensive bibliography in other languages. I have

.made a careful search, both in England and the United

book or manual on this fascinating subject,


M. Langie defines cryptography
as the art of communicating thoughts secretly, and this
certainly appears to me to be a better definition than the
secret writing," as it is perfectly
stereotyped one of
obvious that writing is not the only medium by which
It used to be
secret communication can be effected.
a great problem to travellers and residents in various
parts of Africa how news could be transmitted with
almost incredible rapidity over large distances, and many
were inclined to attribute this to some supernatural
States, for a

but without success.

'

agency. Further investigation, however, proved that


the news was transmitted by beating a drum in a certain

manner, as provided for in a prearranged code, the


message being relayed from one post to another.
Cryptography, in some form or other, has a surprisingly

upon the everyday events of ordinary life.


upon your handkerchief a mysterious little
to you is meaningless, but which in your
which
symbol
indicates
your name and address, and many a
laundry

great bearing

You

will find

fugitive criminal has been brought to justice


159

by such a

CKYPTOGKAPHY

160

slender clue as a laundry mark upon some garment which


he has had to leave behind at the scene of his crime.

Kacegoers may have noticed individuals standing on


the top of a cab or on some coign of vantage, semaphoring
This
energetically, in the intervals between the races.
"

tictacking," and I understand that the


have various codes, which they do their
best to keep secret, for transmitting prices from Tatter-

process

men

sails'

is

called

operating

it

ring to the outside bookmakers.

The marking of cards by sharpers is a form of cryptography in which an amount of ingenuity is exhibited
worthy of a better cause. Playing cards, ostensibly for
conjuring purposes, are sold publicly in the United States,
each card being marked in such a manner that any one

with a

practice can as readily read the card from the


face.
One of the commonest forms of indicat-

little

back as

its

ing the face of a card on its back is in the form of a clock,


as shown in the following diagram:

FIG.

1.

FIG. 2.

At the four corners


Fig. 1 shows the back of the card.
of Fig. 1 will be observed four small rings, an enlargement of which is shown in Fig. 2. This is intended to
represent a clock and the twelve outer rings represent
the hours. A small white dot at one o'clock represents

THE PLAYFAIK CIPHEK SYSTEM,


an

ace,

and

on

so

until eleven o'clock,

ETC.

161

which indicates

a Jack, while twelve o'clock denotes a Queen, and a dot


in the centre spot represents a King.
The suit is indicated by a small white dot in one of the four small circles

around the centre spot. The top dot represents diamonds,


the one on the right clubs, the one on the bottom hearts,
and the one at the left spades. Even after you are informed that the cards are marked, it is surprisingly
difficult for

the uninitiated to detect these marks and, in

cannot refrain from repeating the advice so

passing,
often given that great care

with strangers
Business men find

is

essential

when playing

cards

of

some form

various goods they

by means

of

necessary to

make

extensive use

is most commonly done


remembered word of ten

This

easily

which are used to denote

1,

'

'

instance, the word

to represent

sell.

some

different letters

it

of secret writing to indicate the price of the

bankruptcy
2, and so on.

figures.

An

extra letter

generally used in the case of a figure which

For instance, 11
or BZ/-.

shillings

For,

might be employed,
is

is

repeated.

would be expressed as BX/-

Certain firms use a variety of signs such as


but obviously it would

circles, rectangles, triangles, etc.,

not be a

difficult task for anyone to break this code.


sometimes most important for manufacturers and
merchants that their prices should be strictly secret,

It is

and

have frequently been asked

for advice

and assistance

in this respect, but it is extremely difficult to devise any


system which can be easily written and read, that will

same time defy the efforts of inquisitive rivals to


discover the real figures.
The recent Great War stimulated the general interest in
cryptography, and many and devious were the methods

at the

11

CKYPTOGKAPHY

162

adopted by spies in the various countries involved in the


war to transmit information secretly. It would be impossible in the scope of this work to give more than
passing mention to the many and ingenious devices that

were adopted or to show

how

these efforts were almost

invariably defeated by the ingenuity and resources of


the cryptographers in the various censors' departments.
Secret

communication

is

by no means confined

to

naval and military requirements or to diplomatic offices.


Many important financial houses are well aware that
unprincipled rivals would stick at nothing in order to
be able to tap their messages and to break their cipher.

obvious that the transfer by telegraph of large sums


money must be done with very great care and secrecy,

It is

of

and all the great banks employ elaborate methods to


insure that their secrets will not fall into dishonest hands.
It is well

methods

known

that tramps in

all

countries have their

communicating with each other. This is


usually done by means of chalk marks on the door or
wall of a house which one of the fraternity has visited.
of

The French Police recently captured a copy of the code


used by tramps, full particulars of which were published
in the London Sunday Express of October 9, 1921.
M. Langie apparently considers that the use of invisible
or sympathetic inks is of no value, and is almost certain
I do not altogether agree with this, as
to be detected.

frequently happens that it is of vital importance to the


recipient of a secret message that he should be certain

it

that his are the only eyes to see this message.

When

person has to employ any means of secret communication,


it must necessarily follow that someone is anxious to
obtain possession of the secret information. In case of
an ordinary cipher, the letter may be opened and photo-

THE PLAYFAIK CIPHEE SYSTEM,

ETC.

163

graphed and the cryptogram solved without the rightful


recipient being aware of the fact, and he fondly imagines
that he alone is the custodian of the secret. If a suitable
form of invisible ink is used, the recipient has at least the
satisfaction of being absolutely certain that his are the
only eyes to read the concealed message. There are
varieties

many

of

sympathetic inks, the most


milk, orange or lemon juice, dilute
which are all revealed by the applicathese

widely known being

sulphuric acid, etc.,


tion of heat.
A very simple, although not very wellknown form of sympathetic ink, is to moisten a clean pen

with either saliva or water and write the message either


upon an envelope between the lines of the address, or

between the
newspaper,

lines of a letter or inside the

wrapper of a

may be arranged. The recipient then


ink on the arranged section and promptly

as

pours a little
rubs it off with water.

The scratching of the pen with


moisture has removed the glaze on the paper in such a

even with a powerful magnifying


deposited on the surface it attacks
those portions where the glaze has been removed, thereby
making the words written stand out quite distinctly,

way

glass

that
;

but

it is

invisible

when ink

is

while the surrounding glazed surface


soiled by the application of the ink.

is

merely slightly

There are certain inks which can be made to appear

and disappear at will. A solution of chloride of copper


and water may be used as ordinary ink, and when the
water evaporates the writing will disappear and can be

A solution of nitric
revealed by the application of heat.
acid may also be used, and this can only be revealed by
wetting the paper. After it dries, however, it again
becomes
renders

above-mentioned objection
2 per cent, solution of acid of

invisible, so that the


it

unsuitable.

CEYPTOGEAPHY

164

when used as an ink is quite invisible, and can only


made readable by immersion in hydrogen sulphide

lead

be

This would appear to be a comparatively safe ink


to use, but in the course of some experiments I made in
gas.

New York at the Ledoux Laboratory, Mr. Albert M.


Smoot, their technical director, made the discovery that
after the writing had been made visible by means of
exposure to hydrogen sulphide gas,

it

could be

made

to

with peroxide

disappear again by slightly moistening


Writing done with a solution of potassium
can
ferrocyanide
only be made visible by the application
it

of hydrogen.

some

Probably the safest form of secret


ink
a fairly strong solution of potassium nitrate or
common nitre. Writing done with the resultant ink is
absolutely invisible, and can only be revealed by the
application of a flame which will run along the characters
traced on the paper. Many readers will doubtless have
of

ferric salt.

is

seen this form of sympathetic ink in Christmas crackers.


When making use of this form of secret ink, the writing

should begin at the extreme end of the paper at a pre-

arranged spot.

would recommend that anyone who

desirous of using this form of secret ink should first


make some experiments to see that they get the exact

is

strength of solution required and the right quality of


There is a counterpart to sympathetic inks in
paper.

the form of disappearing inks which, however, are of very


The best known of these is a
practical value.

little

solution of starch with a few drops of tincture of iodine.


The resultant ink is blue and to the uninitiated appears
like

ordinary ink.

Within a short time, however, the

iodine evaporates, and the starch becomes quite dry so


that it leaves the paper without any trace of writing

whatever.

THE PLAYFAIR CIPHER SYSTEM,

ETC.

165

M. Langie states that the cipher invented by Francis


is extremely easy to break, but I am of the opinion
that this system used with certain variations could be

Bacon

made extremely

difficult.

well-known Baconian en-

Fabyan of Chicago, believes that Bacon


a
great deal of secret information into the
interpolated
manuscripts of his various works by means of his cipher.

thusiast, Colonel

The method supposed

to be

employed

is

the use of

dif-

ferent kinds of type, but although a tremendous amount


of research work has been done by Colonel Fabyan and his
assistants

have been unable to obtain any concrete

evidence which would prove that Bacon did employ his


cipher in this manner, and the various claims that have

been made up to the present appear to me to be based


merely upon conjecture.
Students of cryptography should make themselves
conversant with the Morse Code or Alphabet. The
is used
by all countries of the world except
America, where a slightly different form is employed for
inland telegraphy. In practice a dash is equal in length

following table

to

three dots,

and a space between two elements or*

The
signals in a letter is equal in length to one dot.
space between letters in a word is equal in length to three
dots, while the space between words in a sentence is
equal in length to five dots.
It will be seen that there is a liability to error in trans-

mission of messages

adhered

to.

if

the foregoing rules are not strictly


will convert A into E T or N

Bad spacing

T E, and a slight examination of the following table


show many other telegraphic identicals, which all
have to be borne in mind when endeavouring to decipher
into
will

a cryptogram which has been telegraphed.

CEYPTOGKAPHY

166

INTERNATIONAL MORSE CODE SIGNALS.


Letters.

A
B

Figures.

2
3

P
Q --

E
F

...

Q.

K
L

U
V

M --

.-

6
7
8
9

.-

-.

.
-

M. Langie has omitted to give any reference to the


Play fair
cipher, which has been extensively used for
This cipher is one of the substitution
military purposes.
variety, and may be operated with one or more key- words,
"

which
ment.

'

may be located
This square

in the cipher square by pre-arrangedivided into twenty-five separate

is

I.
^compartments, and the letter J is "always represented by
"to
BANKRUPTCY
to
be
the
key-word
Suppose
be distributed between the first and fourth lines of the

square.

Fig.

will

show

their position:

THE PLAYFAIK CIPHEK SYSTEM,


The other

letters of the alphabet

ETC.

167

which are not included

in the above ten letters of the key-word are then added


in alphabetical order, beginning at the first vacant square,
as

shown

in Fig. 2

CRYPTOGKAPHY

1G8
the

first

vertical

column

of Fig. 2, the substitution

would

be S B.
5. When the
pair of letters to be enciphered occurs in
the same horizontal column, substitute the letter at
the right of the plain text letter. When this letter is at

the end of a column, substitute the letter at the extreme


of that column
e.g., to encipher T Y, which are in
the fourth horizontal column of Fig. 2, the substituted
left,

letters

would be C U.

6. When the letters to be enciphered are at opposite


corners of a rectangle, substitute each letter of the pair
by the letter in the other corner of the rectangle on the

same horizontal
be enciphered

line

with

K P, D

it
e.g., on Fig. 2, A C would
would be represented by G I,

RL

by A Q.
The enciphered message may be written in groups
three, five, or eight letters, and the letter agreed upon

and

7.

of

for the

to

fill

purpose of dividing repeated letters may be used


up a group. Should more than one letter be re-

quired to complete a group, as many letters as are required


be taken from a prearranged word, such as STOP,

may

FINISH, etc.
To decipher a message
divide

the

letters

sent in this code,

you simply
the
reverse
and
writing
pairs
The
rules.
decipherer
preceding

into

according to the
or Z, as the
should never neglect to write down the
This simple
case may be, when used as a divisory letter.
time
in
a
lot
of
save
will
often
decoding a
precaution

lengthy message.

The following example

will

show the method

of en-

with
ciphering in accordance with the foregoing rules,
"
as the key-word, distributed in the first
bankruptcy

"

and fourth columns

of the square.

THE PLAYFAIE CIPHEK SYSTEM,

ETC.

169

Suppose the message required to be enciphered to be:

You may

expect relief in three days," and that


to be used to divide duplicated letters.
Divide the

is

plain text into groups of

YO UM AY
CQ TI KP
KE XE DA
AH VG EB

two

letters each, as follows:

CT KE LI EF IN TH
GV VL YC AH ML FG MB YF

EX PE
YS
UZ

and then underneath each group of substituted letters,


shown above. This being done, the message should
be divided into groups of five, and sent as follows
as

CQTIE PGVVL YCAHM LFGMB YFAHV GEBUZ


*

In the same manner, the message


Sell all you have
immediately
may be sent, on the understanding that
the cipher is to be divided into groups of eight letters,
'

'

for

which Z

is

and
The resulting

to be used to divide repeated letters,

STOP

to complete an unfinished group.


will
be as follows
cipher
:

VDQVPEQV QPICEKAL LOWQFELB PFQPWULC


The Playfair system

is

ing, for several reasons.

after

memory,
and receiver

which

to bear in

all

one of the best forms of cipherIt is

that

very simple to commit to


is

necessary for the sender

mind

are the key-words or sequence


the position they are to occupy in

of key-words, and
the square. The key may consist of one or more words,
provided they contain different letters e.g., FAIK

CUSTOM

or

A JUKY OF MEN.

The key-word may be

changed in alternate words or at certain intervals.

For

instance, a message might be sent using as key-words


in the first and fifth columns of the

BANKRUPTCY

CRYPTOGRAPHY
square, CUMBERLAND in the second
TICHBOURNE in the first and third,

170

and fourth, and


and many other

variations will readily suggest themselves to the student.


The message may also be sent in groups of three, four,
or eight letters, so there are abundant opportunities
throwing obstacles in the way of a decipherer who is

five,

for

not in possession of the keys.


An indication of the difficulties to be overcome by the
decipherer will be seen in the first example, where the
six E's in the plain text are represented in the cipher by
G
F
and G, and in the second example the four

VH

E's are transcribed

That these

DLF

difficulties

and F.

can be overcome

is

proved by

the fact that I sent a message ciphered by the Playfair


method to my friend Lieut. -Commander W. W. Smith,

one of the most

skilful

cryptographers in the U.S. Navy,


following account of the

who has kindly given me the

steps he took to solve the cipher, which will be of great


assistance to the student.
We are both of the opinion
that when important messages have to be sent they should

be enciphered with more than one key-word, as by this


method less time is required to cipher the message than
would be necessary if you endeavour to avoid the use of
the commonest digraphs, or to split

them by means

of

divisory letters.

SOLUTION OP THE PLAYFAIR CIPHER.

By

Lieut- Commander

W. W.

Smith, U.S. Navy.

The Playfair cipher may be recognised by the following


characteristics

(a)

It is a substitution cipher.

(b)

It

always contains an even number of

letters.

THE PLAYFAIR CIPHEK SYSTEM,


(c)

171

divided into groups of two letters each, no

When

same

group contains a repetition of the


or EE.
Unless the message

(d)

ETC.

is

letter, as

NN

very short there will be

recurrence of groups, and this recurrence will, in general,


follow the order of normal frequency of digraphs.

has
(e) In messages of length, unless encipherment
been made from several squares of different keys, whole
words are likely to recur in the form of repeated
groups.
In the solution of the Playfair, we need not consider
the normal frequency of individual English letters,

We are, however, very much conE, T, 0, A, N, etc.


cerned with the normal frequency of pairs or digraphs:
th t er, on, an, re, etc., as will

be shown

later.

Before taking up the actual solution of a test message,


let us examine the
system for its inherent weaknesses From
the square of the key-word BANKRUPTCY shown on
:

it is

page 167,

seen that the cipher letters

YE represent th of

and so long as this same key is in use, th plain


can only be represented by YE in cipher. Likewise, on is
always MK, and an NK. (NOTE: Throughout this displain text,

cussion

we

plain text
as above,
as

1,

2,

letters

by capitals and
Also, in referring to equations
designate the letters of the equation

will represent cipher letters

by small letters.

3,

we may
and

of the

4.

Thus

1,

2=3,

cipher group, and

4,
3,

where

and 2 are

4 are plain text

letters.)

Case

1.

Letters at opposite corners of a rectangle:


If

YF=th

then FY==ht

TH=yf
HT=fy

CKYPTOGKAPHY

172

Case 2.- -Two letters in the same line or column:


In line

of the square,

and

NK=an
KN na

But AN is not equal to nk, and NA is not equal


and reciprocity is only partial.
We may therefore note Eule I. as follows:
Kule I. Eegardless of the position of the letters
if

the

made

is

that

to kn,

in the

23,

1,
4,
square,
assumption
the following equation will also hold: 2, 1=4, 3; while
if the letters 1 and 2 form opposite corners of a rectangle,

the additional equations

Now,

be assumed:

4 (cipher) =1, 2 (plain),


3 (cipher) =2, 1 (plain).

3,

and

may

4,

as each letter of a line or

bined with but four other letters of


four letters of its

employed

column can be comown line, and with

its

own column, and

as each letter

when

combined
would appear

at the corner of a rectangle can be

with each of 16 letters to form a group, it


1 is twice as probable as Case 2.

that Case

Now,

in the square, note that

NK=an
FK=gn
MK=on
TK=cn

FA=en

FL=em
also
.

FP=et

FV=ew
FG=ef

From

this it is seen that of the twenty-four equations


be formed when each letter of the square is
can
that
employed either as the initial or final letter of the group,

THE PLAYFAIE CIPHEE SYSTEM,


five will indicate a repetition of a

ETC.

173

corresponding letter of

plain text.

Hence, Kule

it

has been determined, in the

FA=en, there is a probaother


group beginning with
any
indicates e-, and that any group ending in A indi-

equation

one in

bility of

After

II.

2=3,

1,

4, that, say,

five that

cates -n.
or, and en have been
above rule may be of
additional digraphs and partial

After such combinations as

assumed or
use

in

determined,

discovering

er,

the

words.

Kule
equal

3,

In the equation

III.

possibly indicate

er,

eliminating possible

or

2=3,

1,

and 2 can never equal

can never

4,

KK

Thus,
This rule

4.

AY=an.

equations when the

could not
of use in

is

under

is

cipher

investigation.

Kule IV.

In the equation

1,

2=3, 4,
same

if

identical, the letters are all in the

and

in the relative order 3, 4, 2,

NK=em, and

shown,

equivalent to

-ANK-,

the order
or

ANK.

In the square

ANK

is

and 4 are
column

line or

This

is

which

is

a very useful

rule.

Rule V.

same

2=3, the letters of the equation are in the


4,
column, and in the relative order 2, 1,
- 4, 2, 1, -. Thus it is
equivalent to 4, 2, 1, --, or
If

line or

which is

seen that in the square, BR=rfc, and the order


KK.
which is the same as
or B

KRB

is

KB

K,

Some cryptographers claim that from an analysis of


the cipher message, the letters which are found to combine in groups with the greatest variety of other letters
will

may

very likely be the letters of the key-word. This


be of some value provided the key were contained

in the

first

two

lines of the square,

and

if

the key-letters

CKYPTOGKAPHY

174

could positively be eliminated it would be possible to


Unforsolve the message and reconstruct the square.
be
cannot
letters
these
positively eliminated,
tunately,

and the square is not always constructed in a regular


manner. The disadvantage of this system is that it
tempts the student toward guessing the key-word. A
false and usually unsuccessful method of attack.
Eule VI.

Analyse the message

for

group recurrences.

Select the groups of greatest recurrence and assume them


Substitute the assumed
to be high-frequency digraphs.

digraphs throughout the message, testing the assumptions


in their relation to other groups of the cipher.
The reconstruction of the square proceeds simulta-

neously with the solution of the message and aids in


Let us now
hastening the translation of the cipher.

take up the solution of the actual test message given


below:

APBNOH
ANNSXE
KBSNHL
ENPFVB
OAEYSC
FBTACS

VSBNOH

EUXUUO
NESCSD
EAEGCI

ENAOEA
OUUADT
DYPYHS
NVOBNX
SOKTDN
HGQEOA
ENEAEB
QENSXU
VNNSGE

BHNSES
INTHXU
GEGBTE
AEGBIZ

YCNIVK

DADYPY

GIOEEB

WQGEUD

BNOAEP

NIYEEB
WECSFQ

NYSIQC

GNXEOU
KNXDTF

OAFLIG

CIENOM
ECOEOT
QNFLEN
CNOELC

EAEEHN
EXXUUY

In working with the cipher, disregard the above grouping and rearrange the message in pairs of letters.

We

will first analyse the

above message by drawing

a chart of group recurrences (Fig.

.)

THE PLAYFAIE CIPHEK SYSTEM,


FIRST LETTERS OF PAIRS.

CO

Q
o
o
H
Jz;

02

ETC.

175

CBYPTOGKAPHY

176

This chart shows that the following groups occur in the


cipher four times each:

OA

EN

GE

EA

NS

XU

No group
usual.

each.

occurs more than four times.


This is unThe groups BN, CI, and OE occur three times
All of the above groups must represent common

It is well known that


pairs of letters in English text.
the order of frequency of common pairs of letters is as
follows (from a count of 2,000 semi-military letters):
'

th
er

on
an

50
40
39

at

en
es

of

he

38
36
33

in

31

ea

ed

30
30
26

ti

re

nd
ha

or
nt

to
it

25
25
25
25
25
24
22
22
22
20

st

io
le
is

ou
ar
as

de
rt

ve

20
18
18
17
17
16
16
16
16
16

The above table and the frequency chart of Fig. 1


must be kept constantly available throughout the attack
on the ciphered message.
Of the most commonly occurring groups of the cipher,
OA, EA, EN, NS, GE, and XU, we note from Fig. 1
that the reciprocals AO, AE, SN, EG occur only once
each, while NE and UX do not appear in the message.
This is unfortunate, for had one of these reciprocals
occurred, say, three times, we might have begun by
assuming the groups to be er and re (see above table).
Now, as has been shown, the group GE cannot mean
er or or, for the second letter of

the fourth.
III.).

Nor can

EA

or

an equation cannot equal

EN

symbolise

Thus we can eliminate a few

of

re or rt

(Eule
the possible

THE PLAYFAIE CIPHEE SYSTEM,

ETC.

177

But any one of the six groups


The most common jour-letter group in

meanings of the groups.

may

represent

ih.

English is known to
INTH, ENED, TION,

THEE,

be

while such groups as

are very often encountered.


Hereafter, in referring to the groups of the cipher,
let us indicate by a small figure the number of times that
etc..

XU

the group occurs, thus


4
Note that in lines 1 and

BN3 OH EN
XE OU and in lines

recurs:

4,

2,

the following combination


and that in lines 2 and 4 we have
7,

3 and 10

we have

DY PY
2

But

be useless to attempt to guess the meaning of the


two last mentioned groups, as the individual groups are
it

will

not frequently used in the cipher, and occur only with


each other. Thus
and
may indicate

XEOU

DYPY

four unusual letters that recur in the cipher, or they may


be caused by the insertion of nulls between repeated

To guess

their meaning would not greatly assist


in extending our investigation.
Likewise, it is best not
to begin our assumption at the beginning or end of the
cipher, as the sender of the message often purposely

letters.

and ends with unusual words.

The repeated
however,
4
groups
2
2
present opportunities.
Also, note the combinations in the cipher of our most
begins

BN OH EN

common

groups:

EN EA

GE 4 OA
NS 4 GE

NS4 XU4

We
be

must first assume each of these groups in turn to


iher, which is the most common four-letter combina-

tion in English text.


other combinations of

NS 4 may

be

th,

but

Failing to establish this relation,

common

XU

4 is

digraphs will be assumed.


probably not er, as it occurs
12

CRYPTOGRAPH?

ITS

the end of the cipher. in Xl'l'V.

at

bo considered as

Hule

Suppose uHj

kO
Now

nni^t

it

l.V

This

(for.

is

an excellent

as reciprocals of both groups occur in tho

Sumption,

and

message,

bo

to

V,

However.

\S, uK, cannot bo

possibility.

u K cannot

if

hK

0\

and

=/Ji

lUi

cr.

and

-hi

-v.

an nnconnnon digraph, and can occur only


"
"
or
!C-it-ht-ht' or in a combination of
as in
eight

in

".

is

:'

'

rursuing this assumption.


tho combination UN K V

thought."
lino

for

10.

HA

Thon

and

:/,

\l\~ti.

throughout the message and


^

\vo

Substitute

\\o V^
c;ot

for

assume

in

-w~it

-ht.

lui.

those

AO HA

values

in lino

1.

H AH in lino \). ;;
Those do not appear
the
and
after
promising,
carrying
investigation farther
it
\\as decided to abandon the original assumption that

Ttit.

and

for

ii

;:

NOVF: In all work of this nature false assumptions


made, but as tho investigation proceeds they will

will be

ONontuallv be proved
*

tker

false.

false starts

In this case a ^
^reat

many

were made due to unusual conditions,

was abandoned

in

and

favour of such combination^ as

before the investigation resulted in SUOC68S,


For the sake of brevity, those steps will be omitted.
:.

etc..

Assume HN 4 HA^
KN=th
Then
HA--er
and

to be th<r ^lino 7

NB=ht
AR=re.
Unfortunately,
but

(Kg.

one
:

AH.

we have no NK in the
Make those substitutions

cipher,

and

throughout

THE PLAYFATR CIPHEB


AP BX OH RX AO \i\ Gf OK
th

AX \S XR OU

ETC.

he

er

L'A

DT BX OA RP XI

YK RB

r-

KB SX HL DY PY II- XY SI Qr \VR
EX PI- YB XV OB XX GX XR OU OA
r-

CS

PL

IG

h-

OA KY

-0

B(

KT DX KX

XI)

TF

CI

RX OM
th

PR TA CS HQ GR OA BH

B8

EC

OP,

OT

-r

YS

BX OH RX RA
th

fx

uo QE xs

PJ; XT;

TH XU QX FL RX

Rli
e-

IX

XT;

GR GB TR ex OR RC
he

QX

BD YX XS

X-

frh

th

-r

AP,

-t

GB

RA RE HX

IZ

re

RA KG
f-r

YC

CI

f-r

e-

XU VK DA DY PY RX XT

L'Y

eh

First

step

not

and 3rd steps

underscored.

p^^'-^

1^.0

ab.-titutions

and 1S1, aro

of

undf-r-corf-d

2nd
.

dividf-fl into

FIG. 2.

There

is

also a possibility (link-

I.)

that

TH=rn, and

but the.v- occur only once f-ach, and the reciER


and HT do not occur at all.
procals
Xow, if RA=r, we know, from RuP- IY.. that the

RE=ar,

three letters are in the same line or column of the square


and in the ordf-r ERA
.

RX

is

assumed to be

th,

the partial square must be,

CEYPTOGEAPHY

180

depending whether this equation

is

formed from a

rect-

angle or a line:
(1)

or,

E E A

(2)

E E A

H
'

In either case, we have established the fact that


is
somewhere in the same column with E, regardless of the
position of

T and N, and we have

for certain:

E E A

H
Now
Eule
that

note other groups containing E. We know from


that if EA=er, there is a chance of one in five

II.

EB=e-, and EE=e-, and GE=-e.

Now GE

is

used four times in the cipher, and as there

a good chance that it may be -e, let us assume it to be


the highest digraph ending in e in the frequency list.
This digraph is he.

is

Then

Make
same

GE=he
EG=eh

these substitutions in the cipher (Fig.


time adding to the square.

E E A
G H Going back to

at the

(4)

we have:

2),

(1)

and

(2), if

we combine

(2)

with

(4)

THE PLAYFAIB CIPHEE SYSTEM,


E E A
G H
-

From

we may obtain some partial


r, EN=r-,
TH=-r, TA=-r, AN

the above

such as

181

T -

(5)

ETC.

equations,
etc.

Sub-

stitute these values in Fig. 2.


As we have as yet no substitutions that can be extended,

common digraphs to be subcommon groups. In line 9 we


-GE - AE or NS he-re. NS is
groups NS
times, and must be a common pair.
Turning

we must attempt
more

stituted for

have the

to find

of the

used four

back to the

list of normal frequency of digraphs, we find


the group of highest frequency next to th

that on

is

and

Assume NS

er.

If this
1

assumption

= 4, these

correct,

then by Eule IV., since

same line, and in the


It is evident from our partial square
cannot form a column. We may therefore

ONS
ONS

build

up the square

is

three letters are in the

that

order

to be on.

as follows:

(6)

E E A
G H

ONS

NOTE: In the partially completed square the horizontal lines are definitely fixed as shown, for KA=er,
and NS=on. We also know that in the column TEHN,

follows

N follows H, for BN=tfi. But we are not


EH=Tr, as there is a fifth letter to be placed

T and

certain that

can only come below E or below N.


Now substitute in Fig. 3, on for NS and no for SN,
and also the new groups that we get from the square
in the

column, and

above, namely:

it

CKYPTOGKAPHY

182

NG=oh

THE PLAYFAIB CIPHEE SYSTEM, ETC.


In the last two lines of Fig. 3 we now have er\ecannot be th, and if taken from square 6

HN
be

rli.

183

HN er eh.

would
where
assume HN=w/t, which would

This would spell nothing, and as the word

suggests itself, we may


give us an addition to the square as follows

it

"

'

(7)

E E A

H
N

Substitute TB=rrf,

TH=nw,

and

W=wh

in

Fig.

3,

WB=rf.
Line 1 now shows

also

a first word evidently ending in


it
the
but
cannot
be
word having as the square
ing,
does not permit ha to be represented by AP.
BN occurs three times, and must be a common group.
'

Ti was

'

'

but was soon found not to be satisfactory.

tried,

After a few similar suppositions ri was decided upon,


substituted throughout.
and BN

BN=ri, we

see from -the square that i cannot follow


and that B and i must be in the same
separate column as below:
If

GH

in line 3,

(8)

E E A
G

It is evident

now

W
H

that lines 2 and 5 of the square form

the key.

We

are not

able to

determine whether the column

B --- 1 is as shown or is adjacent to the column A --- S,


but will place

it

as

shown

in (8) to avoid confusion.

CBYPTOGBAPHY

184

Substitute in Fig. 3 the equations taken from the

square:

new

HB=r-; IG=o-; IN=-o; GI=-o; NI=o-;


"

'

The word revealed is not these


"
as at first supposed, but is
The ser-on." Few letters
can fill the blank space, and we decide the word is
sermon." Thus GI=om and IG=mo.
Note the

first line.

'

We

have but one of our six commonly-used groups


In line 7 we have
undeciphered namely, XU4
-o nw XU. Keturning to the partial square, X and Z,

left

being uncommon letters, are probably not in the key.


not in the key,
probably follows W, and if U likewise
follows T in the first line,
as," a common digraph,

If

XU="

and

zuXU="

was."

Also, as K and S are already in the key-lines, we may


assume Q to precede T in line 1 of the square, and as i
already appears, GH in line 4 is probably followed by K.

Building up the square as above, we have:

(9)

Q T U

THE PLAYFAIE CIPHEK SYSTEM, ETC. 185


AP BN OH EN AO EA GI OE EB WQ GE UD
clu

ri

ng

th

es

rao

er

ne

-t

AN NS XE OU UA DT BN OA EP
on

wa

NI

se

dt

o-

KB SN HL DY PY HS NY

SI

QC

WE

ma

n-

rs

sq

su

ri

rp

he

jpa

YE EB
ea

-n

no

OS

EN PP VB NV OB NX GN XE OU OA FL
ro

ie

OA EY

SC

SO KT
n-

se

FE TA

CS

ur

in

VS BN
ri

QN

as

ZS

er

g.t

he

th

er

IN

TH XU QN FL EN

j*o

nw

as

nq
th

to

GE GB TE CN OE EC

X~U

IZ

CI

ne

an

SD VN NS GE AE GB

on

ig

OE OT

on.

nt

oq

EN OM

rm

se

ea

IG

om
th

HQ GE OA BH NS ES EC

on
e.h

CI

hs

me

qs

to

EA EG

sq

DN KN XD TF

OH EN EA EB
ng

wa

se

hu

EU XU UO QE NS
at

ho

s.w

FQ

rt

he

as

he

i<3

d-

EA EE HN
er

re

ne

e-

wh

YC NU VK DA DY PY EX XU UY
a.w

st

as

FIG. 4.

Thus, building up the square from the pairs underscored in Fig. 4, we have:

Q T U - P
E E A B D

WX
H K M
N

apparent that column 5 of the square should be


transposed to left of column 1 (this does not at all
It is

affect the equations obtained), 'and the fifth and second


"
considerab."
lines are seen to yield the key- word

CRYPTOGRAPHY

186

P Q T U -

D E R A B

WX

H K M

CONSI

The letters now remaining to be placed are V, F, Z, L,


and Y.
Carrying this process farther, we come to YE eg, and

may

safely

fill

in the square

P Q T U V

D E K A B

We now
"
is

L Y
F G

WX

H K M
S

complete the solution (Fig.

5).

The key- word

considerably."

So long as the relative order remains the same, we may


transpose line 5 to the top of the square without affecting
the equations obtained in enciphering or deciphering.
This would give us

CONSI
PQTUV
DEEAB
LYWXZ
FGHKM

FGHKM
CONSI
PQTUV
DEEAB
LYWXZ

or

of the above squares are equivalent,


probable that the key was used in the second
All

and it is
and fourth

lines as in the last square.


It is

now

interesting to note

how

the Kules held true

message not prepared by the writer. All of the


commonly used groups of the cipher are listed below, with
in a

THE PLAYFAIE CIPHEK SYSTEM,

187

seen that the digraphs


frequency very closely, and that in no

their equivalent digraphs.

follow normal

ETC.

It is

uncommon

case does a repeated group indicate an

EN =th
BA =er
4

BN

CI 2 =is

NS 4

pair:

=ri

on

GE 4 =he
AP BN OH EN AO EA GI OE EB WQ GE UD
Du ri ng th es er mo* ne* da* yt he pa
AN NS XE OU UA DT BN OA EP NI YE EB
rs

on

wa

sq

su

ri

rp

KB SN HL DY PY HS NY
ma

no

fw

el

ql

kn

ow

se

dt

os

eq

ea

SI
ns

QC

WE

CS

FQ

po

rt

in

EN PF VB NV OB NX GN XE OU OA FL
ro

cl

OA EY

iv

it

ie

SO

SO KT

qe

ni

nc

FE TA

CS

se

hd

ur

VS BN
ui

ri

in

ho

wa

se

sq

DN KN XD TF

hu

re

hs

la

CI
is

ph

Id

om

EN OM
th

ig

HQ GE OA BH NS ES EC OE OT
gt

OH EN
ng

sw

gp

IG

th

rm on
EA EB IN TH
er
nw
so
ea
he

se

an

do

ne

nq

XU QN FL EN
as

to

Id

th

EU XU UO QE NS XU GE GB TE CN OE EC
at

as

qs

oq

on

as

He me

nt

io

ne

dn

SD VN NS GE AE GB IZ EA EE HN
me mb er ed wh
ti
re
ca
on He
EA EG CI YC NU VK DA DY PY EX XU UY

QN

ZS

to

xi

er

eh
*

is

lo

st

um

Evidently a group

Q was

br

left out,

el

meant

used as a
FIG.

5.

ql
for

null.

aw

as

"One day."

qx

CKYPTOGKAPHY

188

One claim made in favour


pairs, such as

their

is

normal frequency, due to their chance

prior to enciphering.

up

split

that

common

th, er, on, etc., will not be enciphered in

up when the message


th is

of the Playfair

it

will

is

of being split

divided into two-letter groups


is well based, but when

This claim

probably yield another

common

Furthermore, even though a large percendigraph, he.


tage of these digraphs is split, their frequency in the
It is interesting to examine
cipher is still relatively great.
the following table prepared from the above problem:
Times Occurring

Digraph

th

..

er

on
he
as
ri

..

re

is
'in

Times Represented

in Message.

4
4
4
4
4

..

..

6
6
8
7
7

4
4
3
U

in Cipher.

..
.

,.

..

3
3

2
3
1

CODES
Closely allied to cryptography

is

the use of codes.

In

naval, military, and diplomatic circles, secrecy is the


principal objective, and the utmost care is exercised to
secure the codes from inspection by unauthorised persons.
In commercial codes, the chief aim of the compiler is to
provide an economical means of intercommunication

with overseas business houses. It frequently happens,


however, that important firms have to send messages

where secrecy is essential.


One rough-and-ready method is to substitute for the
actual code word opposite the required phrase another

THE PLAYFAIE CIPHEK SYSTEM,

ETC.

189

code word

so many forward or back as may be arranged.


This system, however, would present very little difficulty
to anyone who wished to break the message, and a much

safer method is to cipher the numbers which appear


against the code words.
This may be done simply by means of a key-word.

For instance,

if

the numbers of the code words you wish

to cipher are:

22350

49861

and the key-word selected

is

Buy another
123

4567890

with Z for repeated figures, the message would be ciphered:

UZYNEAEHOB
A more

elaborate system of ciphering is to have a


for the conversion of the figures with

series of tables
letters,

compiled as follows:
00
01

=AB

-AC
02=AD
03==AE
and

so on,

up

to 99.

obvious that any pair of letters in the alphabet


be used to represent any pair of figures, so that the
variations of this form of ciphering run into millions,
It is

may

even when, in order to comply with the International


Telegraph rules with regard to pronounceability, vowels
and consonants are used alternately. Where economy
in transmission is not an important factor, these variations
can, of course, be increased to an enormous extent, and,
without a knowledge of the code used, a message ciphered
this system is practically unbreakable.

on

190

TEST MESSAGE IN CIPHEE

BNHGYKZJ
KLVEDEST
DZRKICXF
GFIDWTCO
YVPJEGFA
HICKDROQ
SBKEQHJK
EBIZSOBQ
CAXSDPED
OPHYXHEJ
SOBLJGET
FXEZEKLJ
TJEHSQLO
ABEDZFYT
SJETVTYD
CRDYFEZF
SFTYZCBE
DRZLUSIX
EMNYCHCR
NTEFTRYD
RJPARXEW
HLOCKLJE
QUDCAKNS
CUMKJOKS
WLFSONFR
TLXINVHM
UCNZLYDF
OTRIWVAV
RABMVUCL
VUKLXUDI

ELKOCWVD
ABLNYVSG
YTANBECB
WCWLADXE
BLXUBNQC
UZXHVEFG
ORLVDICM
SDATWRAL
RSBIXFEG
ACBNUMOF
EDRXLAVB
UTKYBSEN
WVFRUGUT
SJUJGADV
VLHYZBUK
ICJTFOVS
DLTEMRGU
PEQCFAGF
IVTBIZBC
XPMAQBML
DCYVSOSB
RTCLOCHZ
CYLABRFO
LASPBIQZ
GICLMDUG
EPZLFURS
ALKDENKI
YCANTEDP
FYRLGEXV
FQRBONCD

ARBGEXD!

VIWCOCR
ZBEWGBHJ
RYNQGXY,
EZLGAPHi
DQYNZNY
GIWFYKN,
HVISTVA]
YPTAHQP]
V R D
R T L
ICSZFED]
DNJOCKL,
REHXZBI
XFOBGHI
BRAVENH]
ATLFINC]
KCSJAWE]
JEFGUDZ
MEQMYWL1
ONBPANCi
MLIVCHU1
AMUKRXLI
JDSEXVY]
EVLJMAN5
I

A K A V Z

QUCKIBN2
SCLURZUL
XDUZVNII
ROFUGAL*
FECORXEI

CEYPTOGEAPHY

TDRIQU GHNAZAJG
E ZXFNOK
KWI TI BMD
x DWIMCG
KOWTFXEN
L WOC MTU
ZECWTGOS
L IMJDTO
LBVCUSU^D
o BLNEKU
VJGSITVA
DUNCDLOX
E WQVAVL
N EZVCPA
VFEBNJCU
x RNYFLZ
EQBUNCSI
c FHGUVG
RSEWLFAJ
C R ULMXJO
ZTCBATRL
IW F CGIFXO
ZCZLEGLM
T UDKVYN
PSOJLCDE
F X CIBSGE
NCDLIWDX
E PFLICH
CLATCNDU
E X TASWEL
TZUKBPFA
V A RHITVO
SRCWICHB
A XGPBYN
CTREZCVJ
D MBYCXZ
DARVKINM
L U KRXJEP
LUCKMRIG
K L AZHZYG
VIRVDUXC
R
HNYSPY NCHUVEQR
W M YCHKBE ZBUTSUCR
Y C HLBYVW
ONDEJSQF
P
TVHCYZ NEFCTOGK
C A TGLSET
CLYRTOCL
M D ESMACK
VYCFLNOL
C D SYFLON
LCXODIGF
C ERHBYB
NHOXDLJU
IN

1C

IL

'IS

;A

IS

S J Y Z N F E
Y H D R A J X
J H P U D F N D
Q A W M W X E K
H B A R D T U Z
D F B Y R G P G

X A P M
Q
S U L V Q F
K Y M
A V J B
U B A C R T I M
F G
J

G L A

Y D A J
R Z L J
Q U V
Y R T E D
Q U

A K

M B

A X V

K M

T L D N A J T

N V

F R L

B U P B R G

191

V Q

E Z L T V Y

F E Z J F

F A

N F E

T
S

TSAR

L J A F Y B L
1

M E R
M U R T
L

V U

H D I V S A Q M
B U V L B N I G
H R
B E T Z C
G H V D A N C D

192

CEYPTOGEAPHY
CONCLUSION

I should hesitate to say that a cryptogram can be


invented that will defy solution, provided it is of reasonable length and is not so involved and intricate as to
make its use inexpedient. For practical purposes a

cipher should be upon some system which can easily be


committed to memory, and it should not involve any

expenditure of time in coding messages. The


cryptogram on page 190 has been ciphered in accordance
with these rules on what I believe to be a novel principle,
and will, I venture to think, require a great deal of pains
and patience to solve. In the event of failure to solve
great

the cipher, at a reasonable time after publication the


method of ciphering and the solution may be obtained

from the Marconi International Code Co., Marconi House,


London, W.C. 2, or the Marconi International Code
Corporation, 2236, Park Eow Building, New York, on
receipt of a stamped and addressed envelope.

PRINTED IN GREAT BRITAIN BY


BILLING AND SONS, LTD., UUILDFORD AND ESUER

'.

4-?

...

Wjgmorc Street

LONDON .W.i

HUIliMll>IPHllimilililMi HUH minim

niiiiiiiiniaii

mil

iiiiiiiiiiiiiin

HIM

You might also like