(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Cryptography .."

:?• 









\MEl)NIVER% 




, , ^ ^ o 

<rii]ONvsoi^ 



vvltSANCElfj^j. 




%a3AINn3WV^ 



^^l-LIBRARYOc. 



-v^lllBRARY^ 




^<!/OJI]V0JO^ 



"^.i/ojnvDjo 



r-n 



A\\EUNIVER5'//- 



v^lOSANCElfj> 
o 




<rii3DNVsoi^ "^jiaAiNn-awv 



>;,OfCAllF0%, ^OFCAllFOfi! 




^<?Aav}ian-^^ 



^^Aavaan-^^ 






-^ILIBRARYQc. A^lllBRARYOc 




^\\E•UNIVER% 



^OdllVDJO"^ ^OJIIVDJO"^ %130NVS01^ 




v^lOSANGElfj 




■^/5aaAiNn-3v; 






.^OFCALIF0%, .>;,OFCAIIF0/?^ ,^WE•UNIVERS•//, vvlOSANCElf. 












%a3AiNn]Vi 



% 



^ 



AWEINIVERJ/a 







^10SANCEI% 

§ ^^ — .^ 




%a3AlN^•3W^^ 



^^UIBRARYQ^ -v^lllBRARY<: 




%0JnV3JO'^ ^tfOJITVDJC 



S 
^ 



AV\EUNIVER% ^lOSANCElfj-^ 




<r7iinNV<:m?4^ 






^OFCAIIFO;?^ 




>^;OFCALIF0/i' 







^ 



zini izini i^-^i 1^^ 



>&Aava8ii^^ ^d^AHvaaiH^ 



"^-TiUDNVSOV^^ %a3AINn]WV^ 



AWEUNIVERS/a 




o 



vS:lOSANCElfj> 
3 




"^AaJAINH-l^W 



^IIIBRARYQ^ 



^ILIBRARYQa 




^<!/0JnVDJO>' 



%oi\mi^'^ 



.\\M\imm/A 



^lOSANCElfjv 

o 




<rji]QNvsoi^ %a3AiNn-3WV 



^OFCAllFOff^ 



^OFCALIFO^^ 




:5 -— '> I' § 



^CAavaan-^^ 



v^UIBRARYQ. ., 




^tllBRARY^^ ^\WEIJNIVER% ^lOSANCElfj> 



^<!/ojnv3jor' ^<i/odiiv3jo>^ ^riijoNvsoi^ ^/^a3AiNn-3Wv 




^OFCAIIFO/?^ 




^OFCAIIFO% ,^'rtE•UNIVER5•//i 






vvlOSANCElfjV. 




<rjl30NVS01^ %a3AINn3WV 



AWEUNIVERi-ZA 







^VOSANCElfj> 

CO 




-< 

"^aaAiNH-awv^ 



^v^lLIBRARYQ^ 



^IIIBRARYQ^ 




so 



< 



< 

O 



v^lOSANCElfx> 

O 




^OFCAIIFO/?^ 




.V5o I uwunn ^\Vy■ 



^OFCAllFOff^ 



cm 

Uul 

>■ 



m 



CRYPTOGRAPHY 



CRYPTOGRAPHY 

BY 

ANDRK LANGIE 

TRANSLATED FROM THE FRENCH BY 

J. C. H. MACBETH 

AUTHOR OF " IHI' MARCONI CORE," " MARCONI DICTIONARY," ETC. 






CONSTABLE ^ COMPANY LIMITED 
LONDON BOMBAY SYDNEY 

1922 

%. P. DUTTON AND COMPANt 



Printed in Great Britain 






PREFACE 

1 HAVE no intention of writing a complete manual of 
cryptography. Finality, even very relative, is not 
attainable in the domain of this art. Besides, good 
manuals are in existence on this subject, and the titles 
of some of them will 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 role played by cryptography in history. 

In the second part I relate how I succeeded in decipher- 
ing a dozen cryptograms of various kinds. In some 
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, 
I have, on grounds of expediency, substituted an approxi- 
mate reading for the actual text, and have modified the 
plan, and even radical features of the narrative, in such 
a way as to render abortive any attempt at identification 
and localisation. 



8789C7 



VI PREFACE 

In the third part I give sorae advice in a general way 
on Hnes which have proved profitable to me, and, further, 
a certain number of tables and formulae; but while I 
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 
assistance I have received in preparing this work from 
the late Mr. W. Jarvis, Lieut. -Commander W. W. Smith 
of Washington, U.S.A., Mr. Albert M. Smoot of the 
Ledoux Laboratories, New York, and Miss A. Wishavt 
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 forgiveness. 

J. C. H. MACBETH. 



CONTENTS 

PART I 

I'Aiii; 

Dksciui'tiun ok the Fkincii'ai, Systems of Grypt- 

OGKAl'HY, NVrrii niSlOUICAL NOTICE - - I 

PART II 
Examples of Deciphering 47 

PART III 
Lists and Tajiles 122 

BiBLIOCRAPHY 158 

PART IV 
The P[,ayfaih Ciphek System, etc. - - - 159 



ACKNOWLEDGMENT 

Our thanks are due to the following gentlemen 
in connection with translating the book from the 
original French, working out and substituting 
English " Examples " for the French ones, 
adding additional matter, and seeing the work 
through the Press : 

Mr. J. C. H. Macbeth. 
The late Mr. W. J. Jarvis. 
Mr. H. G. Telling. 
Commander Smith, U.S.N. 
Paymaster- Commander J. E. A. 
Brown, C.B.E., R.N. 

THE MARCONI INTERNATIONAL CODE CO., LTD. 
Makconi House, Strand, 

London, W.C. 2. 
'Ibtli July, 1922. 



CRYPTOGRAPHY 

PART T 

DESCRIPTION OF THE PRINCIPAL SYSTEMS OF 
CRYPTOGRAPHY, WITH HISTORICAL NOTICE^ 

I. 

Everyone has, at some time or other, used crypto- 
graphy,^ or secret ciphers. 

Who has not had occasion to make some note, or to 
correspond with somebody, by dotting letters in a news- 
paper or book ? Even children amuse themselves in 
this way on their school desks. But a pen or pencil is 
not necessarily required to make 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 /lair, his ear, his ^ips, his 
ear, and his none, whereby we understand " Helen." 
We 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 

^ This Part I. appeared in the BibliofMque uvwerseUe et revue 
Suisse in August, September, and October, 1917. 

- From the Greek words Kpi'irTns^ secret; and yf>ii(f)(ii'. to write. 

1 



2 CKYPTOGEAPHY 

make them unreadable to others. Particularly, more- 
over, it enables two persons to correspond under cover 
of complete secrecy — ^at least in theory. A man will 
perhaps invent, on his own account, a system of writing 
by means of which he can write and preserve secrets 
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 w^as as simple as he was noble- 
minded. He used it to note in his diary his qualms and 
trials. The phrases he wrote in this way can be read 
almost at a glance. 

The celebrated Swiss physiognomist, Jean-Gaspard 
Lavater, in his Diary of a Self-Ohserver, constructed 
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. Eight years after his death 
his countrymen had 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 Napoleon I., and which was supposed to have some 
historical importance. Here is a specimen, a part of the 
first line and one word of the sixth : 



PEINCIPAL SYSTEMS OF CRYPTOGRAPHY 3 

My friend had consulted his colleagues : one had declared 
it was not Sanscrit, another that it was not Ethiopic, 
and still others that it was neither Slavonic nor Runic. 
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, 
occurred something over ninety times, while another, the 
fourth, occurred seventy times. 

Now it is well known that in English, French, German, 
and most languages of Western Europe, the most fre- 
quently occurring letter is e; the letter which follows is, 
in French, n or s, according to the writer; in German, n; 
in English, t; in Italian, i; and in Spanish, a. In Russian 
the most frequently occurring letter is o, but i if the 
language is written in Roman characters. In Polish the 
most frequent consonant is z; not uncommonly three 
may be found in llie same word. In Arabic and Turkish 
the letter |, elij, corresponding to the French stopped 
or silent h, occurs oftenc^st. In Chinese — at least, in the 
newspapers — the characters found in order of frequency 
are ;5l {chi, of, genitive), 4^ (puli, not, negative), 
and X i^^ong, work). To ascertain which letters occur 
oftenest in any language, one must " calculate fre- 
quencies." 

The next thing to do is to study which letters most 
commonly adjoin. They are es in French and en in 
German. The most frequent groups of three are ent in 
French, the. in English, ein in German, che in Italian, etc. 
Bulky works have been written on this subject containing 
long lists, more or less complete, of the various articula- 
tions and disarticulations of words. Of course, this 
requires an enormous amount of labour, involving a 



4 CEYPTOGKAPHY 

statistical study of texts containing many thousands of 
letters. 

To revert to our example, I encountered a difficulty at 
the first onset. The sign which came second in order 
of frequency, and which I supposed^ should represent 
either n or s, caused me some embarrassment. 

At last I abandoned the books I had been using, and 
began a new calculation of the frequency of letters in 
certain authors and French newspapers. In the letters 
of Voltaire I noted that the letter u occupied the second 
place in point of numbers, this being obviously due to the 
fact that the words 7ious and vous (" we " and " 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 7i, 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 fol- 
lowed by e occurs principally in the syllable que. It 
could not be the word lequel here, the sixth sign not being 
similar to the first. The group must read: Ce que. A 
little farther on we meet again with the sign represent- 
ing c, followed by the r of tho word heureux and preceded 
by e, a group of letttTS which might, for instance, form 
the words ecran, decret, or, better slill, ecrirc. 

1 Thn ck'cipliermont is based not only on statistics, but also on 
hypotheses. In fact, the fanions expression, " Suppose that . . .," 
is the niotto of the cryptographer 



PKINCIPAL SYSTEMS 0¥ CEYPTOGRAPHY 5 

The result of the deciphering showed that there was 
no question of a conspiracy in this mystical writing, hut 
of the enthusiastic sentiments inspired in the author hy 
some charming person met at a fashionable party. It 
was, perhaps, the rough draft of a letter. The first phrase, 
translated, was as follows: 

" What I am writing you here is merely to relieve my 
heart, since I am writing to the dearest object in my life 
to divert the frightful restlessness of my days. . . ." 

And so on for twenty-five lines. 

Cryptography has provided an entertaining field for 
novelists. They produce heroes who mark in 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 dis- 
covery. 

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 life of an innocent 
victim condemned to death ? 

In A Voijacje into the Interior of the Earth, also by 
Jules Verne, we see a Danish scholar intent on piercing 



6 CEYPTOGRAPHY 

the mystery of a cryptographic parchment which is to 
reveal the route to be followed in order to penetrate into 
the depths of our terrestrial globe. But old Professor 
Lidenbrock seeks too far, and it is his nephew Axel, a 
careless young man, who attains the object simply 
enough by reading the finals of the lines backwards. 

It may be pointed out that the system deciphered by 
Edgar Allan Poe is comparatively simple. He himself 
acknowledges this, and claims to have deciphered keys^ 
" ten thousand times " more arduous. The system 
conceived by Jules Verne in his A Voyage into the 
Interior of the Earth is also very easy. As to that in 
Jangada, the problem is solved, thanks to an incredible 
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. 

A final 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 
been deciphered, probably not many edifying subjects 
would have been found therein. But it was burnt during 
the troubles of the Holy League.'-^ 

Some months ago I received a letter from a foreigner, 
who informed me that he was very interested in crypto- 
graphy, and that he wanted to work on some oliicial 
texts. He begged me to lend him some diplomatic 
documents, of which he would take copies for his use 

^ The ■■ key " in cryptography is the formula which enables a 
text in cipher to be read. 

2 Henri Martin, IJisloire de France, vol. ix., p. -±72, note. 



PKINCIPAL SYSTEMS OF CRYrTOGIiAPHY 7 

and return nie tlio originals. "You do not know me," 
he wrote, " but you can rely on me entirely: I am 
ncuiml.''' 

Despite this excellent recommendation, I had nothing 
to connnunicate. However, touched by his candour, I 
gave him some advice. Living in a large town, he had 
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 
characters, and try to decipher them. It is a crypto- 
graphic exercise as good as any other. There are certain 
methods which enable one to guess which letter means 5, 
which 0, which 9, etc.^ I refrained from pointing them 
out to him, since the value of these exercises lies precisely 
in finding out these methods for oneself. I wonder 
whether he followed my advice, which I consider was 
good. 

We w^ere just now recalling some specimens of secret 
writing whore 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 
inks, the employment of 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 
them immediately. Conventional or shuffled alphabets 
alone are of any use. 

An example of a cryptography wddely in vogue in the 

1 Note by Translator. — This is for decimal currency. In 
ciphers representing £ s. d., the same methods would first disclose 
6, 1, and 0. 



8 • CEYPTOGEAPHY 

Middle Ages is furnished by the so-called alphabet of the 
Freemasons, of which the following is a specimen : 



A 


E 


D 


F 


1 


H 


B 


G 


C 




N 


a 


R 


S 


T 


U 


9 


Y 


P 




The following words will be read without difficulty : 

utnaur"^L_L.urn 



By writing the alphabet in a different order, the 
values of these angles and squares may be altered 
at will. 

In December, 1916, I was given a bundle of papers in 
Spanish cryptography to decipher. It was a private 
correspondence written in the above system, a little 
complicated: there was not only one dot, but two or 
three in most of the letters. 

Among many other ingenious systems may be men- 
tioned that known as the " zigzag," which is constructed 
thus: Take a sheet of paper ruled in squares, and write 
along the top of the vertical columns the whole alphabet 
in any order you like. Having done this, superpose on 
the page a sheet of tracing paper, on which mark a dot 
in the vertical column headed by the letter required, 
taking care not to mark more than one dot at a time in 
each horizontal line. The dots once marked, you join 
them by zigzag lines, and send your correspondent the 
drawing thus obtained. 



PEINCIPAL SYSTEMS OF CRYPTOGRAPHY 9 

An inquisitive intermediary would see nothing in it, 
especially if the document were brief, whereas the 
recipient will place the message received on to a graph 
similar to that of the sender, and will have no difficulty 
in deciphering: 

PFC I OQXDERKYJLSZVMTAUGNBHW 



" I love you dearly." 

He (or she) will reply in the same way, or perhaps by 
means of a thread. This is laid along the cipher alphabet, 
beginning on the left, and wherever the thread passes a 
letter required, it is marked in ink. Arrived at the right 
extreme of the alphabet, the thread is moved a section, 
and a new start is made from the left, and so on in- 
definitely until all the letters required have been marked. 
Thus, to write the words " I reciprocate" would require 
something over seven sections of the thread, each cor- 
responding to the length of the alphabet.^ This method 

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



10 GEYPTOGKAPHY 

of corresponding is very ancient. It is some 2,300 years 
since .Eneas, the tactician, recommended a similar system 
in his Poliorcetes} 

Shorthand is not the same thing as cryptography, of 
course; it is not a secret writing, seeing that hmidreds of 
thousands of persons can make use of it. Nevertheless, 
an artful mind can combine shorthand and cryptography 
in such a way as to form a fairly complicated secret 
writing. 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 shorthand- 
like characters which had resisted the efforts of several 
professional shorthand writers. Here is a fragment : 

It 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 vety well organised postal service. 
Their letter-box was purely and simply the backs of the 
volumes lent them by the prison library. They had 
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 con- 
cealed in the hollow space in the back of the binding to 
fall out. 

They had other hiding-places all ready in case of alarm. 
Their alphabet consisted of more than 200 different 
signs. ^ I will dwell a moment on the contents of these 

^ Chapter xxxi. 

2 The method of clccii)licring iipplied here was to calculate the 
frequency of the various linos and curves. 



PEINCIPAL SYSTEMS OF CKYPTOGEAPHY 11 

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

These two baiulits, M. and S., had drawn up well- 
schemed plans of escape, and were on the eve of carrying 
them out when the unforeseen contingency of the decipher- 
ing occurred. They had organised their future move- 
ments, and projected the burglary of a jeweller's shop 
in ail important Swiss town in order to get on their feet 
before proceeding to effect a gigantic coup at a jeweller's 
in the Kue de la Paix, Paris, or in Eegent Street, London. 

" Take the small cases," wrote M. to S., " but never 
jewels displayed on velvet trays, for jewellers have a 
trained eye, and can at a glance detect whether any piece 
is missing from a tray. The large dealers always have 
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 : 

" I should work in first-class railway-carriages: operate 
on solitary individuals, but never with a dagger, you 
understand ; nor with revolver or chloxoform. 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 friend- 
ship, on Schopenhauer and Nietzsche, on the destiny of 
the soul, etc. Occasionally there is a postscript: "Ask 
our worthy chaplain to borrow my fountain-pen for you. 



12 CEYPTOGEAPHY 

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 
further writes: " Not one out of a hundred shorthand 
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 this afternoon." Its actual reading is: " Achtung ! 
Zange wird nachmittags am 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 seen, cryptography is of service to private 
individuals — that is, to certain private individuals — but 
its main usefulness lies in furnishing a means of cor- 
respondence between heads of States, Ministers, and 
Generals. In wartime, especially, by its aid plans of 
action and secret information can be communicated, 
relief asked for, etc. Cryptography, when employed for 
diplomatic or military purposes, is termed " cipher," 
whether it be in the form of ciphers or figures, letters, 
or any other signs. Obviously, when war conditions 
prevail, only Government departments and military 
authorities are in a position to utilise cryptography,^ 
which is of incalculable value to them. 

^ The author should have said " legitimately." Tt is a matter 
of common knowledge that numberless attempts were made by 
spies to convey information to the enemy by means of more or less 
ingenious cij)hers. In most cases these attempts were foiled by the 
ingenuity of the expert staff of cryptographers employed in the 
various Cipher Departments. — Tkanslatur. 



PEINCIPAL SYSTEMS OF CRYPTOGRAPHY 13 



III. 

The origin of secret writing is lost in the mists of anti- 
quity. To go hack 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 
of LacediTinion l^y means of a message traced on wooden 
tal)li'ts covered with wax. At first nothing could be 
seen on them, and it was Gorgo, the wife of Leonidas the 
King, who discovered the stratagem."^ The Cartha- 
ginians made use of a similar process, which seems to 
indicate the employment of sympathetic ink.^ Hero- 
dotus^ has recorded for us a not very practical system 
which was once employed in the East. " Histiseus, 
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- 
graph had not yet arrived !) "As soon as this occurred, 
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-t?)." 

This rather slow means of correspondence was not in 
current use. At the same period the Spartans had a far 
better system of cryptograph}^ the scyiales, of which 

1 Herodotus, VII. 239. 

2 Aulu-Gelle, Niiifs atliqnes, XVII. 9. 

3 V. 35. 



14 CKYPTOGEAPHY 

Plutarch,^ among others, has left us a description. The 
scytale was a cyHndrical 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- 
pipes. On the ^vrapper thus formed he traced the 
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 slightest difference in the 
diameter of the two rods made 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 
characters at the end of 75,000.000,000 years. If chance 
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 
the end, or, worse 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 com])aratively 
easily, even without Ix'iiig in })()ssessi()n of I he desired 

1 Life of Ly Sander, eh. xix. 



PEINCIPAL SYSTEMS OF CPtYPTOGKAPHY 15 



cylinder. Let us suppose that one of those messages has 
fallen into our hands, and that its twenty-five centuries 
of age have left it preserved in its original state of fresh- 
ness. We begin by making an exact copy, which we 
shall manipulate in our own way, bearing in mind always 
to leave the originals intact. Prom one of the ends of 
this copy let us cut off, say, three fragments, each con- 
taining 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 
after various adjustments our attention is fixed on the 
following combination : 

We observe that of the two groups 
of three letters, W I L is capable of 
forming a part of the word ivill or 
ivild. To test this, we 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 L is 
another L or a D, we are on the 
right track. The trial proves this to 
be the case by yielding L. We now 
make a new copy of the papyrus 
and cut it 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 : 



D 
E 
E 
A 
W 
T 



B 




R 




I 


L 


T 


H 


I 


N 


S 


P 




P 




L 



16 



CRYPTOGKAPHY 



D 










E 






E. 


B 


e. n. 


you r. g u 


A 


R 


d. t h 


e. e n e m y 


W 


I 


h 


1. a 


r r i V e. a 


T. 


T 


H 


e. f 


r n t i e 


r. 


I 


N. 


a. w 


e e k. a n d 


i 


S. 


P 


1 a 


n n i n g. t 




P 










L 







Drawing nearer the Christian era, we are told by 
Suetonius, the biographer of Juhus Caesar, that the latter 
" employed for secret matters a sort of cipher which con- 
sisted in writing, instead of the required letter, the third 
letter from it, as D for A, and so on."^ 

" The Emperor Augustus," says the same historian, 
" when he writes in cipher, puts B for A, C for I), and 
so on for the other letters, and A A for Z."^ 

Julius Caesar's cipher is still in use in our day — that 
is to say, it remains in principle, but with complications 
which make it much harder to decipher. Alfred I., 
King of England, and Charlemagne also used crypto- 
graphy for corresponding with their officers. I do not 
think I am violating a diplomatic secret, a thousand 
years having elapsed, in revealing that in Charlemagne's 



secret writincj 



/ meant i ; / / / d, yC ^' 



Ccesar, eh. Ivi. 



2 Augustus, ch. Ixxxviii. 



PEINCIPAL SYSTEMS OF CRYPTOGEAPHY 17 

etc.-^ The Governments of Venice, Florence, and other 
Italian republics made 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 do Vigenere, the French diplomatist, whose 
ingenious table is still useful to-day, either for coding 
or decoding. Cardinal liichelieu, the great statesman, 
frequently resorted to cryptography. Louis XIV. used 
so complicated a cipher for corresponding with his 
Minist(>rs when they were absent from Versailles, or 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, Rome, 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 Chiffrierhuro, staffed 
by professional experts, whose mission is to find new 
ciphers, both complicated and safe, and to decipher the 
secret documents of the enemy. The newspapers in- 

^ G. Selonua, Cri/plomem're, p. 282 (Alriiin). 



18 CEYPTOGRAPHY 

formed us that in February, 1916, the Department at 
Vienna employed twentj'-six cryptographers. 

" Cryptography," said one of the most genial of Swiss 
Army commanders to me the other day, "is a German 
science. You must be a German, wear gold spectacles 
and a 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 re- 
course to one of those little States which they so utterly 
despise.^ 

Each step in the progress of cryptography is accom- 
panied by a corresponding step in the art of deciphering. 

History has preserved the names of some celebrated 
decodists. Thus, the geometrician Francois Viete suc- 
ceeded 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.^ 
The latter angrily denounced Viete to the Holy See as a 
wizard and a necromancer. According to them, he could 
only have entered into possession of the secret by calling 
up the spirits of those who had known the cipher during 
their earthly career. But the Pope was a man of humour : 
he submitted the plaint to examination by a commission 
of Cardinals, " with urgent recommendation." The 
Cardinals understood the hint, and the examination is 
still unfinished. 

^ See the Ziircher Post, Februar}' 28, 1910, midday edition, and 
the Bund, Febniary 29, 1916, Sup. No. 100. The military Court 
at Zurich, after seeming to hesitate subjeetively over this point in 
a paragraph of its judgment, a(hni1 ted it objectively in a.notli(>r 
paragraph. 

- I)c 'riioii. I/lsloire imiverspllp.. Book 129, year 100,3. 



PRINCIPAL SYSTEMS OF CRYPTOGEAPIiY ID 

During the reign of Louis XIII. another decodist, 
Antoine Rossignol, made himself known, to the dis- 
comfiture of the Huguenots. 

" It was in the year 1626," says Charles Porrault,^ 
" at the siege of llealmont, a city of Languedoc, then in 
possession of the Huguenots, that he first gave proof of 
his talent. The city w^as besieged by the army of the 
King, commanded by the Prince de Conde, and it opposed 
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 
in the art of deciphering could make nothing. It w^as 
given to M. Rossignol, who deciphered it forthwith, and 
said that the besieged were sending to the Huguenots 
of Montauban to say that they were short of powder, 
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 
being reported to Cardinal Richelieu, he invited M. Ros- 
signol to the Court, and the latter gave such astonishing 
proofs of his skill that the great Cardinal, despite that 
extraordinary disposition which prevented him from 
admiring many things, nevertheless could not forbear 
expressing his surprise. He (Rossignol) served very 
usefully during the siege of La Rochelle, discovering 
the enemy's secrets by means of intercepted letters, all 
of which he deciphered with scarcely any trouble." 

He continued his activities under Louis XIV., who 
held him in such high esteem that once, on the way 
back from Fontainel)leau, he called on him at his country 

' Lea Hnmmes ilhiMres qui ont jinrv. ■poxlanl cc {llUi) ftiecle. 
Vol. i. Antoine EnssignnL 3Iaislre des Com'ples. 



20 CEYPTOGKAPHY 

house at Jiivisj to which he had retired. The poet 
Bois-Kobert addressed many of his epistles in verse to 
Kossignol, in one of which, in accordance with the wishes 
of Cardinal Eichelieu, he extols Kossignol's skill, regard- 
ing him as a redoubtable prodigy. The following is a 
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 secret plans of kings undone. 
This 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 thj' pay; 
Hell's unseen imps their packets steal, 
Their secrets to thine eyes reveal." 

There is a certain amount of truth underlying this 
extravagant eulogy, not that an Antoine Eossignol 
would wish it. Colonel Schaeck, of the Swiss General 
Staff, has stated that " a good decipherer must have 
both natural and acquired qualifications, the former 
necessarily playing a predominant part. The natural 
qualifications are insight, the spirit of observation, 
patience, and perseverance. If a person be happily 
gifted in any d(»gree for this kind of work, and finds an 
opportunity of developing his natural talents, he may 
attain by study and practice a surprising degree of skill. 
For this he will have to devote himself to a profound 
study of the various systems of cryptography, have a 
thorough knowledge of mathematics, and especially the 
calculation of probal)ilities, and be acquainted with 
languages and their lilcraturos." 



PRINCIPAL SYSTEMS 01^^ CKYPTOGKAPHY 21 

Two remarks may be added to this statement: First, 
in default of mathematics, wo may be satisfied with 
aritlnuetic; secondly, one thing is indispensable, which 
Colonel Shaeck possessed, although he modestly refrained 
from mentioning it — common sense. I have heard of 
a case where iifteen months of assiduous research failed 
to produce any result, while, a lit lie later, by the exercise 
of a. little connnon sense, the goal was reached in two 
days. 

In 1G45 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 
game.-"- 

On July 2, 1673, Louvois, the then French Minister 
of War, paid GOO livres, equivalent to £120 sterling, to 
one Vimbois for discovering the cipher of certain con- 
spirators; four days later he prescribed a similar fee to the 
Sieur de la Tixere for a discovery of the same kind.^ 
If those lines meet the eyes of any cryptographers, they 
will regretfully admit that the remuneration for their 
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 Courtelary, a village 

^ Encydopcedia Britannica, art. Cryptography. 

- Valcrio, De la cryplographie, vol. ii., p. 11. 

^ The amour-propre of the cryptographer does not always meet 
vvitli the respect due to it. For instance, a cryiitogram which I was 
charged officially to decipher in May. 1917, resolved itself into 
" . . . for the fool who reads these lines." 



22 CEYPTOGEAPHY 

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 ihc Royal Academy 
of Science and Literature of Berlin} 

It was by methods used in cryptography that Miinter, 
a Dane, and Grotefend, a German, succeeded in 1802 
in deciphering a part of the alphabet of the Persian 
cuneiform inscriptions. One group of angles or arrow- 
heads struck them by its frequent repetition. Miinter 
pronounced it to be equivalent to the word " king " 
[Kh-shayathiya in the harmonious language of the time), 
and this supposition was eventually confirmed. 

Mention may be made also of Bazeries, a French of&cer, 
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. 
Thus, for example, the word "mine" could be written 
in these four ways — i.e., 



I. 


46. 


144. 




II. 


230. 


59. 


125. 


[II. 


514. 


184. 


374. 


IV. 


535. 


229. 


146. 



and by still other figure combinations. 

As we have seen, cryptography has at all times been 
extensively used by conspirators, revolutionaries, and 
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 Pontainebleau, which they had bought 
1 Year 17i58 (1765). pp. 369-389, witli two plates. 



PEINCIPAL SYSTEMS OF CEYPTOGRAPHY 23 

for tlio purposo of meeting secretly. Their plottings, 
however, were frustrated, as their house was raided, and 
among other suspicious objects were found a quantity of 
letters in cipher which revealed the conspiracy."^ 

" Among the papers of the Chouannerie," says M. G. 
Lonotre,'-^ " are to be seen a number of sheets written 
in bizarre characters 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, 
after interminable efforts, have had to abandon — accord- 
ing to plan ! A curious circumstance is that texts 
written in cipher are encountered even among the hiero- 
glyphs. A certain inscription of Esneh contains a 
profusion of crocodiles, in groups of as many as eighteen 
at a time, tiie meaning of which is not apparent. The 
most hardened Egyptologists have not yet succeeded in 
forcing the teeth of these redoubtable saurians apart 
and making them disgorge their secret. Certain mysteri- 
ous languages — perhaps Etruscan, for instance — might 
yield to cryptographic methods of decipherment. 



If the " black cabinets," or postal espionage offices, 
which were extensively used in France during the reigns 
of Louis XIV., Louis XV., Louis XVI., and Louis XVIIL, 
unsealed letters to feed the police reports and to furnish 
gossip to the Court camarillas, the black cabinets of the 
German Empire in the eighteenth century w'ere centres 

^ Dulaure. Singiilarith hisloriques, p. 303. 

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



24 CEYPTOGEAPHY 

of cryptography. Count Briihl, Prime Minister of 
Augustus III., Elector of Saxony, organised a completely 
equipped establishment at Dresden. All the messages 
received or sent b}' 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 terri- 
tory, 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 
to unseal. The Court locksmith was under orders to go 
to the Legation and, with the connivance of the Prussian 
Secretary, force the lock of the chest in which the Prussian 
Minister kept the kej^s of the ciphers.^ 

Thanks to their laudable activity, Saxony was aware of 
the plans of Frederick II., and, when needful, comnmni- 
cated them to Austria and Eussia. Count Briihl, how- 
ever, gave the game away at an official dinner, when he 
indiscreetly mentioned something he had learnt through 
his perverted laboratory. Frederick II. changed his 
systems of cryptography, and thenceforth entrusted his 
correspondence solely to functionaries who were abso- 

• SchWzers Staatsanzeigen, I'lul (52, p. 12'J el spq. 



PKINCIPAL SYSTEMS Oi^^ OliYPTOGEAPHY 25 

luti'ly hi'yoiid suspicioji.^ IWii he did not complain, for 
he himself had for some time carried on the same kind 
of espionage, which gave him a tangible advantage over 
his opponents during the Seven Years' War. 

Austria, moreover, did not lag behind, and — a master- 
stroke — her black cabinet was operated in a wing of the 
Imperial Palace of the Stallburg, at Vienna. The staff, 
who were Neapolitans and well versed in work of this kind, 
directed their energies to the correspondence of all the 
Ambassadors. On one occasion the deciphered copy 
was placed in the official cover addressed from Madrid 
to the Spanish Ambassador, instead of the original letter 
extracted therefrom. The Spanish dij[)lomat lodged a 
complaint with the Austrian Prime Minister, Prince 
Kaunitz. The matter was serious, and might have 
involved grave consequences, so the Prince severely 
reprimanded the negligent official. 

The work done in the black cabinets cannot be 
accm'ately termed " cryptography," as they merely 
deciphered cryptographic documents by means of the 
key, which they were quite incapable of discovering 
for themselves. 

***** 

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 
this subject. I need say no more than that, of all those 
1 have read, the most substantial is the work of a French- 
man. 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 

^ The secretary changed his name and souglitother fields for his 
talents. 



26 CEYPTOGEAPHY 

documents which are in any degree complicated, any 
more than the best of 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 them completely, so as to 
produce chaos. 

1. In the Substiiutional class — that is 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 and the thread, and of 
the two thieves. 

Here are some others: The Hebrew cabalists had 
several cryptographic ciphers, which they used principally 
to discover the hidden meaning of certain passages in 
the Bible, Thus the Atlibasli, the Hebrew spelling of 
which forms the key — ^A.Th.B.Sh — consisted in writing 
the last letter of the alphabet T\ {thaw) instead of the 
first letter X (alej^li), and the last but one ^ {sldn) 
instead of the second 1 {beih), and so on. The applica- 
tion of the Athbash resulted, among other instances, in 
identifying under the place-name Bheshak^ that of Babel, 
or Babylon. 

Another system, Alham, consisted in replacing the first 
letter of the first half of the Hebrew alphabet J< {alejjlL) 

* Jer. XXV. 26. 



PKINCIPAL SYSTEMS OF CllYPTOGEAPHY 27 

by the lirst kitter of the second half of tliat alphabet 
^ (lamed), and the second letter of the first half 2 [beth) 
by the second letter of the second half D {mem), etc. 

In a third system, the Athakh, the interchange of the 
hitters was based on their numerical value. But I shall 
not dilate further on this, as that clever Hebraist, J. 13ux- 
torf, has explained the whole thing far more clearly 
in Latin than I can in a modern language. Those 
desiring further details are referred to his book.^ 

IJacon 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 AAAAB AAABA 

for ABC. The method of deciphering a document 
written in this way is obvious enough: the frequency of 
the groups must be calculated instead of that of the 
letters. In the example given below, representing the 
last letters of a message, and, according to the most 
plausible supposition, the termination of a feminine 
Christian name, 

ABAAA BBBAB ABAAA 

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

^ De Abbreviaturis Hebraicis, Baslo, 1613, pp. 24, 27, and 37. 



28 GEYPTOGEAPHY 

E=A, 0=B, V = C, P=D, H-E, etc. The difficulty 
then becomes apparent. 

By making use of the cipher square, or Vigenere's 
table, it is possible to write in cipher by means of 
several secret alphabets, as many as fom-, five, six, or 
even ten or more at a time, in periodical succession. 
Here are the first few lines of Vigenere's table^ : — 

lABCDEFGH IJKLMNOPQESTUVWXYZ 

Aabcdefghij k Imnopqrstuvwxyz 
Bbcdefghi jk Imnopqrstuvwxyza 
Ccdefghij klmnopqrstuvwxyzab 
P I d e f g h i j k 1 m n o p q r s t u v w x y z a b c 
B e f g h i j k 1 m n o p q r s t u v w x y z a b c d 

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. 
We now write the word " hieroglyphics," and under 
each letter add a letter of the word BEG in consecutive 
order, thus : 

hieroglyphics 
B E G B E G B E G B E G B 

The next thing is to look in the above table for the 
letter H in the horizontal line of capitals, and for B in 
the column of capitals on the left ; at the point where the 
two lines commanded by these letters intersect we find 
the letter i, which we write as the first letter of our 

1 The complute lablu will hv ioiiiid ou page 155. 



PRINCIPAL SYSTEMS OF CRYPTOGRAPHY 29 

ciphered word. The same operation ior the letters I 
and E yields m, which we take as our second letter, and 
so on. The finished result appears as follows: 

hieroglyphics 
BECBECBECBECB 

i ni g s s i m c r i ni e t 

Thus the word ." hieroglyphics," written hy the aid of 
ihe key-word BEC.is transformed into " imgssimcrimet." 
It will be noted that of ihe three i's in the cryptogram 
only two stand for the same letter in the plain text; 
the same is true of the three wi's, wdiile the two s's also 
represent different letters. 

Let us examine another cryptogram of the same order: 

tapvccigrqduprbhitvcc a c e o e a o s c 
a 1 i e c c. 

Given the knowledge that this text has been ciphered 
by means of Vigenere's table and that the key-word is 
PIANO, we operate by reversing the process described 
above — that is, we write the key-word PL\NO repeatedly 
under the letters of the cipher; then, looking for the 
tirst letter of the key-word P in the vertical column to 
the left of the table, we find in the line corresponding 
thereto the first letter of the cipher t, and at the head 
of the column in which this occurs we note the capital 
letter 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 ciplicr. I and a respectively, 
we obtain S, and so on until we have before us the whole 
text deciphered as follows: " Espionage compensation 
scraps of old iron." 

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



30 CEYPTOGEAPHY 

do not know it, however, there are certain methods, a 
httle too long to explain here, which permit of its dis- 
covery almost mechanically. All that can be said is 
that in the above cryptogram, as well as in any other 
secret document, the first thing to do is to find the 
vulnerable point in the armour and attack it with the 
weapons at your command. 

Now, in the text we have just deciphered, the weak 
spot is the doul)le letter cc, repeated three times, and it 
is this which will help us pierce the mystery. After 
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. 

Let us now examine a somewhat different example. 
We are handed a document to decipher which reads as 

follows : 

M A S E G X 
I S OM OX 
AMOXEX 
G K Y Y M N K 
YKOSE K 

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

'The letter occurring most frequently is 0, which, 
according to the rule of frequency, should represent e, 



PEINCIPAL SYSTEMS OF CEYPTOGRAPHY 31 

hut it is in vain tliat we try to decipher the second line, 
in which it appears three times. If we adopt the hypo- 
thesis that we have before us a list of proper names, we 
arc; compelled at the same time to recognise how little 
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 
is the pivot in deciphering will not carry us very far, 
so that another method must be adopted for deciphering 
proper names — not only those we have just enumerated, 
but names in general. The freqiienqi of tlieir termina' 
tions in each language must be taken as the basis. In 
French, for instance, 8 per cent, of proper names end 
in er or ier — e.g., Mercier, Fournier, Garnier, Beranger, 
Boulanger; 7 per cent, in on — e.g., ond, ong, ant : Masson, 
Champion, Dupont, Leblond, Long; 6 per cent, in au 
— e.g., eau, aud, aut, aux : Boileau, Eousseau, Moreau, 
Clemenceau, Nadaud, Caillaux.^ In Russian, ov and ev 
terminate 35 per cent, of names; shj, 25 per cent.; in, 
9 per cent.; itch, 6 per cent., etc. 

Those who wish to take up cryptography and to indulge 
in these interesting calculations without undue mental 'H<vr uau. v. 
fatigue should confine their energies to Turkish family aJ' '^^*^' 
names — a by no means complicated task, for there ^''^^^T^^^^^XJ^, 
none ! In the Ottoman dominions all that is necessary, %j^ ^ 
even for official records, is to say that one is called John / 
the son of James, or Ali the son of Mustapha. I once ^'^"'^"^l 

1 Tt must 1)0 understood that tliesc proportious are only approxi- 
mate. 



32 CEYPTOGRAPHY 

asked a friendly Greek, 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 
Mustaphas. My interlocutor seemed surprised at my 
question, and answered: " Oh, there is no trouble at all 
in identifying anybody." 

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 
fathers. For instance, a certain Osman has a son named 
Taleb, 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, 
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.), 
(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, 
represents i. At the end of the third name we find two 
of these supposed ■i's separated by a letter not yet identi- 
fied. Now, as our study of proper names has gone con- 
siderably beyond the rudiments set out above, we know 
that ini is the most likely ending : Bellini, Rossini, Mazzini, 
])i Uudini, etc. We therefore assume that E=w. 

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



PRINCIPAL SYSTEMS OF CRYPTOGRAPHY 33 

in ntl (example: cantl, conti, santi), and that it might be 
a heading or the title of the list, perhaps agenti. Acting on 
this assumption, wo make the required substitutions — 
ready, of course, to try other suppositions if this fails 
us — and our cryptogram assumes the following form: 

AGENTI 
? E ? ? A ? I 
G A ? I N I 
T ? ? ? A ? ? 

? ? ? E N ? 

A moment's reflection induces us to substitute the 
letter r for 0, which occurs tliree 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 
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 w^e decide to replace it by m, and finally 
we have the following version: 

AGENTI 
FERRARI 
G A R I N I 
T MM AS 
I\I R E N 

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. G5. 13. 79. 52. 32. 75. 88. 79. 43. 22. 

stand for " Travaillez, prenez ..." (" Work, take 
. . ."). The numbers 54 and 09 each moan a; 13, 88, 

3 



34 CEYPTOGEAPHY 

and 43, e; number 52 means nothing; the first 15=v, 
the second 75 =r. The methods of deciphering here are 
so delicate, fragile, and awkward to explain that I prefer 
to leave them to the innate sagacity of the reader. 



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 word 
occupied by the letter in question, thus: 127.6.4.2. 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 
letter required. If you are using a French book, for 
instance, you may have to dispense with a fe. True, 
you might use a c instead, but this would sometimes 
lead to confusion. Suppose you want to write: " Kiel 
is besieged." " del (heaven) is besieged " is scarcely 
the same thing. Neither would your correspondent ever 
guess that in the phrase, " His Majesty ill ; cocher 
(coachman) summoned to general headquarters," the 
word " cocher " was intended for the famous surgeon 
Kocher. 

In Eussian books the letter / is also infrequent, while 
in Italian pul)lications w and y are rarely seen. 



Correspondence has sometimes becMi carried on in the 
following manner: Most dictionaries are printed with 
two columns on a page. Instead, therefore, of writing 



PRINCIPAL SYSTEMS OP CBYPTOGRAPHY 35 

the required word, you adopt the word appearing on the 
same Hiir in ih(! parallel column, thus: 

WADE instead of VISION 

THERMOMETER „ TERRIFY 

BELLICOSE „ BEARER 

ESTUARY „ EQUAL 

TORRENT „ TO 

OMIT „ OCCASION 

The word "terrify" appears here, but not "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 
cannot all be shown. Thus, with the aid of any dic- 
tionary you can say " SentZ letter," but not " Sent 
letter," which two phrases are diametrically opposed in 
meaning. 

Special dictionaries have been compiled, each page 
containing fifty words in current use. Thus, for instance: 

(page) 17 

23 GRADUAL 

24 GRANT 

25 (4RAYE 
2G GREEK 
27 GREEN 

If it is rccpiired to send the word " Greek." you write 
the number which precedt'S tliat word and the nninbcr 
of the pagi\ 17. '21. or the whole in one number, 1724. 

Much more voluminous dictionaries have been utilised 
or compiled, in which all the words are accompanied in 
.the margin In' numbers ranging, say, from 1 to 100,000. 
Let us endeavour to decipher the following crypto- 



36 CEYPTOGEAPHY 



gram, coded from a dictionary 


of 


25,000 numbered 


words : 








24133 


24029 




1512S 


21682 


01643 




21531 


05070 


24127 




09043 


21531 


02432 






01174 


15311 







The first thing to do — and it is not easj^ — 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 must be sought decoding 
a ciphered document. 

Let us assume that we have discovered the meaning 
of the last two numbers in the above : 

21531 =THE; 09043=GENEEAL. 

It will bo 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 
to stand for words Ix'ginning 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 
are W and T. Tjeaving this on one side for the moment. 



PllINCIPAL SYSTEMS OF CEYPTOGKAPHY 37 

however, we will study the end, where the last two words 
are assumed to have been definitely established as THE 
GENEEAL. Immediately preceding these, we note two 
numbers in the fifteenth thousand, which occur numeri- 
cally about half-way between those representing 
GENEEAL (09043) and THE (21531). We accordingly 
look in a dictionary, and find that the corresponding 
position is among the O's. Of words beginning with 
hkely to precede THE GENEEAL, we observe OF, ON, 
OPPOSE, and OE, and provisionally select the first, OF, 
corresponding to 15128. We now have 15311, another 
presumed 0-word, occurring later than OF in the dic- 
tionary. There are OPTION and OEDEE, of which 
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 OEDEE OF THE GENEEAL. 

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 
make the most of the " landmarks " already more or less 
identified. The fourth word, 21531, is known to be 
THE, and the word following is probably, though not 
necessarily, a noun. We note that the number repre- 
senting it, 01174, is the lowest of all, occurring doubtless 
among the A's. The message being of a military nature, 
we immediately think of AEMY and AETILLEEY, and 
look for a fm'ther clue. The next number, as well as 
the eighth, is 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 
a verb expressing something in connection with the 



38 CEYPTOGEAPHY 

supposed army or artillery. Consulting the dictionary 
under W, we are attracted by the word WITHDEAW 
or WITHDEAWN. If the latter is correct, it should 
follow some part of the verb " to be," and, in fact, the 
seventh number, 01643, occupying numerically a position 
something over a third of the distance between AEMY 
(or AETILLEEY) and BY, would seem to represent the 
w^ord BE itself. The sixth number, 24029, is a W-word, 
and both from the context and its numerical position 
a little earher than WITHDEAWN (24127) in the 
dictionary, it excludes any other reading but WILL. 

We have now to tackle the first three words of the 
cryptogram. The first number, 24133, closely follows 
that representing WITHDEAWN (24127) in numerical 
order, and the dictionary offers us as "probables" 
WITHIN or WITHOUT. After further study on the 
lines described, we produce WITHIN THEEE DAYS 
as the first three words of the cipher. All that remains 
is to decide whether the fifth number, 01174, means 
" artillery " or " army." The words occur so closely in 
the dictionary that this is no easy task, but after careful 
calculation of the distances separating " be " and " by " 
from the beginning of the dictionary, we plump for 
"artillery," and our complete message reads : WITHIN 
THEEE DAYS THE AETILLEEY WILL BE WITH- 
DEAWN BY OEDEE OE THE GENEEAL. 

It should be added that in practice such documents 
are not often found with the numbers written in this 
straightforward way. Usually the figures are transposed 
and all sorts of complications interspersed. 

A common method is to rearrange the order of the 
figures in each group upon a prearranged plan. Thus, 
24133, 21GH2, etc., are transformed into 13432, 02182, 



PRINCIPAL SYSTEMS OF CRYPTOGRAPHY 39 

etc. 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 
deciphering. It is a case of baling the ocean ! 

The principal inconvenience of those numbered dic- 
tionaries, kn(nvn in diplomacy as " codes," is that when 
they are lost or stolen, in most cases others have to be 
compiled, and works of this kind cannot be made in a 
day. Even under the most favourable circumstances, 
when a fresh code is held in reserve for contingencies, 
considerable delay must ensue before instructions for 
their use can reach those concerned, and the enemy 
reaps the benefit. The following is one instance, among 
others, of this disadvantage: During the Russo-Turkish 
War in 1877, the Ottoman Field-Marshal, Osman Pasha, 
entrusted one of his Generals, Selim Pasha, with a con- 
fidential 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 
so promptly on his mission that he forgot to leave the 
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- 
'positional, in which the actual letters are not changed, 
but are mixed together or shuffled, and in effect really 
amount to anagrams. Some anagrams are very short: 
veil for live, are for ear, more for Rome, wander for AndreW: 
Angelus for Galenus, Vaussore^ for Bousseau, etc. In 

* The pseudonym adopted by Rousseau when giving his famous 
concert at Lausanne. 



10 CRYPTOGEAPHY 

cryptography, however, we encounter anagrams of 
100, 200, 300, 500, and 1,000 letters. I have seen one 
comprismg nearly 6,000 letters. It may be added that 
the longest are not the hardest to decipher; quite the 
contrary. Among these systems, which are very 
numerous, are included the " scy tales " of the Lace- 
daemonians, which we have already considered. 

A system easy enough to decipher is one which 
the cryptologist Vesin de Romanini called an " aerial 
telegraph cipher." The first letter of the text is wTitten 
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 is made at the top, the letters being written 
in the same order as before, and immediately to the 
right — or left — of those already put down. A crypto- 
grapher will have no difficulty in reading a text ciphered 
in this way : 



ER 


TO 


HP 


SO 


BE 


TT 


NT 


GT 


00 


UE 


OH 


GW 


TA 


TK 


HE 



which means: THE STRONG OUGHT TO PROTECT 
THE WEAK. It was a simihir cipher which Jules 
Verne used for his cryptogram in A Voyage into the 
Interior oj ilie Earth. 

Let us now pass to the " grille," or " lattice," a well- 
known form of cipher. The grille is a square piece of 
stiff paper or cardboard in which a certain number of 



PEINCIPAL SYSTEMS OF ClIYPTOGllAPHY 41 

holes arc cut. The square thus perforated is super- 
imposed on a sheet of white paper, and a letter is written 
in each hole. This done, the grille is turned 90 degrees 
to the right, so that 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 
operation is continued until all four corners of the grille 
have occupied the same position. It need scarcely be 
said that, when 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. 

The following example can be read quite easily by 
means of the appropriate grille: 



U-..-.f 



wz/jy/MM/mm m 


; M 




E 


w ' 


|S C 






o . ' 



f. 


M 








p! 


f-: 




j 




L 


1 


"/// 


E 




T 




\ 


E 

1 


; D 


L 




1 



LLi 




Deciphered, this reads: KOME WAS COMPLETELY 
DESTROYED BY THE GAULS. 

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



42 



CEYPTOGEAPHY 



when one lias the proper grille. Xevertheless, even 
without the grille the difficulty of deciphering is not 
very great, and in the second part of this volume I shall 
explain the mechanism of the process by which it can 
be done. To complicate this cipher, a high Austrian 
officer had the idea of minglmg a number of blank or 
meaningless letters with the others; but this did not 
increase the difficulty of translating, as is proved by the 
fact that such a system is scarcely ever used. 

The method of employing dividers is much to be 
preferred. It consists of cutting vertical slices in a text 
and mixing these columns of letters. Here is a very 
short example — just three names: 

M A D E I D 
V I E N N A 
P A E I S 

which we divide into vertical slices, proceeding then to 
write first the letters of the second column, next those 
of the last, fourth, first, fifth, and third — or in some 
other order as agreed on. In this case the key will be 
2, 6, 4, 1, 5, 3: A I A D A E N I M V P I N S D E E. 

How is 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, he proceeds to draw the following graph: 




PlilNCIPAL SYSTEMS Oi^ CliYPTOGllAPHY 4:i 

crossing through tho last square, after which he writes 
in the letters of the cryptograms in the order indicated 
by the key — the second column, the sixth, the fourth, 
etc. — when the three names will be restored. On no 
account must he forget to cancel the last square, for if 
ho 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 PARIS, for 
instance, would be changed to lAMDR. This system 
can be complicated indefinitely. 

Let us examine another example: — 

V H I I N R 
U R S I N H 
P WP LT N 

K T D S S F 
Z E I C 
L E M B E 

This contains thirty-six letters, and is therefore very 
short; nevertheless, the numher of possible changes of 
position which one might effect in the letters, to ascertain 
th(nr meaning, is so enormous as to be practically in- 
calculable. For the sake of comparison, take wheat- 
flour. If we could isolate a particle, and under a micro- 
scope count how many such scarcely perceptible mole- 
cules could be contained in a cubic millimetre, we should 
find, let us say, 100,000. Now, to form a sphere as 
large as our terrestrial globe, it would require a number 
of these particles equal to that of the combinations which 
it is 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 is the same as that of the 
preceding cryptogram. Accordingly, the plain text must 



U CKYPTOGEAPHY 

be written six letters to the line, and the first letters of 
our example will form the second vertical slice of the 
graph. The plirase 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 
the present victories of our enemy are insignificant. ■^ 
He succeeded in a much more magnificent enterprise: 
he won the admiration of the enemy peoples. Decipher- 
ing by the graph produces the following: 

K V L P Z U 
T H E W R 
D I I\I P S 
S I B L E I 
S N T I N 
FRENCH 

The first line is composed of blank letters, intended 
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 mi decipherable cipher 
par excellence " — is very difficult to decode when one has 
to deal with texts more complicated than the elementary 
specimens we have just presented.^ It may be pointed 

^ Note by Translatok. — When M. Langie wrote this, the 
Germans were inflated with their niilitar}^ successes. 
2 The actual words are: " Le mot impossible n'est pas fran^ais." 
^ One method of complication, calculated to exercise the patience 
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 oiler absolute security, and one 
runs the risk of entangling both oneself and one's corresjjondent. 



rRINCIPAL SYSTEMS OF CllYPTOGRAPHY 45 

out that the second and longer example, being regular, 
is less difficult to decipher than the first, \vhich, though 
shorter, is irregular. 

There are systems in existence which are literally 
undecipherable, the ciphers being sometimes composed by 
moans of ingenious machines resembling the cash regis- 
ters of the shops. But — there is a " but " — it is probable, 
.111(1 often certain, that systems absolutely undecipher- 
al)le 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 
is nearly always urgent. As soon as it is a question of 
urgency, resort has to be had to telegraphy^ or radio- 
telegraphy. Now, in a long alignment of letters which 
are nieaningless to him, the most skilful of telegraphists 
will commit involuntary errors — due to inattention? 
fatigue, or reflex movements. And when a telegram 
runs into a number of lines and has to be retransmitted 
several times, the case is worse. 

It is stated that not 10 per cent, of telegrams in cipher 
are free from errors on their arrival. In the first place, 
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 (.), 

or the two letters m ( ) and t ( — ) into the single 

letter o (— ). It is precisely these extra or missing 

letters that do all the mischief. 

One error is sometimes sufficient to make the whole 
text meaningless, as we have seen in tlie example 
MADRID. Hence, if one must use systems very diffi- 

^ Inoidentaily it may bo pointed out that the telegraphic alphabet 
is nothing else but a sj'stem of cryptography. 



46 CEYPTOGEAPHY 

cult to decipher, it is none the less indispensable to 
choose keys in which one error will not cause a reper- 
cussion throughout a document. Furthermore, it is not 
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 be- 
wildered. 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 om^ efforts, the more profound will be the satis- 
faction we experience when we reach the goal, and. like 
Archimedes, exclaim " Eureka !" 



PART II 

EXAMPLES OF DECIPHERING 

A Consultation. 

One day a gentleman sent np his card and was shown 
into my office. 

"It is to an unfortunate accident that I owe the 
pleasure of making your acquaintance," he said, very 
affably. " What has driven me to seek enhghtenment 
from you is this : I have been sent here on a mission ; you 
will excuse me from going into details. Arriving this 
morning, I had scarcely entered my hotel when a postman 
brought in a letter addressed to me. Now, it was an 
understood thing that those who sent it must write in 
cipher all communications of any importance. It is a 
wise precaution, for you will see, if you examine the 
envelope, that it had been opened by steam, stuck down 
again, and immediately dried. By whom ? None of the 
hotel people could or would throw any light on the 
subject. 

" The cipher agreed upon between us is a grille. T did 
not bring the grille — it might have gone astray — but I 
had noted on an old letter a method for quickly recon- 
structing 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 



48 CRYPTOGEAPHY 

to read the secret message transcribed on to another 
square placed under the perforated sheet. Every night 
I destroy papers which I no longer want, and I fear I 
may have inadvertently burnt the letter containing the 
key in question. I was able to get your address without 
difficult}', and am come to beg you to bring all your skill 
into play, so that I may know the contents of the message 
without delay." 

While saying these words, he handed me the ciphered 
text, which ran as follows^ : 

a i t e g f 1 y t b o e e h r e a u w n a n o a r r d r t e e t 
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 first idea to occur to me. 

I begged my visitor to call again at the end of an hour, 
and immediately set to work. First I copied the text on 
to a square divided into sixty-four sectional squares, like 
that appearing below. I numbered the four corners in 
Roman numerals, and further added Arabic figures to the 
sixty-four squares for the purpose of easy identification. 

The principle of the grille system has already been 
explained on p. 41. I revert to it merely to point out 
that the grille, numbered at the four corners in Roman 
figures, should fit exactly over the text, the corner of the 
grille numbered 1 covering corner I. of the text, corner 
2 of the grille corner 11. of the text, and so on. If, in 
our example, the first hole in the grille exposes the letter 
A (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 boon jiunlifi -d to 
meet the roquiromonts of the English tr,inslatioi\. 



EXAMPLES OF DECIPHERING 



49 



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



1 

A 


2 
1 


3 

T 


4 

E 


5 

G 


6 

F 


7 

L 


8 

Y 


9 

T 


10 

B 


1 1 



12 

E 


13 

E 


14 

H 


15 

R 


16 

E 


17 

A 


18 

u 


19 

w 


20 

N 


21 

A 


22 

N 


23 




24 

A 


25 

R 


26 

R 


27 

D 


28 

R 


29 

T 


30 

E 


31 

E 


32 

T 


33 

H 


34 




35 

s 


36 

H 


37 

F 


38 
P 


39 

E 


40 

T 


41 

A 


42 
P 


43 




44 

T 


45 




46 

Y 


47 

H 


48 

L 


49 

R 


50 

E 


51 

T 


52 
1 


53 

H 


54 

E 


55 

N 


56 

E 


57 

M 


58 

G 


59 

A 


60 




61 

A 


62 

R 


^3 

N 


64 

T 



IV 



III 



In this case, however, the grille was missing, and my 
visitor, in his embarrassment, had asked me to reconstruct 
it. How must I proceed ? 

I take a piece of tracing paper, cut a little larger than 
the above square, and sup(n'inipose it thereon. On the 
tracing paper I reproduce the outline of the square as it 

4 



50 CEYPTOGKAPHY 

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 A T 
(squares 1 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, 
not a quarter only, but a half round, so that it is now 
reversed, and corners 1 and 2 cover corners III. and IV* 
of the text. The marks made on the tracing paper now 
coincide with E T (squares 62 and 64). This is a very 
good w^ord-ending, and it is evident that from the last 
two lines we could easily extract the word H E A E T — 
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 ATE (or F) T (or 0) E. This not being 
very satisfactory, I abandon the combination and try 
another. 

Having seen that the tracing paper is in its original 
position, I turn my attention to the second line, and 
decide to mark THE (squares 9, 14, 16). This is con- 
ceded by investigators to be the commonest trigram in 
English, and is almost certain to occur in a text of sixty- 
four letters. Eeversing the tracing paper as before 
brings the marks to the corresponding squares 49, 51, 56 
in the sixth line, indicating the letters E T E. This is 
quit(i a promising combination, and I look for the vowel 
which should precede it. The first that meets the eye is 
(square 45), while three scpiares farther back (42) 
appears P. We now have the group P C) E T J*j, which 
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 68, 64) are possible. 

I now reverse the paper to ascertain what letters corre- 
spond to the new marks, and bring to light T (square 3) and 
N (20, 23). We now have the group T THEN 0, the 
first letter being doubtless a final, and the last two the 
beginning of a new word. The next proceeding is ten- 
tatively to begin the construction of the grille, which I 
do by drawing squares round the letters PORTER 
(42,45, 49, 51, 56, 62). 

Let us now take crayons in four colours — say blue, 
red, green, and yellow. With the blue crayon we make 
a small mark in the text itself against each of the above 
six letters. We now^ turn the tracing paper, but this time 
only a quarter, and our six marked squares now cover the 
letters I U S A G (2, 11, 18, 35, 41, 58). To these we 
attach a brown mark. Incidentally, it may be noted 
that the results of the quarter turn, unlike those of the 
complete half, are not necessarily to be read consecutively. 

A further turn brings us to the group T T H E N 
(3, 9, 14, 16, 20, 23), which- we mark in red; and a final 
turn produces L A E H E N (7, 24, 30, 47, 54, 63), which 
we indicate in green. 

We have now ncnitralised 24 out of the 64 squares, 
thereby narrowing considerably our field of research. 
Coming back to our original group PORTER, we look 
for a likely word to precede it, and are favourably inclined 
towards THE (32, 36, 39). There are two H's between 
the T and the E, and we adopt the second experimen- 
tally. Marking these and reversing the tracing paper, 
we find the three corresponding letters to be R T H 
(26, 29, 33). This enlarges our red group to T THE 
NORTH — a result which proves that we are on th(> right 



62 CEYPTOGEAPHY 

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 
s<5;uares. For instance, on scrutinising our two red words, 
THE NOETH, and the unmarked letters following them, 
we quickly discern a P and an 0, and think of " Pole." 
These letters are 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, 
we find that only twelve squares remain to be 
accounted for. 

The materials for the grille are noAV 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, now appear as 
follows : 

Eed: TTHENOETHPOLE 
Green: ELBEWA E FT HE AN 
Blue: EANDTHEPOETEE 

Brown: IF QUAE S AYINGA 

A glance at our text shows Y (square 8) to bo the only 
unmarked letter between the F and in the brown line, 
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 com- 
pile the cutting of the grille. 

At this juncture my visitor appears. " Have you dis- 
covered anything ?" he asks eagerly. 



EXAMPLES OF DECIPHERING 53 

" Here is your grille," I reply. " Let us read it to- 
gether: IF YOU AKE STAYING AT THE NORTH 
POLE HOTEL BEWARE OF THE MANAGER AND 
THE PORTER." 

But our friend suddenly looks grave. 

" Good heavens !" ho exclaims, " I have some important 
papers in my portmanteau." And with a bound ho dii- 
appears downstairs. 

It is to bo hoped he arrived in time. 

Where is the Money ? 

The Chief of the French Secret Service Department had 
invited me to call upon him, and after giving me a cordial 
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 
years' imprisonment. Had he taken scrip payable to 
order or bearer, we should have been at ease, but he has 
confined his attention to cash. 

" Now, a fellow of his stamp would hide the stolen sums 
in such a wa}' as to be able to find them intact on regaining 
his liberty, and he is too keen a psychologist to have 
confided his secret to an accomplice. He has w'orked 
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 w^ore 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 
justice. 

" Unfortunately, six days elapsed between the robbery 



54 CEYPTOGEAPHY 

and the arrest. We have been able to trace his move- 
ments dm'mg the last two days, but are still in the dark 
as to how he employed his time dm'ing the first four. 
Being determined to leave no stone unturned to recover 
the money, I have continued a strict investigation. 
Yesterday I visited the central establishment housing 
am 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 ivishes of the 
prisoner, and in his presence. 

" The text of the will was somewhat to the effect that 
its author bequeathed his watch, a ring, the contents of 
his purse, and his personal effects to his brother and sister, 
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 figm'es. 
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 ! 

" Here is 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', 
Pastoure has written sums in addition, subtraction, 
multiplication, and division, with erasures everywhere. 
Please take the document, examine it at leisure, and let 
me know at your convenience what you think of it." 

I put the will into my bag, and, having arrived home, 
studied it with eag(^r curiosity. 

I began by verifying the results of the arithmetical 



EXAMPLES OP DECIPHERING 55 

operations occurring therein. At first glance the whole 
was a confused medley. The numl)er 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- 
gested by a preliminary examination. I counted all the 
figures on the page, and found there were 144. I then 
made a list of them, column by column, as follows: 

201 01 0458787243667351 87()93()1 771)52984 
847201 224675294851 85272364522361 21 1 1 
588643110225011813414375659521401593 
579131544939454441021252448100642212 

These 144 figures were distributed in the following order of 
fi'equency : 

Figure: 12 4 5 3 7 6 8 9 

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

I became more and more convinced that there was a key 
to be discovered here. But the frequency of the ten 
figures led to nothing. In French, even in a long phrase, 
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. The famous line 
of Voltaire's, " Non, il n'est rien que Nanine n'honore," 
where the ?i's and ^'s form nearly half the total number of 
letters, absorbs 12 different letters. It is scarcely possible, 
except perhaps in Runic, to write phrases with a total 
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 



56 CEYPTOGEAPHY 

were never more than two groups alike: two lO's, 
45's, etc. Twenty-one groups were repeated, and 
occurred only once. 

Matters were not making much headway. The letter 
occm-ring, in French, on an average once in six letters, 
or 17 per cent., I had no means of deciding which of the 
21 repeated groups could represent that letter. I then 
divided the 144 figures into groups of tliree; there were 48. 
But my disappointment was, if possible, still greater, 
only 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. 

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 
certain groups, which followed at intervals of 1. Thus: 
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 : 
984. 9 + 8 + 4=21. The groups following this in value 
gave: 952, or 9 + 5 + 2=16; 939, or 9 + 8 + 9=21. 

"Hallo!" I said to myself, "none of these groups 
seems to exceed 21 when I add the three figures composing 
it." And I 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 
usage being v, the 22nd. Perhaps, after all, each of the 
three-figure groups represented the order of a letter in 
the alphabet. 

Accordingly, I made the trial, and added the figures 
of each group. This resulted in the groups 010 and 100 



EXAMPLES OF DECIPHERING 57 

. .producing a total sum of 1; Oil and 110, each 2; 

111, and the two 201 's, each 3. The group which 

iuced the largest sum was not, as I had at first sup- 

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

ly new hypothesis, the plain text probably contained 

the letters a to v, if the language were French. I was 

disconcerted, however, to find that the most frequent 

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

equalling i, according to my theory. " It might be Latin," 

1 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 
in French, where, after e, the most frequent letters are 
n, a, i, r, s, t (onl}^ the i of our three supposed letters has 
any place hero), but also in Latin, where the letters 
should occur in the following order : i, e, s, u, a, n, o, r, 
etc. Still we have here the i, u, and o ; and, in any case, 
it was desirable to put our supposition to the test. 

The first three figures, 201=3, meant c; the next three, 
010=a; then 453=12=1. The complete text proved 
to be: " Calvisius, Opus Chronologicum. Bihliotlieque 
Municipale."^ 

" So he has deposited the sequel to his secret in a 
volume," I thought. " And with a psychological fore- 
sight by no means stupid, he has not trusted to his 
memorj^ having had experience of the transformations 
which memory can effect in a name after a certain number 

1 Note by Translator. — There are two errors in the cr^-pto- 
gram, the final letter in Calvisius being represented by the group 
936= 18 = r, and the third letter in Municipale by the group 
454=13 = m. 



58 CEYPTOGEAPHY 

of years. Furthermore, he has chosen, as the receptacle 
of his confidences, a kind of work which is among the 
least consulted in a collection of books. Old books on 
law or theology are sometimes referred to, but ancient 
manuals of clironology are generally allowed to sleep in 
peace." 

The same day I repaked to the Bibliotheque Munici- 
pale, and 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 Pas- 
toure must have made some secret entries on the first or 
last pages. There could be no question of sympathetic 
ink, since the necessary manipulations to make it visible 
were scarcely possible in a public reading-room. I 
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 
latter completely open. I perceived no note or anything 
else. Still reflecting, I looked at the inside cover. I 
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 witli these objects, I rL'tuniud for my Calvisius 
and, operating in the same way as Pastoure must have 



EXAMPLES OF DEOIPHEKING 59 

done, but vice versa, I slipped the little sponge, soaked 
ill 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 its hiding-place a small safe key, 
which bore the name Bauche and a luimber. I re- 
gummed the paper, and, having returned the volume, 
wont 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. 

" Why, yes," I replied. " They have enabled me to 
find this little metal object." 

Picture the astonishment, then joy, of the Chief ! He 
made me describe point by point the development of 
my discovery. Then he started on the chase, accom- 
panied by his sleuth-hounds. Two days later, on 
opening my newspaper, I learnt that the thirteen hundred 
thousand francs which had been stolen had been re- 
covered from the strong-room of a bank, where a com- 
partment had been rented for fifteen years by a client 
about to start for Australia ! 

Arabic Numerals. 

I will now give another instance of success in the dis- 
covery or key to a cryptogram. It was in Arabic nu- 
merals. One day I received in the usual buff envelope 
the following text:^ 

67534 34959 61496 54860 46495 14564 46496 25350 
65646 04950 45664 45966 49664 56649 60494 96646 
59665 06249 50536 65060 57496 85849 

^ See footnote on p. GO. 



60 CKYPTOGEAPHY 

I began by arranging the numbers in order of importance, 
from the lowest to the highest : 

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 tw^o-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 .■■• 

A striking peculiarity is the preponderance of 6's and 
4's, which occur 30 and 24 times respectively, whereas 
1, 2, and 7 each occur only twice, 8 three times, and 3 
four times. 

Perhaps it is possible to group the figures differently. 
Can we divide the text into groups of four figures ? No, 
because there are 110 figm-es. Neither can we form 
groups of three. We can, however, form 55 groups of 
two figures: 

67 53 43 49 59 61 49 65 48 60 46 49 51 45 

64 46 49 62 53 50 65 64 60 49 50 45 66 44 

59 66 49 66 45 66 49 60 49 49 66 46 59 66 

50 62 49 50 53 66 50 60 57 49 68 58 49 

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. 
Wo will, therefore, assume provisionally that 49=^. 

^ NoTK BY Translator. — At first siglit this ca-yptograni would 
be hard to distinguisli from the dielionary eij)her described on 
p. 35. 



EXAMPLES OF DECIPHERING CA 

We next note the frequency of the numbers 49 (supposed 
e) and 66 in the series 66, 49, 66, 45, 66, 49, 60, 49, 49, 66. 
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 c d, but the sequence would be: ? <? 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 ?ien?ne?een. The 
letter t 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 
tail-end of the preceding word. This promising result 
yields us the following four equivalents: 

49=e; 45=i; 66=w; 60=i. 

We proceed to make trials with these four letters, and 
observe that, preceding the word " nineteen," there 
occurs the group te9in?9ne (60, 49, 50, 45, 66, 44, 
59, 66, 49). The combination leads us to the idea that 
a date is in question, in which case there can be no hesi- 
tation in filling in the three blanks thus: " ted in June." 
Our theory is confirmed on examining the series following 
" nineteen "—viz., ?und^ed (46, 59, 66, 50, 62, 49, 
50), which is obviously " hundred." 

Before going any farther, we summarise the results 
so far obtained, to wit: 

d=oO; e=49; li=4Q; i=-45; j=44; n=66; r=62; t=GO; 

and are immediately struck ])y the fact that the numbers 
proceed in two regularly descending sequences, so that, 
without further trial, we are able to construct our 
alphabet : 



62 



/ 








>• 


CEYPTOGEAPHY 




a =53 


h=46 


0=65 


v=58 


b=52 


i=45 


p=64 


w=57 


c=51 


j=44 


q=63 


x=56 


d=50 


k=43 


r=62 


y=55 


e=49 


1=68 


s=61 


z=54 


f=48 


m=67 


t=60 




g=47 


n=66 


11=59 





The complete text of the cryptogram is then found to be : 
" Make use of the cipher adopted in June, nineteen 
hundred and twelve." * 

Once more I would ask my readers to refer to the 
Preface to this volume. 



In the Flour. 

Towards eleven o'clock one night, as I was about to 
retire to bed, I heard a violent ring at the door, and a 
moment later my maid appeared and smilingly informed 
me that a policeman was asking for me. She knows I 
am always pleased to see a policeman, whose visit usually 
leads to certain little expeditions, which generally result 
in something unexpected with a spice of adventure. 

I opened the door half-way and asked: "Who is 
there?" An unknown voice replied: "May I trouble 
you to come at once to you know where ? You will then 
receive instructions." I hastened to the rendezvous — 
an office in which were congregated officials, detectives, 
and policemen. We exchanged greetings, and then, a 
large motor-car having drawn \ip at the entrance, some 
of the party — myself included — took our places in the 
vehicle, which moved off at great speed. 

I tlioroughly enjoyed the moonlight trip. We rushed 
through villages and a wood and climlx'd a hill, con- 
versing in low tones all \]\o time. SuddtMily the car 



EXAMPLES OF DECIPHERING G3 

stopped; we got out and proceeded on foot to a solitary 
liouso, whore a bright hght gleamed from a window on 
the ground floor. Sonu'body knocked at the door. 
While waiting for it to be opened, the leader of the ex- 
pedition took me apart and informed me that a domi- 
ciliary search was about to be carried out in the rooms of 
a man who had been arrested that day, and w'ho had 
strongly protested, saying he had been entrusted with a 
diplomatic mission. In the papers to be submitted 
to me for examination I was to search for proofs of that 
statement. 

At last we were inside, and I was installed before the 
drawing-room table, on which documents in various 
languages were being piled. I buried myself in the 
tedious and wearying task of selection, putting on one 
side the screeds which seemed to deserve a more minute 
examination. I had no knowledge of the case or of the 
allegations against the arrested 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 
notes. It appeared that a considerably larger supply of 
provisions had been found in the rooms than was author- 
ised 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 so, kept separately in a writing desk, 
had puzzled him. 

Contemplating the white powder with an air of absorp- 
tion, the detective murmured: " Old flour sometimes 
contains worms; I wonder whether there are any here." 
And while " Mr. Deputy" and other functionaries looked 



64 CEYPTOGKAPHY 

on with interest, he passed the flom- through a sieve which 
he had just procured, letting the fine powder fall on to an 
open newspaper. Suddenl}^ a small cylindrical object 
appeared, a sort of case for steel pen nibs. With obvious 
deUght the detective examined this object, cleaned and 
opened it, and, to our astonishment, produced therefrom 
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 : 

YOUWOULDHAEDLYKNOWTHEEE 
WASAWAEPEOVI S I N SAEEPLEN 
T IFULA'NDQUITECHEAP. 

(" You would hardly know there was a war. Pro- 
visions are plentiful and quite cheap.") 

The other text began with the letters USLAAVI, 
followed by several more without any apparent signifi- 
cance, though the words IDOL and SHEBA stood out 
among the meaningless array. 

Eeturning to my seat, I glanced from time to time at 
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 
carefully pressed the two bands of paper into their case 
and put the latter into his pocket. 

Two hours had elapsed since our arrival. The exami- 
nation was finished and the seals affixed. The " Deputy " 
disappeared, and we rejoined our motor. Going off" at 



EXAMPLES OF DECIPHERING 05 

a good pace, utidt'f (lie ligl'' <d' (lie moon, I reached home 
al)Oiit two in the moniinj^. 

The word uslaavi kept running tlirough my head. 
Surely this was a Slavonic word meaning " glory." 
And could one connect it with the words " idol " and 
" Sh(d)a " ? It might be some ritual. On the other 
hand, the initial word of the mysterious writing might 
be a variant spelling of the Russian 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 
station 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 
without preamble into the business. 

" You see," he said, " these are the two documents 
seized the other night. Each contains sixty-five letters. 
One is evidently tlie transcription of the other, to which 
it has been attached in error. I have compared the 
frequency of the letters in each, and here is the result : 

Letter : A V. C D E F GUI ,1 K L M N P Q K S T U V W X Y Z 

Plain text :7013610340140453153341-1020 = 65 
Cipher : 71 3 57013402303710273310011 = 65 

We have to discover the key. I will not hide from you 
i\\e fact that it will be difficult. 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 wliich we can 
decipher them. 

" For instance," he continued, " both texts have an 



66 CEYPTOGEAPHY 

equal number of A's, H's, I's, and T's. There are n 
J's, M's, or X's in either. The cipher contains B, G, and 2 
which are absent from the plain text, and my theory i; 
that these and other redundant letters, such as sever 
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 in with what eagerness 
may be imagined. AYith the respect due to a relic, I drew 
the precious paper from my pocket case and began to 
study it. It read as follows: 

USLAAVIPICASDHOIOTOEIDOLY 
SHEBAHADADSTCESEENESONEZ 
T U K U K D G E L A C S N E 

A rapid glance led me to the conclusion that the tlu-ee 
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 
temporarily decided against the theory of a grille, which 
usuahy requires a square number. The pairs OE, AD, 
ES, NE, and UK, which were repeated, gave me the iclea 
of looking for a key-word (see p. 70), but the intervals 
between the repeated groups furnishing no satisfactory 
indication, I passed on to another hypothesis. 

I noticed that the letter occurred three times in a 
sequence of five letters, thus: — — 0; and that the same 



EXAMPLES OF DECIPHERING 67 

hing happonod with A: A — ^A — A. Tliis favoured the 
dea that the cipher had been composed with the aid of 
lihe system known as " dividers " (see p. 44) — that is, 
the required phrases had been written in very short hues 
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 I juxtaposed with the ol)ject of re- 
constructing the original text. As a nucleus I to(jk 
the two groups just mentioned, and arranged them 
in vertical columns, thus: 






A 


I 


H 





A 


T 


D 





A 



Of these pairs only OA could form part of an English 
word, but the other two pairs could each be the final and 
initial letters of separate words. I first tried the word 
" board." There was only one B in the cryptogram, 
and the group containing it was YSHEB. Juxtaposed 
with the above, this produced the series: YOA, SIH, 
HOA, ETD, BOA. The first trigram being unsatisfac- 
tory, I abandoned the word " board," and tried " coa(st)," 
" roa(d)," etc., but obtained no good result, one or other 
of the trigrams produced always proving an impossible 
combination. 

It then occurred to me that the five pairs alreadv 
marked off need not necessarily be consecutive letters. 
The pair IH, for instance, was very unpromising, but 
the insertion of S or T, making ISH or ITH, would 
yield a far more hopeful basis. Accordingly, I decided 



68 CKYPTOGEAPHY 

to interpose a third group between the lii'st two, and hit 
upon STCES. This produced: 

S A 
il T H 
OCA 
TED 
S A 

The second Hne could not be " wi</t," for there was no 
W in the cryptogram, but it might be " J f/iink." Adopt- 
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: 

S A V E U 

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

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 
series, as yet untouched. I looked again at the partial 
reconstruction. Certainly the words "save," "think," 
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, 
substituting the two groups OEIDO and USLAA. These 
could only be adjusted by moving them down one line. 
The word " think " was now preceded by OU instead 



EXAMPLES OF DECIPHERING 69 

of I, so I completed the word " you " by adding the group 
YSHEB, and now read the following: 









S 


A 


V 


E 


u 


Y 





U 


T. 


H 


I 


N 


K 


S 


E 


s 


C 


A 


P 


E 


u 


H 


I 


L 


E 


D 


I 


S 


K 


E 


D 


A 


S 


A 


C 





D 


B 





A 













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, 
following on from the ciphered text. This made the last 
line read: BOARDANG. Assuming that ANG was part 
of the name of a ship, the word " on " seemed the proper 
word to precede " board." To introduce this, I pre- 
j&xed 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 
sliding the column up two lines, the K of " think " was 
preserved and the third line became " has escaped." 

Success was now a foregone conclusion. It turned out 
that the original text had been written in lines of eleven 
letters, and had then been divided into vertical sections, 
of which the fifth had formed the first letters of the 
ciphered text, the eighth forming the second series, 
and so on. The first line, SAVEU, had to be 
abandoned, and the complete reconstructed text proved 
to be: 



70 . CKYPTOGEAPHY 



D 





Y 





U 


T 


H 


I 


N 


K 


Z 


H 


A 


!S 


E 


8 


C 


A 


P 


E 


D 


T 





C 


H 


I 


L 


E 


D 


I 


IS 


G 


U 


I 


S 


E 


I) 


A 


S 


A 


c 








K 





N 


B 





A 


K 


D 


A 


N 


E 


U 


T 


K 


A 


L 


V 


E 


S 


S 


E 


L 





(" 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 whom the message was intended, any more than 
the name of the person referred to as " Z." I concerned 
myself only with forwarding the whole — ciphered and 
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: 

i p b V d d z o b g q w w n z s c c z a f s t x 

V i y s d s X p t f h k t d d d s k 1j p f V p c 

V p a f s V k z f e j t V y b i p (1 o a a s y b 
a c r }) w h s m 1 s n c t g k n i y s x f v y c 
i plddlahvwccvpzdqagtcwdj 

There are 120 letters in the text. I note the following 
repetitions: ip, dd, cc, Jv, ds, sx, vj), vy, yp, ajs, 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 letters between the marks. I do the same 

^ Tlio reader is again rcfentd to tin- Preface and to the footnote 
on p. 48. 



EXAMPLES OF DEGIPHEKING 71 

with the other identical groups, and draw up the following 
table: 

From ip to ip G3 letter?^, or 3 X 3 X 7 



ip „ 


ip 33 


>> 


„ 3 X 11 






dd „ 


dd33 


5> 


„ 3 X 11 






dd „ 


dd 62 


)) 


„ 2 X 31 






cc „ 


cc90 


>> 


„ 2 X 3 X 


3 


X 5 


fv „ 


fv 48 


>> 


„ 2 X 2 X 


2 


X 2 X 3 


ds „ 


dsll 


>> 


„ 11 






sx „ 


sx 61 


?> 


„ 61 






vp „ 


vp 3 


>) 


„ 3 






vp ,, 


vp60 


J> 


„ 2 X 2 X 


3 


X 5 


^T M 


vy 33 


J5 


„ 3 X 11 






yb „ 


yb9 


>> 


„ 3 X 3 






afs ,, 


afs 31 


}) 


„ 31 






iys „ 


iys63 


>J 


„ 3 X 3 X 


7 




cvp „ 


cvp 60 


5> 


„ 2 X 2 X 


3 


X 5 



It will be noted that the factor 3 occurs in eleven out of 
the fifteen lines, so it is fairly safe to assume that a key- 
word has been used in coding the text, and that this word 
contains three letters. The question is: Can we discover 
this key-word and successfully decipher the text ? We 
begin operations by copying the whole of our text into 
three columns — that is, in lines of three letters, numbering 
each line to facilitate reference: 



(1) 


i 1) b 


(11) xpt 


(21) 


v y b 


(31) sxf 


C-^) 


vdd 


(12) fhk 


(22) 


ipq 


(32) V y c 


(3) 


z b 


(13) t d d 


(23) 


a a 


(33) i p 1 


(4) 


gqw 


(14) dsk 


(24) 


sy b 


(34) d d 1 


(5) 


w n z 


(15) bpf 


(25) 


a c r 


(35) a h V 


{^) 


sec 


(16) v p c 


(26) 


p w h 


(36'> w c c 


(7) 


zaf 


(17) vp a 


(27) 


s m 1 


(37) V p z 


i^) 


St X 


(IS) f s V 


(28) 


s n c 


(38) d q a 


CJ) 


V i y 


(1<)) kzv 


(29) 


tgk 


(39) g t c 


(10) 


s d s 


(20) e j t 


(30) 


n i y 


(40) w d j 



7-2 CKYPTOGEAPHY 

The first column begins with the letters i v z, and ends 
with d g IV ; the second column begins with p d o, and 
ends with q t d ; and the third column begins with h dh, 
and ends with a cj. 

The next thing is to calculate the frequencies in each 
column, which gives us the following table: 

First column: s, v, 7 each; d, i, lo, 3 each; a, J, g, i, z, 2 

each; h, e, k, n, o, p, x, 1 each. 
Second column: y, 8; d, 5; c, ^, 3 each; a, h, i, n, q, s, t, 

2 each; g.j, m, o, iv, x, z, 1 each. 
Third column: c, 6; h, 4; a, J, k, I, v, 3 each; d, t, y, z, 

2 each; li,j, q, r, s, w, x, 1 each. 

According to the law of frequencies, E is the commonest 
letter in English, followed by T or 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 
hesitate between s and v, either of which may repre- 
sent E. How can we arrive at a decision ? 

Looking down our table of numbered lines, we note 
that p (col. 2) is preceded by v three times (lines 16, 
17, 37). If, therefore, v (col. 1) represents E, we get 
the combination (lines 16 and 17) EE?EE, which seems 
unlikely. Eecalling that one of the commonest bigrams 
is HE, let us substitute H for E as the value of v in 
col. 1. In our list of repetitions we find the group 
c V p. If we adopt HE as the value of v p, we may 
easily infer that c v p equals THE, and this combination 
does, as a matter of fact, occur in lines 16-17 and 36-37, 
the c in col. 3 and v p in cols. 1 and 2 on the succeeding 
lines. 

If we are satisfied that we have established one equiva- 
lent in each column, we can immediately ascertain the 



EXAMPLES OE DECIPHERING 73 

key-word used from one or other of the ciphering tables 
at the end of this book, and armed with this word decipher 
the cryptogram automatically. The process will be 
explained in due course. 

Meanwhile, it will be interesting to see whether it is 
possible to effect the decipherment without knowing 
the key-word, and without reference to the ciphering 
tables. We will suppose that, for some reason or other, 
we have not at our disposal such useful adjuncts for 
finding a key-word, and that we are without any clues 
outside the cryptogram itself to help us in the deci- 
l^herment.^ 

So far, then, we have established the following: 

V (col. 1)=H; f (col. '^)=E; c (col. 3)=T. 

Our copy of the cryptogram, ^^Titten 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 H to all the u's in col. 1, 
E to the p's in col. 2, and T to the c's in col. 3. Lines 
16-17 attract our attention at once Avith the group 
HETHE, which looks like a part of the word " whether." 
We therefore tentatively add W as the equivalent of / in 
col. 3, and R as that of a in the same column, duly 
marking accordingly all the similar letters in the column. 
The next thing we notice is the group WH?T in lines 
31-32, and we decide to fill the blank with A, attaching 
this value to the three ?/'s occurring in the middle column. 
For the moment we cannot go any farther in this 
du'ection, so W'e fall back on the law of frequencies, which, 
however, might easily prove a pitfall if we did not recog- 

1 Note by Translator. — This experiment is not in the French 
edition, but is added here to amplify the example. 



74 CRYPTOGEAPHY 

nise the possibility of numerous exceptions. It will be 
remembered that the letters s and v each occur seven times 
in col. 1, so that we could not at first decide which was 
likely to stand for E. However, having eliminated v by 
attaching to it the value of H, and noted that the next 
letter after s and v in order of frequency in the column 
occurs only three times, 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, 
and another in line 31, and that in each case it is preceded 
by the letters i y (in the second and third columns of the 
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 H as the value of i (col. 2) 
and y (col. 3) respectively. 

We are now able to resume the thread of our internal 
clues with the group (lines 7-10) WE??HTHE, which we 
construe as "weigh the," thus obtaining two new equiva- 
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. 

s=E . i;=T a=R 

ij=H 2)=E c=T 

f=I /=W 

2/=A ic=G 

//=.H 

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 i=T and, vice versa, f=I. We can 
soon ascertain wlu'ther this principle applies throughout. 
The result of a trial, as far as wo can go, confirms this 
hypothesis, and we quickly arrive at some gratifying 



EXA.MPLES 0¥ DECIPHERING 75 

results. Our attention is first diroctud to tlio j^^roup 
(lines 20-21) S?CHA?, which we identify as " such as." 
Isolated groups begin to join up, as, for instance, 
E?T?YWEIGHTHE?B?EC?, which can scarcely be any- 
thing else but " ently weigh the object," ENTLY being 
part of an adverb yet to be discovered. Always sub- 
stituting the new equivalents as we establish them, we 
continue to build up words and phrases. From line 20 
we can now read SUCH AS ?E ??Y REAS?NA??Y 
E??ECT E??? 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 
paper), we still do not know the key -word by which the 
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 wliere we 
had established only one equivalent in each column — viz.: 

Col. 1, 'y=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. 

Turning to Vigenere's ciphering table on p. 155, 
we first look along the top line of capitals for the letter H, 



76 CEYPTOGEAPHY 

from which we proceed directly downwards in the column 
immediately below until we arrive at the letter v; and on 
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 
with T and c, which gives us J as the third letter. Accord- 
ing to this, then, the key-word is OLJ. 

Thus armed, and with Vigenere's table before us, we 
refer to the text of the cryptogram, and proceed as de- 
scribed on p. 28. We first write the key-word repeatedly 
under the text, thus: 

i p b V d d z o b g q w, etc. 
OLJOLJOLJOLJ 

Starting from the capital in the column to the left of 
the table, we follow the horizontal line w4iich it commands 
and stop at the letter i, the first letter in the ciphered 
text. From this i we ascend the column containing it 
until we reach the top line of capitals, wdiere we find the 
letter U. This should be the first letter of the plain 
text. We continue in like manner with the second letter 
of the cryptogram and of the key-word, p and L, which 
produces E, and so on. We thus decipher as far as the 
following: u e s h s u 1 d s s f n. 

But here we stop, for this array of letters makes no sense 
at all. We are evidently on the wrong track. What is 
the next thing to be done ? Fortunately, Vigenere's 
table is not the only ciphering instrument known to 
cryptographers. Possibly the table used was that of 
Porta, which will Ix; found on p. 158, 

To use Porta's table, wo take oiu- first pair of equiva- 
lents — i.e., v=}i — and we look in the top line for which- 



EXAMPLES OF DECIPHERING 77 

ever of the two letters belongs to the first half of the 
alphabet — in this case h ; we then descend until we 
encounter in the same column the second letter of the 
pair, V. At the left of the double line containing the con- 
junction of the two letters will be found two capital 
letters, Y and Z. Either of these — it is immaterial which 
— will be the first letter of the key-word. We proceed 
similarly with the second pair, e and p, which yields as 
the second letter of the key-word E or F, while the third 
pair, c and t, gives us S or T as the last letter. 

We will say, therefore, that the key-word is YES. 
As before, we write it repeatedly under the text of the 
cryptogram, and, following the instructions accompany- 
ing Porta's table, proceed as follows: 

i p b V d d z o b g q w" w n z s c c z a f, etc. 
YESYESYESYESYESYESYES, etc. 
weshouldsuffi cientlyw, etc. 

In this way the complete text is deciphered easily: 
" We should sufficiently weigh the objects of our hope, 
whether they be such as we may reasonably expect from 
them what we propose in their fruition." 

Our readers will doubtless recognise this as one of 
Addison's obiter dicta. 

A good cryptographer would have detected at once that 
Porta's table was the more likely to have furnished the 
key-word, for the three initial pairs of equivalents which 
gave the clue to the cipher consisted of letters belonging 
to different halves of the alphabet, and Porta's table is 
so 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. 

In order to decipher quickly by means of a table it is 



78 CEYPTOGKAPHY 

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 first letter of the key-word — as, for instance, 
Y in YES — follow^ed by those under the second letter, E, 
and finally those under the last, S. In the case of A^ige- 
nere's table a set square is a useful aid. 

A BiLLET-DoUX. 

A gentleman called upon me and complained that the 
behaviour of his son was not giving him entire satisfac- 
tion. 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) : 

"JDBR DE 3L DJL RbZl 

1 2 3 4 5 6 7 8 91011 12, 1314 

nBR3R J : 00 LQEI 

15 16 17 18 19 20 21 22 23 24 25 

nRZlDHJL bJJL 

26 27 28 29 30 31 32 33 34 35 

niiiUZlRIlL" HidHCJ 

36 37 38 39 40 41 42 43 44 45 46 47 

LLi zjLL^zi njnuH 

48 49 50 51 52 53 54 55 56 57 58 



EXAMPLES OF DECIPHEEING 79 

" Very good," I said to the anxious father; " will you 
kindly call to-morrow about two o'clock ?" 

Loft to myself, I began to study the cryptogram. 

The signs 1-42 are between quotation marks. The 
most frequently occurring sign is No. 7, which is repeated 
nine times in all — about the normal frequency of the 
letter E 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 
information — the text cannot be French, a language which 
does not contain words beginning with doubled letters. 

Examples of such words occur in English — eel ; in 
German, Aal (eel), Aar (eagle), Aas (carcase). Leaving 
aside Gaelic,^ a language not very extensively used, the 
two principal languages which contain a considerable 
number of words of this sort are Kussian and Spanish. 

In Eussian, a whole series of words begin with vv, the 
commonest being vvcdienie (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, sign 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 
odyio (one) or okno (window). But n is one of the most 
frequent letters in Eussian, whereas sign 52, supposed 
to represent it, occurs only twice in the whole text. 
Furthermore, the word formed by signs 7 and 8, which, 
according to our supposition, should be od or ok, is mean- 
ingless in Eussian. We must, therefore, abandon that 
language. 

1 Note by Translator. — Mention might be made of Dutch, 
with oog (eye), een (a, one), uur (hour), and other similar words. 



80 CEYPTOGEAPHY . 

Let us now pass on to Spanish. Here the only letter 
which can be represented by the double initial 43 and 44 
is I, and, in fact, II forms the beginning of a large number 
of very common 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 letters a and e, we observe that signs 
7 and 41, representing e, 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 
este (this). In our text we count nine e's, seven a's, and 
seven s's. According to the rules of Spanish crypto- 
graphy, occurs as frequently in that language as s, if 
not more frequently. 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 
0. We should theh have for the word 23-25 SO?, doubt- 
less son (are). With n tracked down, we identify 5-6 
and 21-22 as no (not). 

. In Spanish, the commonest group of three letters by 
far is que; the word 12-14 ends with e, and, its first two 
letters being so far unknown to us, might well be que 
(that, than). This seems probable, for then 48-49 will be 
tu (thou, you). If 9-11 is mas (more), 36-42 =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 
of a square, with this differt'nc(>, that the square m is 
blank, n contains one dot, and o two dots. Comparing 
these with the][other angles and open squares, with and 



EXAMPLES OF DECIPHERING 81 

without (lots, 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 : 



A 


J 


S 


D 


M 


V 


G 


P 


Z 



B 


K 


T 


E 


N 


X 


H 


Q 


, 



C 


L 


U 


F 





Y 


1 


R 


,, 



The deciphered text proves to be as follows : 

" Amor no es mas que porfia: 
No son piedras las mujeres." 
Lleva tu este papel. 

(" Love is nothing more than a squabble: 
< Women are not stones." 

Keep this jiaper.) 

When the fond parent returned, he was not particu- 
larly pleased to learn that he had had his son taught 
Spanish only to find him receiving such sentimental 
lessons as this on satin and scneted paper. I had a 
presentiment that his offspring w'as going to have a very 
uncomfortable interview. I retain one pleasant memory 
of my visitor: to his thanks he added a phrase which I 
hear none too often: " Don't forget to send me a note of 
your charges." 

A Little German. 

In the German original of the Diary of a S el J -Observer, 
by the celebrated physiognomist Jean-Gaspard Lavater, 
of Zurich, dated January 17. 1773, wo find this sextet in 
cipher : 



82 CEYPTOGEAPHY 

" Efs ekftfo Tubwc efs Fsef hkfcu, 
Fs xbs hftffhofs woe hfmkfcii. 
Fs ibssf efs Wotufscmkdilfku. 
Ko tfkofs Obdiu tkdi pgu hfgsfwii ! 
Ft gsfwf tkdi, xfs ekftft mkftu, 
Ebtt Fs, hmfkdi kin, wotufscmkdi ktu." 

We begin by calculating the frequency of the letters 
in the text. 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., 
or an average of 1 in 5| letters. As this verse contains 
156 letters, we ought to have here proportionately 
18+10=28 E's. The proportion of E's, or letters 
supposed to be such, is therefore somewhat higher than 
the normal average. 

According to an Austrian authority, Colonel Fleissner 
von Wostrowitz, the letters following E in order of fre- 
quency in German are: N I E S T. We will suppose, 
therefore, that s, 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 
these should signify I, the other E. Then comes u 
(eleven times), probably equalling S. For the letter T 
we have the choice between c, i, and 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 frccpient terminal letter. Now, out 
of the thirty-three words comprised in the verse, ten do, 
in fact, end with our supposed N. Indeed, nine out of the 
ten end with I^N, which is also in conformity with the 
rule. 

Now that wo have at om" disposal two practically 



EXAMPLES OF DECIPHERING 83 

certain letters, let ns substitute the pliiiii letters for the 
ciphered ones standing for them. We shall Ihen have 
33 E's+n N's, making 50 known letters out of a total 
of 156. The undecipliered letters are replaced by dots: 

".en ..e.e en En.e ..e.., 

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 not much 
harm is done, since we have only just started. Where 
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 
word in German beginning with e can only be eh, ei, er, 
or es, apart from such imported expressions as en hloc, 
en gros, and en-tete. 

Can the word in question, then, be Ei (egg) ? No, for 
it occurs three times in the sextet, and " egg " is not a 
term likely to be repeated so often in the poetic style. 
True, if this were the case, the fifth word in the first line 
might be Eile (haste), but the first word in the line, 
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 Beh (roebuck), we 
should have " roebuck " occurring twice in tlu^ same line, 
which is incredible; besides, the second reli is not written 
with a capital letter, and cannot, therefore, be a nt)un in 
German. Furthermore, the text would contain a dis- 
proportionate number of words ending witli eh. Neither 
can the word be Es, for while the fifth word would then 



84 CEYPTOGEAPHY 

be Espe (asp), the first Avord would be Pes, which is also 
non-existent in the German language. 

Only Er is left, and we find this meets the case well. 
The fifth word in the first line now becomes Er . e, which 
can be no other than Erde (earth). The first word will be 
Der (the, who). Assuming, therefore, that s=E and 
e=D, the first line reads: 

"Der d. e. e der Erde .. e..," 

Further trial favours the idea that the second word must 
be diesen (this), the value i=S being arrived at from the 
first word in the fifth line: Ff=probably ES. That 
k=l in the w'ord diesen is confirmed by the second word 
in the fom'th line, which, with the letters so far ascer- 
tained, gives us tfkofs=seiner (his), and by the preceding 
word, Ko=In (in). 

From the letters already deciphered we make the 
following discovery: /=E — that is, the ciphered letter 
stands for the letter preceding it in the alphabet; s=E, 
e=D, the same remark applying in each case. Perhaps 
it will be the same for the whole of the alphabet. We 
accordingly make the trial, checking the result of the' 
equations from time to time : 

a=Z b=A c-B d=C e=D f=E 

g=F h=--G i=H k=I 1=K m=L 

n=M o=N p=0 q=P r=Q s=E 

t=S u=T w=U x=W y=X z=Y 

It will be noted that the letters a, q, r, y, and z of the 
secret alphabet, corresponding to the plain letters Z, P, 
Q, X, and Y, are absent from the sextet, which wc now 
read as follows: 



EXAMPLES OE DECIPHERING 85 

" Der diesen Staub der Erde giebt, 
Er war gesccgner und geliebt. 
Er harre der Unsterblichkcit. 
In seiner Nacht sich oft gefrcut. 
Es freue sich, wer dieses liost, 
Dass Er. gloicli ilim, imsterblich ist." 

In English: 

" He who gives this dust to earth, 
Was blessed and beloved. 
He waits for immortality. 
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 
usual in German texts to find a proportion of pure French 
words varying from 5 to 8 per cent., or more. 

A Short Cut. 

I have just received a picture postcard from a young 
friend who signs himself " M. J." It depicts a pretty 
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: 

n>Aj + vrLv<no + xrn = Anooj 

12 84 50789 10 11 12 13 14 15 16 17 18 19 20 21 

+ v-nv I +x + ncoTfn>n<VLO 

22 23 24 25 26 27 2s 29 30 31 32 33 34 35 36 37 38 39 40 41 

L I +voo±Lvr + Aoon + <"inrA 

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



86 



CEYPTOGEAPHY 



We begin by constructing a numerical table of all the 
signs in order of frequency: 



Sign. 


Times. 


Si(jn. 


Times. 


Sign. 


Times 


+ 


9 


00 


4 


J 


2 


n 


8 


< 


3 


X 


2 


V 


7 


n 


8 


T 


1 


r 


4 


• 1 


2 


± 


1 


L 


4 


o 


2 


= 


1 


A 


4 


> 


2 


— 


1 



We assume that -}-=E. According to the laws of 
letter frequency in English, "1? the second sign in the 
table, should be S, T, or A. Signs 30 and 31 would 
then represent either ES, ET, or EA, any one of 
which is possible. There being no other similar juxta- 
positions, we are unable for the moment to establish 
the point. 

We observe that tliree of our supposed E's are followed 
by the sign V? which may, according to order of fre- 
quency, stand for E, D, N, A, S, etc. This sign occurs no 
fewer than seven times in the cryptogram, but as yet we 
are unable to establish from its connections more than 
that it must be a consonant, and that the sign [_ (Nos. 8 
and 48) is i)r()l)ably a vowel. 

Let us now try a trail that has proved very useful in 
deciphering other examples — the discovery of the trigram 
THE, which, of course, cannot be said with certainty to 
be present in the text, but is so frequent a group in P^nglish 
as to make its presence a very reasonable assumption. 
We liiiNc ill the cryptogram nine triplets ending with E, 
of which 28. 21). 30 begin with I*j and 42.43.44 begin with 
a supposed vowel. This leaves us with seven groups, 
two of which start with the sign Vj which we have 



EXAMPLES OF DECIPHERING «7 

already noted as a consonant often following E, and 
three begin with the sign 00. These tliree triplets are 
20.21/22, B2.83.B4, and 53.54.55. Assuming that one 
of tlie tri])lets is THE, we have for the value H to choose 
between the signs J, Xj '^i"^ D- The first, _|, occurs 
twice before E, while the last, Q, occurs once before and 
once after E. The second sign, "J", 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 T, it occurs to us to search for an external 
clue. Turning the card, we observe the name " Pen- 
zance," which suggests a short cut. Our young corre- 
spondent has possibly mentioned the name in his message. 
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 
are represented by the sign V at 23 and 26. The initial 
P, however, proves to be the sign J, which we had assumed 
to be H. This letter must, therefore, be one of the two 
signs □ or T — that is, if the trigram THE occurs in 
the text. 

The results so far established are as follows : 

n-A, I=C, +=E, V-^^ J=P, 00-T, - = Z. 

Having marked the equivalents in the cryptogram as 
far as we have gone, we note that the group following 
" Penzance "—i.e., 29-35— is ?EAT?E?, which it does not 
require much imagination to transform into WEATHER. 
From A?EW (H-H)-" a few" to ?A?N'?F?CENT 
(36-46)=" magniticeiit," we reach om- goal in tliree or 



88 CEYPTOGEAPHY 

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

" Am spending a few days at Penzance. Weather 
magnificent. Kindest regards." 

A Dictionary Code. 

The following cryptogram is handed to me : 

5761 8922 7642 0001 9219 6448 6016 4570 4868 7159 

8686 8576 1878 2799 6018 4212 3940 0644 7262 8686 

7670 4049 3261 4176 6638 4888 4827 0001 8696 6062 

8686 2137 4049 2485 7948 0300 9712 0300 4212 9576 

2475 8576 8337 0702 9185 

In practice, this kind of cipher, which is very commonly 
used, is subject to arbitrary complications, and it may well 
prove quite a long task to restore each number to its 
original integrity, the sender having probably shuffled the 
four figures throughout in accordance wdth a formula 
agreed upon with the recipient. 

But as it is always best to proceed from the simple 
to the complex, we will act on the preliminar}- assumption 
that the above numbers have not been changed, and are 
to be read just as w'e see them. We begin by making 
a list of the forty-five numbers, of which the low-est is 
0001 and the highest 9712, arranging them in numerical 
order : 



0001 


2485 


4212 


6062 


8576 


0001 


2799 


4212 


6448 


8576 


0300 


3261 


4868 


6638 


8686 


0300 


8696 


4570 


7159 


8686 


0644 


8922 


4827 


7262 


8686 


0702 


8940 


4883 


7642 


9185 


1378 


404!) 


5761 


7(;7() 


9219 


2187 


4049 


6016 


7!>4S 


9576 


2475 


4176 


6018 


8837 


9712 



EXAMPLES OE DECIPHERING 89 

We apparently have to deal with a dictionary code, 
numbered from 1 to 10,000. Faced with a system like 
this, so simple and regular, one has to be on the alert 
lest it should conceal a trap. On one occasion, in an 
example which seemed quite as clear, I produced the 
reading: " Either X or Y w-armly recommended." 
But subsequently I ascertained that the numbers had 
been " cooked " in the cipher, and that the true reading 
of the phrase was: " Both X and Y absolutely unknown." 

Assuming in the present case, how^ever, that the 
numbers are unaltered, w^e make the following observa- 
tions: The number 0001 occurs twice, as do 0300, 4049, 
421 2, and 8576, w'hile 8686 appears three times. The 
following pairs occur wdth 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 will now suppose that the number 0001 represents 
the letter A. We 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 
in Webster's Dictionarj^ classified according to their 
initials.-^ 

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. 

^ See p. 138. 



90 CEYPTOGRAPHY 

We have begun by supposing that 0001= A. The 
next thing is to look for certain words which one would 
expect to find in most texts, as, for instance, the prepo- 
sitions " of " and " to," the conjunction " and," the 
article " the," etc. Now, we learn from the table that 
in Webster's Dictionary, theoretically divided into a 
hundred equal sections, words beginning with are 
comprised between the 58 and 61 per cent, marks. If 
the dictionary w^ere 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, om' attention is drawn to 
the words closely following it: " off," " offend," " offen- 
sive." Surely this last — a common military term — is 
the equivalent of our second presumed number, 6018. 
At any rate, the close proximity of the two numbers is a 
promising indication that our surmise is correct. 

It will be useful now to seek such words as " the " and 
"to." The dictionary table shows T's in Webster to 
fall between 8715 and 9298 (substituting the 10,000 divi- 
sion for the percentages). As already noted, the number 
8576 appears twice in the text and 8686 three times. 
These numbers are outside the T limits, and fall in the 
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 consecutivelv — 8686, 8576 — 



EXAMPLES OE DECIPHERING Ul 

which favours the assmii[)ii()ii that the first equals TO 
and the second THE. 

Another number occurring twice is 0300. The dic- 
tionary shows A to extend to 6"43 per cent, of Webster, 
or 643 per 10,000, and as " and " is about half-way through 
the A section, this word may well be the reading of 0300. 

There are two other pairs of duplicate numbers — 4041) 
and 4212. These fall somewhere about H, but there are 
so many likely words with this initial, such as HAVE, 
HAS, HE, HIM, etc., that it is difficult to favour any 
isolated word without the assistance of the context. 

It will bt' as well at this juncture to endeavour to 
construct a part of the text by using the words so far 
obtained as a skeleton. Can we fill in the lacunae in 

the group TO THE OFFENSIVE, for 

instance ? The two missing words are represented by 
the numbers 1378 and 2799. This latter falls among 
the E's. We have, then, E . . . OFFENSIVE— doubt- 
less " enemy offensive.'' It happens that the other 
nundier, 1378, ^vhich falls under C, is almost half-way 
between 0000 (A) and 2799 (ENEMY), and the only 
likely word in the dictionary occurring in this position 
is " coming." We ma\', therefore, not be far w-rong 
in reading this group: TO THE COMING ENEMY 
OFFENSIVE. 

Another group that attracts our attention is A XI) .... 
AND. This is followed by the number 4212, which occm's 
again after the word " offensive.'' Numerically, the 
number is nearly half-way to 8576, to which we have 
attached the reading THE. Allowing for a small margin, 
as we did ill the case of the T's, No. 4212 should coincide 
with the beginning of I's rather than the H's. Tenta- 
tively adopting the pronoun " I " for this number, we 



92 CRYPTOGEAPHY 

next note that the number occurring between the two 
AND's is 9712, the highest in the cryptogram. As this 
is near the end of the alphabet, the pronoun " you " 
seems to be indicated, and we have: AND YOU AND I. 

The number following " i " in the above group is 
9576, the second highest, and therefore probably a 
W word, perhaps WERE or WILL. It is followed by 
2475, an undoubted D word, and the next is THE. 
What can this D word be ? Alphabetically it occurs 
somewhere between " coming " (137H) and " enemy " 
(2799). The interval between these two is 1421, and the 
difference between 1378 and 2475 is 1097, or roughly 
three-fourths of the interval. This brings us among the 
DI's or DO's. There is 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 
be DO. Suppose we give this reading to the second 
fro tern., and look for a word closely preceding it which 
will suit our context. The dictionary shows us " di- 
vulge " and " divide." The group we are studying may 
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. 
We shall, of course, make a false step occasionally, but 
every word established strengthens our foothold, and the 
context guides us more and more surely as we till in the 
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 sug- 
gestion in th(^ context, coupled with the a]»proximate 
dictionary positions of tlui nunil)ers, cl'tVclivcly narrowing 
our choice. After going on to establish some G and H 



EXAMPLES OF DECIPHERING . 93 

words, such as HIM, HAVE, and GOOD, only a very 
slight imagination is required to convert such a group 
as A GOOD . . . TO into " a good opportunity to," 
and eventually we produce the complete reading of the 
cryptogram as follows: 

"Mi . . . has secured a valuable piece of information 
in regard to the coming enemy offensive. I have been 
requested to send him five hundred pounds. It is a good 
opportunity to denounce him. Do so, and you and I will 
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, 
by a minute investigation, identify so much of the name 
as to reveal the nationality of its owner, but it does not 
matter much to us. The person who gave mo 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 P»,ules. 

A copy of the Berliner Tagehlatt 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 harbom'ing enemy agents. 

Opening the journal, I observe an article witli big 
headlines announcing an enemy victory. The article is 
heavily marked with red crayon. Concluding that this 



94 CKYPTOGKAPHY 

is a mere blind, I scrutinize every page, column after 
column, until, on the last page, among the Stock Exchange 
quotations, my attention is attracted by a certain number 
of figures marked with dots in ink. 

Taking a sheet of plain paper, I make a careful copy of 
all the figures marked in order as follows : 

1856 2 9593769367418742 2 742555 
37 5 42 86943673562 2 166 2 6708567 
375 8 39 6 93 2 44 3 2 682979 3 6634163 
3174 2 559 2 8 683 2 77811966 2 363 2 8 
76296532763160256136 8 0227617 
22762272427425 63 3 16135591858 
425642671879 2 3693872376236 6 3 
2 4681866 5 32473 2 67639612166 

Altogether there are 222 figures. Have we here a 
dictionary code ? No, because 222 cannot be divided 
by 5 or 4. It is, however, divisible by 6 or 3. With 
six figures a dictionarj'^ of a million words (including 
000,000) can be constructed, Init this would be too many. 
With three figures one might compose a dictionary of a 
thousand words (including 000), but this is obviously 
too few for practical purposes. 

A dictionary code being apparently out of the ques- 
tion, we 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 
into groups of three, which we arrange in order from the 
lowest to the highest. 

Of the seventy-four groups thus obtained, we note 
that six are duplicated — viz., 16(), 267, 276, 425, 532, and 
742. 

If we admit that each of those seventy-four groups of 
three figures represents a lett(>r, we shall rofpiire a pro- 



EXAMPLES OF DECIPHERING 05 

portion of at least niiio ll^'s, and the total nuiiil)or of 
repeated groups does not reach that. Tlien how are 
we to got over the great difficulty of identifying the 
alphabetical value of the sixty-two groups not re- 
peated ? 

Let us put aside for the nioniont our notes on the 
three-figure groups, after adding thereto the observation 
that the list shows a certain number of groups which 
differ from each other only by single units, to wit : 135-136, 
255-'25t), 267-268, 296-297, 316-317, 366-367, 591-592-593, 
622-623, 762-763, 868-869. 

While being almost certain that this* will not be of 
much use to us, we will hold it in reserve as a possible 
forlorn hope. It is just possible, too, that these three- 
figure groups may stand for syllables, but even so the 
repetitions should still be more frequent. 

It then occurs to us to add the figures of each group to 
see whether the totals will correspond to the numerical 
rank of the letters in the alphabet. Putting the larger 
numbers to the test, we get the equivalents 938 (9+3+8) 
=20, corresponding to T; 879=24=X, etc. 

So far, so good. On trying the small numbers, how- 
ever, we meet with a check, there being no A, B, C, D, 
or even E; in fact, no number produces a lower total 
than 7 (160) or 8 (035). Now, a text of such a length 
without a single E will scarcely be found in any language 
of Western Europe. 

It is evident that we must pursue our researches in a 
different direction. TIk^ number 222 is divisible by 2. 
We will, therefore, divide our cryptogram into sections 
of two figures, classifying them in numerical order. 
This enables us to produce the following table of fre- 
quencies : 



96 



CEYPTOGEAPHY 



17 once 

18 5 times 

19 once 

21 twice 

22 4 times 

23 twice 

24 „ 

25 4 times 

26 twice 

28 3 times 

29 3 „ 
31 3 ,, 



32 4 times 

35 3 ., 

36 4 „ 

37 4 „ 

38 once 

39 twice 

41 once 

42 3 times 

43 once 

44 ., 

53 ., 

54 ,, 



55 once 

56 twice 

58 ., 

59 3 times 

60 once 

61 3 times 

62 twice 

63 5 times 

65 once 

66 4 times 

67 3 ,, 

68 3 ., 



69 4 times 

70 once 

72 twice 

73 once 

74 5 times 

76 5 „ 

77 once 

79 twice 

80 once. 

81 „ 



Out of the 111 groups of two figures obtained, we ought 
to find one repeated at least a dozen times to represent E. 
But no group occurs more than five times. There are 
four with this proportion — ^viz., 18, 63, 74, and 76. Can 
they all mean E ? This would make a total of twenty E's, 
a proportion which, while not beyond the bounds of reason, 
is yet somewhat too high for the length of the text. 

Have we to deal with a cipher of several alphabets, as 
the total of forty-six different numbers might lead us to 
suppose ? But how many alphabets ? We must en- 
deavour to find out. While doing so we note that the 
consecutive groups 74.25 are repeated at two other places, 
that the consecutive groups 21.66, 36.63, 63.31, and 
22.76 each occur twice, and that the groups 67.35 and 
32.68 are repeated, but in inverse order. 

If we calculate the intervals between the repeated 

pairs, as described on p. 70, we find thirty-two numbers 

between the first and second pairs, 74.25, and — a curious 

coincidence — a similar number between the second and 

third pairs, 74.25; thirty-six numbers between 63.31 and 

63.31; 87 between 21.66 and 21.66; 58 between 36.63 

and 36.63; and 3 between 22.76 and 22.76.^ 

^ The interval is calculated, as already explained, from a stroke 
dividing the first pair to a corresponding stroke dividing the second 
pair. 



EXAMPLES OF DECIPHERING 97 

Out of tlio 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, there- 
lore, represent as manj^ different letters. Furthermore, 
no number has a higher frequency than three in any of the 
colunms, and, with the exception of 59, which is the only 
nund)er occurring three times in col. 1, and might, 
therefore, stand for E, the frequencies of 3, 2, and 1 are 
too evenly dispersed to furnish any clue as to their 
significance. 

The solid features to w^hich we must revert are the 
repetitions of the groups 74.25, 21.66, 36.63, 63.31, 
and 22.76. 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 
reveals a peculiarity which may help to put us on the 
right track. It will be observed that there is no number 
lower than 17 and none higher than 81. The cryptogram 
may, therefore, 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 made to 
coincide with different figures on the numerical slips, and 

7 



98 CEYPTOGKAPHY 

by this means a great variety of secret alphabets repre- 
sented by numbers can be formed. 

Where the respective positions of the strips as adopted 
at the beginning remain michanged to the end of the 
cryptogram, the system is that of numerical fixed slides. 
When 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 parchment are very fragile 
and easily torn, we will replace the whole by small rules 
of plain wood, two long and two short. We graduate 
all the rules by 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 of the short rules we inscribe the alphabet 
in the usual order, and on the other the alphabet 
in reversed order: Z, Y, X, etc. The diagram will 
better illustrate the part which the four rules can play 
in ciphering. 



(hk 


cIdIeIf 


-^ 


1 1 J Ik 


lImIn 


o|p|q|r|s|t|u 


v|w|x|y|2|) 




(hl2 


sUlsle 


7 8 


9ll0|l l|l2|l3|l4 


1 5|l 6|l 7|l 8|l 9|20|2 l|22|23|24|25|26t27|28i2| 




(l5l|52|53|54|55|56|57|58|59|60|6l|62|63|64|65|66|67|68^9|70|7li72|73|74|75|76|77|78|79|8^ 




(l^lY 


xiw|v|u 


T|S 


r|q|p 


o|n|m 


l|k| J 1 1 |h|g1f 


EiD|c|B|AD 



As an example, to cipher the word " deed " from the 
upper alphabet, we can employ at will either of the two 
numbers falling under each letter, transcribing the word 
04.56.06.55, for instance. Using the lower alphabet, 
we have the choice between the two lines of figures above 
it, and may produce 73.'22.72.2;-5. The two different 
letters in the word " deed " may, therefore, be represented 
by eight different numbers. 



EXAMPLES OF DECIPHERING 99 

As regards tho system of movable rules, we give 
farther on, in the chapter entitled " Spilt Ink," 
a proximate instance, coded by letters instead of 
numbers. 

In what way must we adjust our wooden rules or slides 
in order to decipher the cryptogram which wo are now 
studying ? We have already noted that the double 
group 74.25 occurs three times in tho cryptogram. 
Let us now endeavour to ascertain whether it corresponds 
to the frequently occurring bigram TH. For this pur- 
pose we adjust the upper alphabet in such a way that T 
is above 74 in the lower numerical strip. We then move 
the upper numerical strip until the 25 thereon falls below 
H in the upper alphabet. 

Beginning from the first pair, 74.25, which occurs in the 
first line of tho cryptogram, we decipher AT as the two 
letters following TH. The next number, 54, falls outside 
the range of the alphabet. Ignoring this for the moment, 
we proceed: KOZMllHDLIP. Plainly we are on the 
wrong track. 

Suppose we try another of the repeated groups, 21. GG. 
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 promis- 
ing than W0YR9CTH. 

There are other duplicated pairs open to investigation, 
but the fact that our first essay, though a failure, pro- 
duced initially the combination THAT induces us to 
restore the rules to the position 74.25=TH. This time 
we take the second pair, which occurs in the fourth line 



3 g ^ 

S ^ |00 CKYPTOGEAPHY 

J' ^ §)f the cryptogram. We get as far as THEKN0W9BLFI, 
? <d aand are again baffled. Yet there are the initial letters 

^ that seem so promising. 

^^' Suppose we revert to the combination THAT (pre- 

- ^ sumed) in the first line and try the letters preceding it. 



^ "^ We get ETATSOTELBA. Eeversing this, we recognise 
^ 5 " Able to state." At last we are making definite pro- 
*? Z gress. But we have now reached the beginning of the 
7 text ; and when we attempt to go in the other direction, 

we get a mixture of comprehensible and incompre- 
hensible groups, with occasional numbers which have no 
corresponding letter. Such numbers are 17, 44, 53, 54, 
and 81. 

These perplexing numbers must be either pvmctuation 
marks, blank letters, or — as we are beginning to suspect — 
" changes of alphabet." In two cases, certainly, such 
numbers separate intelligible from non-intelligible groups. 
Perhaps Nos. 54 and 81, occurring, as they do, on the 
lower numerical rule, are intended to indicate that the 
groups following are to be read from the loiver alphabetic 
rule, in which case 17 and 44 will refer to a change to 
the upper alphabet. 

On putting this theory to the test, working from 
No. 54, we are agreeably surprised to encounter the 
group PLANISWOEKING. By continuing to follow 
the indications given by the key numbers, we are 
very soon in possession of the plain text complete, as 
follows : 

" Able to state that plan is working well. Only six 
in the know. Your people must have everything ready 
by May fourth. Signal three two. 

What the further history of this interesting plot was 
I am unal)le to state. We may at least suppose that 



EXAMPLES Oi^^ DECIPHEKING 101 

tlu! interception and disclosure uf the message went far 
to bring it to an untimely end. 

A Contribution to History. 

The post brings mo a letter; I recognise in tlie address 
thc! 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 
of paper containing a cryptogram in the same writing. 
Who would have expected a communication in cipher 
from such a man ? Decidedly, everybody is taking 
up cryptography now^adays. Let us see what he has 
to say: 



o 


u 


s 


z 


e 


h 


n 


s 





b 





n 


1 


h 


h 


i 


c 


m 


a 


c 


P 


k 


s 


c 





u 


s 


e 


e 


V 


r 


i 





X 


g 


e 


g 


u 


u 


e 


u 


u 


u 


h 


s 


s 


d 


u 


y 


u 





u 


c 


n 


s 


a 


c 


P 


i 


a 


m 


e 


g 


u 


V 


b 


a 


i 


k 


a 


s 


s 


d 


f 


a 


P 





a 


r 


i 


j 


a 




e 


V 


a 


f 


n 


u 


s 


r 


t 


r 


r 


c 


c 


m 


c 


J 


c 


a 


f 


w 


s 


i 


u 


t 


u 


i 


i 


k 


i 


i 


u 


u 








u 


i 


n 


1 


i 


k 


a 


d 


n 




o 


h 


h 


a 


cr 





V 


j 


i 



The cryptogram contains, in all, 13G letters. A scrutiny 
of the text shows the following repetitions: OU (4 times), 
US (3), NS, NL, CM, AC, CP, EV, EI, GE, EG, GU, 
UU (4), SS, SD, UO, BA, IK (3), KA, AG, AF, lU, UI, 
II, GO. 

• We divide the two letters of the first OU with a stroke, 
do the same with the second, third, and fourth, and then 
count the letters between the strokes. Proceeding like- 
wise with the other repeated pairs, we establish the follow- 
ing table (the figures represent the number of letters in 
the interval from one pair to its repetition): 



102 



CRYPTOGEAPHY 



ou-ou 24 or 2x2x2x3 



US-US 

ns-ns 

nl-nl 

cm-cm 

ac-ac 

cp-cp 

ev-ev 

ri-ri 

ge-ge 

eg-eg 

gu-gu 

uu-uu 



26 
66 
24 
63 
47 
108 
79 
37 
37 
55 
48 
48 
26 
26 
3 



2x13 
2x3x11 

2x2x2x3 

3x3x7 

47 

2x2x3x3x3 

79 

37 

37 

5x11 

2x2x2x2x3 

2x2x2x2x3 

2x13 

2x13 

3 



uu-uu 1 or 1 

72 ,,2x2x2x3x3 

ss-ss 26 ,, 2x 13 

sd-sd 26 ,, 2x13 

uo-uo 64,, 2x2x2x2x2x2 

ba-ba 64 ,, 2x2x2x2x2x2 

ik-ik 42 ,, 2x3x7 

12 ,, 2x2x3 

ka-ka 54 ,, 2x3x3x3 

ag-ag 49 ,, 7x 7 

af-af 15 ,, 3x5 
iu-iu 8 ,, 2x2x2 

ui-ui 10 ,, 2x5 
ii-ii 3 ,, 3 

go-go 5 ,, 5 



It will be noted that the factor 2 is common to 19 out 
of the above 31 intervals, and if there were no other 
important factor we should bo tempted to assume a 
key- word of two letters; but the factor 3 is common to 
14 of the intervals, which is nearly a half, and in any 
case there are bound to be a considerable proportion of 
2's, because that factor must appear in every even 
number. The probabiHty that a key-word of three 
letters has been used is strengthened by the fact that 
there are some repeated trigrams in the cryptogram, two 
of which, OUS and IKA, have intervals divisible by 
three. 

Assuming, therefore, a key-word of three letters, we 
copy the whole of the text into three columns — that is, 
in sections of three letters, each forming a line, and number 
the lines for convenicint reference. A sufficient margin 
should ])G left to attach the transcription, as described 
in the chapter on " Ciphering by Means of a Key- 
Word." 



EXAMPLES OE DECIPHEliING 103 



(1)0 


u 


s 


(17) y 


u 


o 


(33) m 


c 


i 


(2) z 





h 


(18) u 


c 


11 


(34) c 


a 


£ 


(8) n 


s 


o 


(19) s 


a 


c 


(35) w 


s 


i 


(4) b 





n 


(20) p 


i 


a 


(36) 11 


t 


u 


(5)1 


h 


h 


(21) 111 





g 


(37) i 


i 


k 


(G) i 


c 


m 


(22) u 


V 


b 


(38) i 


i 


u 


(7) a 


c 


P 


(23) a 


i 


k 


(39) u 








(8) k 


s 


c 


(24) a 


s 


s 


(40) u 


i 


11 


(9) o 


u 


s 


(25) (I 


f 


a 


(41) 1 


i 


k 


(10) 


c 


V 


(26) p 


() 


a 


(42) a 


d 


n 


(11) r 


i 


o 


(27) r 


i 


J 


(43) g 


o 


h 


(12) X 


g 


e 


(28) a 


g 





(44) b 


a 


g 


(13) g 


u 


u 


(29) V 


a 


f 


(45) 


V 


J 


(14) e 


u 


u 


(30) n 


u 


s 


(46) a 






(15) u 


h 


s 


(31) r 


t 


r 








(16) s 


d 


u 


f32) r 


c 


c 









Our text is thus arranged in three columns, the first 
beginning o z n, the second u e s, and the third s h o. 
Col. 1 is presumed to have been ciphered by the first 
letter of the key-word, which remains to be discovered, 
and cols. 2 and 3 by the second and third letters 
of the same key-word. 

The best way to find the key-word is to ascertain, if 
possible, which letter represents E in each cclumn, or, 
failing that, to establish at least one letter in each column. 
Now, although E is the most frequently occurring letter 
in English, it is followed so closely by T and A that 
allowance has to be made for one or other of these pre- 
dominating in a short text. In looking for E, it should 
be borne in mind that this letter very commonly follows 
H, also that TH is a very frequent bigram and that THE 
is the commonest trigram. 

Now it happens that the first three letters, o u s, are 
repeated in line 9. The word " the " is not an unlikely 
beginning, and the fact that s is one of the two letters 



104 CEYPTOGKAPHY 

having the highest frequency in col. 3 favours the 
supposition that it stands for E. In Unes 1, 9, and 30 
the letter follows u, whence we may draw the legitimate 
Inference, subject to correction, that us=WEi and ous= 
THE. 

Armed with these three letters, we will now consult 
Vigenere's table on p. 155, and endeavour to recon- 
struct the key -word. From the capital letter T in the top 
line of the table we descend the column which it heads 
till we reach 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 
letters in the body of the table are the ciphered letters, 
and the column of capitals to the left are intended to form 
the key-word. This relationship must always be borne 
in mind when ciphering or deciphering from Vigenere's 
table. 

Proceeding in the same way with H and u, the second 
letters in the supposed plain text and the ciphered word 
respectively, we obtain N as the second letter of the key- 
word, and, continuing, from E and s we obtain 0. Accord- 
ing to this, then, our key-word is VNO. 

We must next write out a portion of the text of the 
cryptogram, and, underneath, the key-word repeated 
continuously. By means of the table we proceed to the 
decipherment, with the following result : 

o u s z e h n s o b o n 1 h h i c m a c p k s c 
VNOVNOVNOYNOVNOVNOVNOVNO 
tlieertsfagbzqut 

Plainly, it is useless to go any farther; wo have struck 
a false trail, and must patiently go over our ground anew. 



EXAMPLES UE DECIPHERING 105 

We cannot resort to Porta's tabic i'or eiiliglitcnniunt, 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 ows=TIIE. There is another 
duplicated trigram in the cryptogram which might repre- 
sent THE— i.e., ilxa. This occurs in Knes 23-24 and 41-42, 
i being in the middle column, k in the third, and a in the 
tirst column on the following line. This group, ika= 
THE, does not prejudice, nor is it prejudiced by, ous= 
THE, the difference being due to the fact that the letters 
did not fall to be ciphered under the same alphabet. 
There could, of com'se, be still another form for THE if 
three alphabets were used. 

On putting ika=ll'EE to the test by means of Vige- 
nere's table, we produce the key-word PDW, but this 
merely proves another failure. There can be no doubt 
that tlu-ee alphabets were used, and, as we are unable to 
get any assistance from a key-word, the obvious conclu- 
sion is that we are faced with a cryptogram ciphered by 
means of three irregular alphabets. This makes our task 
rather more complicated, and we shall have to discover 
the moaning of the letters one by one. 

We make a beginning by assuming that ous in line 1 
in the columnar table represents THE, and, in addition 
to marking this word in the margin, w^e mark T opposite 
every o in col. 1, H opposite every u in col. 2, and 
E opposite every s in col. 3. We must always proceed 
in this way, going through the colunms, and marking the 
appropriate transcription tlu'oughout, every time we 
establish the value of a letter. By this means we obtain 
our clues and build up the solid fabric of the plain 
text. 

In the present instance this marking, besides bringing 



106 CKYPTOGEAPHY 

out the word THE repeated in line 9, shows H to occur 
in two consecutive Hnes, 13-14, followed in each case by 
the ciphered letter u. This cannot be meant for E, since 
we have already given this value to s; neither can it be 
T, which would produce the combination HT?HT. The 
choice is limited to A and I, either of which is far more 
likely than 0. If 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 
the same column. 

On the other hand, if we assume the letter following 
H to be I, we have the possible group (W)HI(C)H I(S), 
and as neither of the parenthetical letters usurps the 
.position of T, we will boldly adopt this reading, which 
gives us the equivalents: (/=W, e=C, u—S (all col. 1), 
w=I (col. 3). 

We have already noted another trigram in the crypto- 
gram wdiich appears likely to represent THE — i.e., ika. 
It occurs isolated in lines 41-42, but enables us to produce 
THE(R)E in hnes 23-24 and TH(A)T IS in hnes 37-39. 

The word " of " would naturally be expected in a text 
the length of our cryptogram, perhaps more than once, 
and probably preceding " the." The group THE in 
lines 41-42 is preceded by nl, and as this bigram occurs 
twice in the text, we may not be far out in ascribing to 
it the value OF. It occurs isolated in lines 4-5, but in 
lines 37-42 it gives us the very substantial result: THAT 
IS (EA)ST OE THE. The parenthetical letters cannot 
be WE, because E (col. 3) has been estaljlished as the 
equivalent of .s, whereas the ciphered letter here is o. 

Let us pause here a moment to sunnnarise our dis- 
coveries : 



EXAMPLES OF DECIPHEKING 107 

Col 1. Col. 2. Col. 8. 

o=T, 3 letters u=H, G letters s=E, 5 letters 

1=E, 2 „ s=R, 4 „ o=A, 4 „ 

i=A, 4 , „ o=E, 4 „ n=0, 4 „ 

a=E, 5 * „ i=T, 8 „ u=I, 5 „ 

e=C, 2 „ k=H, 3 „ 

g=W, 2 „ 

u=S, G „ 

Total: 67 letters out of 13G, which indicates good pro- 
gress. It is as well to summarise results in this way 
from time to time, as it shows how far the realm of hypo- 
thesis 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 
that the repeated ciphered letter a occurs in col. 3. 
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, 
C, 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 sand- 
wiched between two T's, and not yet accounted for, is I. 
This enables us to submit the group to the following trans- 
formation: THE TIT(L)E (OF), the parenthetical letters 
requu-ing 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 9 to 



108 CKYPTOGEAPHY 

line 17 now appears as: THE CevITAx OF WHICH IS 
liEsdly. The second word can be no other than CAPI- 
TAL, while as to the last, the name of a capital having 
six letters, of which the second is E and the fifth I, there 
need not be much hesitation in pronouncing it BEKLIN. 
The cryptogram holds no further terrors for us. 
BEELIN makes us think of (P)E(U)S(S)IA (lines 35-37), 
and eventually we have this table of all three alphabets: 

Plain text: ABCDEFGHIKLMN O PRSTUVWY 

Cipher Col. l:ide-alnprvxzymwsuokb gc 

,, ,, 2:eh o ua-d-cg-stiv--f 

,, ,, 3:op-mseckurjbfnvhgai--- 

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. — This statement by my correspondent, who is 
not a 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: Hermaini Schulze, Ludwig von Roennc, and Ludwig 
Bornhak, who, with considerable reticence, acknowledge that the 
Margrave of Brandenburg is only, and has never been other than, 
" king in Prussia." The immense kingdom of Prussia, as we Icnow 
it to-day, is only a fiction; its existence has no serious historical or 
juridical basis. 

Spilt Ink. 

This mornhig's post brought me a letter from a well- 
known professional man who has been utilising his spare 
time in constructing a safe ci[)lu'r. He sends me a 
specimen, and warns me that he has submitted it to 
several amatem's, who failed to decipher it, the last re- 



EXAMPLES OF DECIPHERING 109 

turning it intact with tho remark that it was so much 
" spilt ink." " Perhaps you will have hettor luck," adds 
my correspondent, no douht smiling up his sleeve as he 
wrote. 

Let us glance at this cryptogram. If I do not succeed 
in deciphering it, I will frankly admit it without any 
false shame: 

sseguhckxzdgzggzszdsj nfj 

w^ p h f X q u g o g g h z n y s 1 p ays f i m 

owl s w n n o z cl cl f h v m p x k q I3 z h h t 
i z n h k d r y y t x f s n r e x b e m f 

Total: 93 letters. I prepare a list of frequencies, and 
find that the most numerous letters in the text are h, s, 
and z, which each figure eight times. Can any 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), 
/, n (six times), and d, x (five times). These figures are 
too close for a simple alphaljet cipher. Besides, what 
double letter could the initial ss stand for ? If w^e assume 
a name beginning LL, they would have to be followed 
by or E, and as e, the cipher equivalent, appears only 
tlu'ee times in the whole of the cryptogram, it is useless 
to go any farther in this direction. 

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 de- 
scribed 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, 2d. for instance, occurring three 
times. Accordingly, we insert a stroke between the two 



no CEYPTOGEAPHY 

letters in each pair and count the intervals, making the 
f ollo^\^ng tabulation : 

gz-gz 3 letters, or 3 
zd-zd 8 „ „ 2x2x2 

,, ov „ ,, oXl3 
gg-gg 20 „ „ 2x2x5 
ys-ys 5 „ „ 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 tliree equal factors, 
3 is the most likely to indicate a key-word, if any, leads 
to no result. 

The cryptogram contains several double letters — viz., 
ss, gg (twice), nn, dd, lifi, and ijy. A close scrutiny reveals 
the fact that the gg in one case is followed by h in alpha- 
betic sequence, and in like manner nn is followed by o. 
This detail gives me a clue to the right track. The system 
of ciphering used appears to be that known as the " Saint 
Cyr SHdes." 

By means of the Saint Cyr slides we can obtain twenty- 
six different alphabets. Anyone can make these slides. 
All that is necessary is to obtain two rules or strips of 
plain wood, one long and one short. On the short rule 
mark equidistant divisions, and in them inscribe the 
twenty-six letters of the alphabet. Proceed in the same 
way with ilic long rule, but with tlic difference that two 
consecutive alphabets — i.e., fifty-two Ic^ttc^rs, A-Z and A-Z 
— must be marked hero, in order that, when the smaller 
rule is moved up or down in juxtaposition with the 



EXAMPLES OF DECIPHERING 111 

longer, it will always be in contact with twenty-six letters 
on the latter, as shown: 

gThI i|j|k|l|m|n|o|p|q|r|s|t|u|v|w|x|y|zD 



|a|b|c|d|eT7 



|e|f|g|h|i|jRT^ 



ivi|n|o|p|q|r|s|t|u|v|w|x|y|z|a|b|c|d|e|f|g|h|^ 



When the short rule is moved so that its A is above Z 
on the long rule, the Z of the former will coincide with Y 
on the latter. The short rule represents the alphabet of 
the plain text, and the long rule the cipher alphabets. 

In the above example A is represented by G, B by H, 
and so on. If we want to change the cipher, we have 
only to slide the small rule to the right or left, and a new 
ready-made secret alphabet is produced on the long rule 
underneath. In this way we have twenty-six different 
alphabets at our disposal. 

It is even possible to make a change of alphabet with 
every letter ciphered, and that without risking a mental 
breakdown. Suppose, under this scheme, we wish to 
cipher the word " gun." The G on the short rule is seen 
to be over M on the lower rule, so we write M as our first 
letter. We now decide to use a new alphabet based on 
this M; for this purpose we merely slide the short rule to 
the right until its A is above M on the long rule. To 
cipher the second letter of our word, we look for U in the 
upper rule, and below, in the new alphabet, we find G. 
Having written this, we again change the alphabet by 
sliding the upper rule to the position where its A will be 
over G, and now liud the ciphered equivalent of N, our 
last letter, to be T. Thus, the word " gun " becomes 
MGT by means of tliree different secret alphabets, one 
for each letter, obtained automatically in the way de- 
scribed. 

Is this the method by which our cryptogram was 



112 . CEYPTOGEAPHY 

ciphered ? If we knew the secret of the first letter, 
everything coiikl be unfolded mechanically; but we do 
not, and this is the mystery of the " spilt ink." 

However, there are ways and means. When the groups 
ggh and nno attracted my attention just now, I consulted 
a notebook in wdiich I record rules which appear dedu- 
cible 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 A {mm—^A). 

B. When two like letters are followed by the next 
letter in alphabetic sequence, the second and third letters 
in the trigram represent AB {mmn=?AB). 

C. When an a occurs in the ciphered text, the letter 
which follows is identical with the corresponding plain- 
text letter (ae=?E). 

I have noted some further rules on the subject, but 
these three will be ample for our purpose. Let us apply 
them to our cryptogram: 

Eule A. In ss the second s equals A 





» yg >' 


M 


9 




A 




„ nn „ 


>5 


11 




A 




„ dd „ 


>J 


d 




A 




„ Jih „ 


)» 


h 




A 




,, hh ., 


; J 


h 




A 




M yy „ 


j; 


y 




A 


Eule B. 


In (jgh the 


last 


two 


l(>tters equal AB 




„ nno ., 


,, 


JJ 


j> 


„ AB 


Eule 0. 


In (III the 


cilcl 


• V <' 


quals 


Y 



By these ruk^s we would a])])ear to liave accounted for 
twelve letters {(jg occurring twice). Using these as a 



EXAMPLES OF DECIPHERING 113 

check, it now remains to decipher the remaining eighty- 
one 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 plciiiifid choice of starting- 
points. With the Saint Cyr slides to oiu- hand, we select 
for a beginning the first group containing two known 
letters — i.e., ggh. From what has gone before, we infer 
that the ciphered letter h is the first letter of a new alpha- 
bet. 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 
section is reproduced for convenience: 

gghznyslpay 

The letter in question is z, which we find to correspond 
to S. Adding this to the two letters already known, we 
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 
rules, we identify n as the equivalent of 0. Again chang- 
ing the alphabet by giving the value of A to n, and con- 
tinuing similarly with each letter, wo decipher the above 
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 have 
no means of kno^^^ng what the original value of this 
initial s was. However, we have the context to guide us 



114 CKYPTOGEAPHY 

where our formulae are impotent, and effectively the series 
of letters from the second prove to be, AMCONVINCED, 
so that we may safely conclude that the mysterious first 
letter is I. The complete transcription is as follows : 

" I am convinced that the present system is absolutely 
undecipherable. Accordingly, I am proud to have in- 
vented it." 

With the view of softening the disappointment of my 
correspondent, to whom I have communicated the 
deciphered text, I am able to inform him in a covering 
note that this system offers safeguards by no means 
negligible, since, for example, the word "am" appears 
in two different disguises, se and ht; the "in" of the 
words " convinced," " accordingly," and " invented " 
is dissimulated under three separate forms: kx, xk, and/s; 
while " ent " in " preseni" and " invented " is ciphered 
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: " You are fond of solving 
difficult problems. Here you are, then ! I wish you 
joy." The cipher text, which contained fifty-two letters, 
was as follows: 

ylirxqjzmpatcmovzngrqlfkve 
w n o d s d s c k u t i u h p f y w b h e g b x j a 

I began by calculating the letter froquoncios, and, to 
my stupefaction, found two a's, two V^, two c's — in fact, 
two of each letter of the alphabet, neither more iior less. 
Only one group, fZ.s, was repeated, and, that being the 
case, it was useless to seek a key-word. 



EXAMPLES OF DECIPHERING 115 

There could be no question either of a grille or " di- 
viders." In any case, these two letters of every kind 
were a strange coincidence, though instances almost as 
curious are encountered from time to time. I reflected 
on the possibility of a dictionary code. There are con- 
ventional codes composed of three-letter groups: aah 
ivkj, etc. 

By combining the letters of the alphabet in threes, a 
large number of groups can bo obtained, sufficient to 
replace the words 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 
of the alphabet, followed by another letter, similarly 
yields twenty-six combinations. In this way, 26x26= 
676 different groups of two letters can be formed, and 
676x26=17,576 groups of three letters. 

I might have made some investigation in this direction 
but for two obstacles: (1) The fifty-two letters of the text 
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 I 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 — hlieg and hxja. But all 
my efforts proved futile. I could not identify these 
groups approximately with such frequently occurring 
words as " and," " at," " be," " but," and others with 
initials in the early part of the alphabet. 

Not being al)le to obtain the faintest clue as to the 
method of ciphering employed, I called on my friend, 



116 CEYPTOGEAPHY 

informed him of my lack of success, and begged him to 
acquaint me mth the key. 

" Quite simple," he said. " This is how I wTote 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 twenty- 
six. I dropped the whole fifty-two slips into my hat, and, 
after shaking them up, took them out one by one hap- 
hazard, and noted them down just as they came to hand. 
The text thus formed I sent to you." 

" Then it has no meaning !" I exclaimed. 

" Of course," he replied, " and I must ask you to forgive 
my little trick; but you are so clever at cryptography 
that if I had submitted you a text with any meaning at 
all, you would probably have deciphered it far too 
quickly !" 

The Antique Dealer's. 

A friend had asked me to meet him at the tramway 
terminus. I 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 
lost all right to repeat the famous remark, " I almost 
had to wait," attributed gratuitously to Louis XIV., 
who appears to have said quite the opposite.^ 

How should I pass the time ? There was no news- 
vendor. Besides, when one is in the habit of reading the 
news at certain regular hours, it is just the same as with 
meals — one has no appetite between. 

^ " Why scold him 1 Don't you think he is sorry enough to have 
kept me waiting ?" {(Euvres de J. Racine, Hachctte's edition, 1865, 
vol. v.. p. 12.5). 



EXAMPLES OE DECIPHEKING 117 

l>ul 1 descried iui anti(|uc 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 (juickly fell upon the prices 
inscribed on the labels, and, succumbing to the fascina- 
tions of my innocent hobby, I set myself the task of 
deciphering the values of the letters which took the place 
of Arabic numerals. Drawing out my notebook, I took 
note of a number of the articles exposed for sale in the 
four or five windows, as well as their mysterious prices 
— viz. : 

1. Bronze statue (Psyche emerging from bath) z.r.p 

2. Incense burner . - - . i.mp.p 

3. Colom'ed engraving (national costumes) - m.mp.p 

4. Large double mirror, old frame, gilt much 

rubbed ----- mi.mp.p 

5. Inlaid card table - - - . e.p.p 

6. Iron dagger .... mp.p 

7. Small picture (glacier), white wood frame - mr.p 

8. Small engraving (Marie Antoinette), black 

frame ----- nii.z 

9. Small picture (The Elirt), worn gilt frame- mi.z 

10. Black and gold metal -tray, flowers in centre mf.z 

11. Old barometer - - - - m.r.p 

12. Old picture (rustic scene) - - - f.z 

13. Eoiu: old engravings : the four - - m.mf.z 

14. Grandfather clock - - - - mr.p.p 

15. Card table, with inlaid draughtboard - b.r.p 
US. Glass cheese dish - - - . f.^ 

17. Small Elemish painting, copy ("? of a copy) i.mp.p 

18. Large trunk, much patchc(l - - z.r.p 

19. Devotional picture, cloisonne worked on 

wood - . - - . mi.z 

20. Head of lion in bronze (door-knocker) - mi.z 

21. Concave shield, bas-reliefs, in gilt frame - z.r.p 

22. Large oval metal tray, Watteau subject 

in centre - - . . nif.z 



118 CEYPTOGEAPHY 

23. Small mirror, large black wooden frame - b.p.p 

24. Mirror, brown carved frame - - m.mr.p 

25. Tin candlestick - - - - r.p 

26. Oval silver tray, tarnished - - m.mp.p 

27. Bronze bowl on three feet - - - z.p 

28. Large bronze hand lamp - - . - i.mp.p 

29. Small china vase, coloured and gilt - - m.mp.p 

30. Engraving (the Signing of Magna Charta) - r.p 

31. Engraving (Friends till Death) - mp.p 

32. " Library of Famous Men," volume wdth 

49 plates . . . . m.p.p 

33. Silver bell - - - - - m.i.z 

34. Small silk mat, silver fringes - - o.p 

35. Silver strainer - - - - z.p 

36. Old decanter, silver stand - - m.i.z 

37. Locket, wdth cat's-eye and amethysts - mf.z 

38. Chased silver egg-cup - - - mf.z 

39. Copper seal .... m.m.p 

40. Old silver chafing dish - - - o.mp.p 

41. Old decanter .... mi.z 

42. Liqueur stand, with two flagons - - b.b.p 

43. Old beer mug, colomred stoneware - mp.p 

44. Bronze medal set in ring of chased silver - o.p 

The first thing I noticed was. the large number of p's, 
reaching a third of all the letters I had noted. It occurs 
among the ponce and shillings, but never among the 
pounds. 

We all know the role played by zero in arithmetic 
when high numbers come into play. Zero, the value of 
which is defined as nil, then assumes an extreme im- 
portance, provided it appears on the right side — that is, 
to the right. This is the figure that best gives the notion 
of infinity, if repeated to a sufficient extent. If I were 
a mathematician and were commissioned by the Board 
of Eesearcb, I would willingly write a book on TJie Value of 
Naught. 



EXAMPLES UE DECIPHERING 119 

From the above it follows that, our letter j) never 
occurring to the left of any of the prices, but being often 
repeated to the right, we may boldly conclude p—O. 

Next we observe that the only other letter in the pence 
column besides p is 2, 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. 

Of the bigrams occurring in the shilling column, the 
first letter is always m, and since numbers in this column 
do not go beyond the teens, m can mean no other than 1. 

Parenthetically it may be noted that there are only 
nine different letters in the price list, so that one figure 
out of the ten is unrepresented. This is most likely to be 
9, a figure that is rarely seen in prices of antiques. Nine 
or 19 shillings or pounds is very unusual. Prices 
hovering in the region of the 9'3 are wandermg asteroids 
which usually succumb to the minor attraction of the 
smaller planets 8 or 7, or the increasing attraction of the 
larger planet 10. 

We have presumptively disposed of the four figures 0, 
1, 6, and 9, and almost certainly know that the price of 
the coloured engraving (3), tarnished tray {'2(j), and china 
vase ("29) is £1 lOs. each; of the " Library of Famous Men " 
(32), £1 ; of the copper seal (39), £1 Is.; of the dagger (6), 
engraving (31), and beer mug (43), 10s. each; and of the 
bronze bowl (27) and silver strainer (35), Gs. each. 

The incense burner (2), small Flemish painting (17), 
and bronze hand lamp (28) are all the same price — that 
is, 10s., plus a number of pounds indicated by i. The 
double mirror (4) is the same price, augmented by £10 — 
i.e., mi.mp.p. Further, the engraving (8), picture (9), 
door-knocker (20), and decanter (41) are each priced 
at mi.z. These articles seem to me quite dear enough 



120 CEYPTOGRAPHY 

at 12s. 6d., so that there is no need to ascribe a greater 
value than 2 to i. 

This brings the price of the incense burner, etc., to 
£2 10s., and that of the double mirror to £12 10s. There 
are also two items — silver bell (33) and decanter with silver 
stand (36) — ^priced at £1 2s. 6d. 

The letter i 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 shilhngs; 
the barometer (11), £1 plus r shillings; grandfather clock 
(14), 10 plus r pounds; brown-framed mirror (24), 30 plus 
r shillings; and the candlestick (25) and engraving 
" Magna Charta " (30), each r siiillings. So ubiquitous 
a letter can scarcely be intended for anything but 5. 
Certainly the two last-named objects would not fetch 
more than 5s. each, while sucli quotations for the other 
articles as 15s., £1 5s., £1 15s., £6 5s., and £10 5s. are 
commonly seen. We therefore attach the value of 
5 to r. 

Our attention is now attracted to the letter h. There 
is a mirror (23) at h pounds, a liqueur stand (42) at b 
guineas, and a card table (15) at b pounds 5 shillings. 
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, tlie liqueur stand at b guineas is an obstacle to the 
placing of b at any higher value than 3. Accordingly 
we appraise b at 3. 

Summarising, we have established six out of llic nine 
digits. Those remaining to be discovered are 4, 7, and 8. 

The letter / occurs among the shillings, and is always 
accompanied by z {—(J) pence. This latter factor induces 



EXAMPLES OF DECIPHEiUNG 121 

us to ascribe the value of 7, rather than -1 or S, to f. 
On this assumption, prices are as follows: Black and 
gold tray (10), devotional picture (19), Watteau tray (22), 
locket (87), and chased silver egg-cup (38), 17s. Gd. each; 
rustic scene (12), cheesi^ dish (16), 7s. Gd. each; four 
engravings, the four (13), £1 17s. Gd. 

The letter o occurs three times. The silk mat (34) is 
marked o.p. It was originally marked b.z., or 3s. Gd., 
wliich price has been crossed out. The value is scarcely 
likely to 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 
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 £9. 
'The letter e occurs nowhere else, so W'O have no means of 
drawing any reliable inference. 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 
coniii-mation 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 
(luarter late, greeted me in breathless tones: 

" So sorry, old fellow^; but, you know " 

" Yes, yes; I know," I interrupted. " If you had kept 
me 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 advance. But don't worry. I have not 
wasted my time." 



PART III 

LISTS AND TABLES 

Note. — This third part consists of a series of calculations 
of letter frequencies and combinations in English and 
certain foreign languages. 

Of the practical value of these lists, compiled as a 
result of 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 
persevering study. 

The third factor is flair, or insight. This need not be 
regarded as purely instinctive or in the nature of a lucky 
gift. A 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 
stand an egg in a position contrary to the laws of physics. 

ENGLISH.— I. 

Order of Letter Frequency. 

According to Edgar Allan Poe: E A I D H N R H T U Y, 

etc. 

According to Vesin do Romanini : ETAONIRSHD 

L C W U M, etc. 

122 



LISTS AND TABLES 123 

Norimil frequoiicy table (Hitt) : E T A N I II S 11 \) L 

U C M P E Y W G B V K J X Z Q. 
Telegraphic frequency (Hitt): E A N 1 11 S T D L H U 

C M P Y E G W B V K X J Q Z. 

Order of Frequency of Final Letters. 

According to Valerio: E S D N T E Y E A, etc. (See 
also English. — III.) 

The Commonest Bujrams {Valerio). 

TH, HE, AN, EE, ON, EE, IN, ED, ND, AT, OE, OE, 
HA, EN, NT, EA, etc. 

Freqaencij of Double Letters. 

EE, 00, EE, LL, SS, etc. 

According to Valerio: SS, EE, TT, LL, MM, 00, Fl^, etc. 

The Most Frequent Two-Letter Words {in Order). 

OE, TO, IN, IT, IS, BE, HE, BY, OE, AS, AT, AN, SO, 
etc. 

ENGLISH.— II. 

The Commonest Trigrams (Valerio). 

THE, AND, THA, HAT, EDT (triED To, carriED The), 
ENT, EOE, ION, TIO, NDE, HAS, MEN, NCE, 
OFT, STH. 

The Co?nmonest Three-Letter Words. 
THE, AND, then FOE, AEE, BUT, ALL, NOT, etc. 

The Commonest Four-Letter Words. 
THAT, WITH, EEOM, HAVE, THIS, THEY, etc. 



124 CRYPTOGEAPHY 

Words of One Letter. 
A, 1, 0. 

Proportion of E (Yalerio): 13 per cent. 
Proportion of vowels (Valerio) : 40 per cent. 

ENGLISH.— III. 

(Compiled by Translator.)^ 

Order of Letter Frequency in BeJation to Position 
of Letter in Word. 

Initial letters: TAOMHWCIPBES, etc. 
Second letters: HOBIAUNRT, etc. 
Third letters: E S A R N I, etc. 
Antepenultimate letters: I T E A H N 0, etc. 
Penultimate letters: E N A E H I L C 0, etc. 
Final letters: E T S D N E Y G, etc. (See also Eng- 
lish.— I.) 

Consonant Birjrams at the Ends of Words {Order 
of Frequency). 

NG, Nn, NT, DS, KS, ST, TS, TH, HT, RT, 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) : 

11 w^ords ending HE. 
10 words ending ED. 

^ This and the following sections up to page 138 have been com- 
piled specially for the English Edition. 



LISTS AND TABLES 125 

8 words each ending EE, NG. 
7 words each ending OH, EE. 
G words each ending AT, ND. 
5 words ending NT. 

4 words ending LY. 

3 words each ending AN, DS, EN, ES, LE, ON, EY, SE, 
TY. 

2 words each ending AD, AS, CE, HT, ID, IS, KE, KS, 

ME, NE, OT, OW, ET, SS, ST, TS, TH, VE. 

1 word each ending AL, AP, AE, AY, CH, CT, DE, EE, 

EM, ET, EW, EY, GE, GH, HY, IG, IL, IN, IE, LD, 
LL, LS, LT, NS, NY, OM, OU, EN, ES, SH, TE, UE, 
UL, UE, US, UT, WN, WO, YS. 

Final Trigrams. 

An Enghsh text of 1,000 letters contains, on an average 
(excluding three-letter w^ords): 

5 words ending ING. 

3 words each ending ENT, HAT. 

2 words each ending AVE, EEE, GHT, ION, lED, 

NDS, PLE, ETY, VEE. 
1 word each ending ACT, AIL), AND, ANT, AET, ATS, 
EEN, END, EEY, ESS, EST, HED, HEN, HEE, 
HIS, ICE, lES, ISE, ISH, ITH, LLY, LOE, NCE, 
NED, NTS, OKS, OEE, EED, TED, TEE, UND. 

Initial Consonant Bigrams {Order of Frequency). 

TH, PE, WH, CH, FE, SH, TE, CL, SP, CE, PH, PL, 
BE. GL, SC, SM, ST, WE. 



126 



CRYPTOGKAPHY 



ENGLISH.— V. 

Like Letters at Equal Intervals {separated hy two 
Letters). 



A b b A c y 


AmbAssador 


ArA 


b i 


A 


A m i 


Able 


A 


b 1 


A t i V e 


— A m m 


A — 


A 


b 


Ard 


d A m n 


Able 


— A 


c i 


A — 


c A m p 


Aign 


pA 


c k A g e 


t r A m w 


Ay 


sA 


c r 


A m e n t 


— A n d A — 


e V A 


c u 


At e 


— terrA n e 


An 


rA 


d i 


A — 


m A n i 


A 


heA 


d 1 


And 


— A n s 


Act 


qu A 


d r 


Angle 


A n t 


Agonist 


gr A d u 


Al 


J A n u 


Ary 


A d V 


A — 


c A n V 


As 


h e A d w 


Ay 


c h A p 1 


A i n 


A 


f f 


Air 


A p p 


A — 


A 


f r 


Aid 


— A p t 


A — 


pA 


g e 


Ant 


A q u 


A — 


— A 


g g 


A — 


— Arc 


A — 


m A 


g n 


A — 


w h A r f 


Age" 


— A 


g ^ 


A — 


— A r g 


A — 


A 


h e 


Ad 


— A r i 


A — 


a V A 


i 1 


Able 


r e m A r k 


Able 


clA 


i m 


Ant 


A r m 


A m e n t 


— A 


i n 


A — 


c A r n 


A t i n 


c 111 p 1 A 


i a 


Ant 


— A r r 


A — 


br e A 


k f 


As t 


p A r t 


Ake 


— A 


1 i 


A — 


s t A r V 


A t i n 


— A 


1 1 


A — 


— A s c 


A — 


s i g n A 1 m 


An 


A s b 


A m e d 


pA 


1 P 


Able 


A s i 


A 


— A 


1 t 


A — 


— A s s 


A — 


vA 


1 u 


A — 


d e V A s t 


Ate 


— A 


1 V 


A — 


c A s n 


Al 


A 


1 w 


Ays 


A t 1 


As 



LISTS AND TABLES 



127 



— A 


t t 


A — 


— E a n E — 


— A 


u d 


A — 


c h E a p E r 


r e s t A 


u r 


Ant 


— E a r E — 


cA 


V e 


At 


E a s E 


— A 


V i 


A — 


E a t E — 


dr A 


w b A c k 


— E a V E — 


tA 


X p 


x\ y r 


— E b 1 E 


lA 


y m 


An 


— E c r E — 


B 


a r 


B — 


— E c t Ed 


B 


111 


B 


r E d e E in 


B 


r i 


Be 


E d g E 


suB 


u r 


B 


expE d i Ent 


coB 


w e 


B 


neE d 1 E 


C 


a 1 


C 11 1 a t e 


W 111 d n E s d a y 


— C 


a r 


C — 


r E d r Ess 


C 


a t 


Ch 


d E e p En 


— C 


e n 


Cy 


E f f Ect 


Che 


Ck 


dE f i Ed 


C 


i r 


C — 


— E f 1 Ect 


C 


1 


Ck 


nE g 1 Ect 


sC 


r 


Ch 


s E g m E n t 


SC 


t 


Ch 


— E g r E — 


— c 


r a 


C — 


— cE i V E 


seC 


r e 


Cv 


— E 1 i E — 


C 


r i 


Cket 


— E 1 1 E — 


C 


r 


C — 


E 1 s E 


C 


r u 


C — 


— E 1 t E — 


s p e C 


t a 


Cle 


— E 1 V E — 


— C 


t i 


C 


— Emb Er 


s t a 11 D 


a r 


1) 


p r E mi E r 


D 


e a 


D 


— E m p E — 


D 


e e 


D 


t h E m s E 1 V e s 


D 


i a 


D e m 


— E n c E 


D 


i e 


D 


— E n d E — 


h u n ]) 


r 


D 


— E n g E — 


reD 


U 11 


]) a n t 


c n V E n i Ent 


dE 


a d 


En 


— E n n E — 


E 


a g 


Er 


— E n s E — 


spE 


a k 


Er 


— E n t E — 


— E 


a 1 


E — 


E n V Elo p 



128 CRYPTOGEAPHY 



gE m 


Etry 


m or E v 


Er 


— E p h E — 


— E p 1 


E — 


— E p r 


E — 


— cE p t 


Ed 


— E q u 


E — 


— E re 


E — 


AbE r d 


E en 


— E r f 


E — 


-E r g 


E — 


e X p E r i 


E n c e 


cheE r 1 


Ess 


s E r p 


Ent 


— E r s 


E — 


— E r t 


E — 


— E r V 


E— , 


— E s c 


E — 


bE s i 


Ege 


— E s p 


E — 


— E s s 


E — 


— E s t 


E — 


n i n E t e 


En 


— E t h 


E — 


mE t r 


E 


— E t t 


E — 


b E t w 


E en 


— E u t 


E — 


r E V i 


Ew 


somEwhEre 


n E w y 


Ear 


E X c 


E — 


Exp 


E — 


— E X t 


E — 


F i t 


Fill 


For 


F e i t 


F u 1 


Fil 


G a n 


G 


G a u 


Ge 


-Gg a 


Ge 



— G 


i n 


G — 


neG 


1 ]• 


Gent 


G 


n 


G 


G 


r 


Ge 


— G 


r e 


Gate 


a n G 


u a 


Ge 


sH 


e p 


Herd 


wH 


i c 


H 


H 


i g 


H 


rH 


y t 


Hm 


— cl 


a 1 


I St 


— I 


a t 


Ion 


— I 


c t 


I — 


m I d n 


Ight 


b e s I 


e g 


I n g 


dl 


f f 


I cult 


fl 


f t 


leth 


— I 


g n 


I — 


— I 


1 d 


I — 


— I 


1 1 


I — 


— I 


m m 


I — 


— I 


m p 


I — 


— I 


n c 


I — 


— I 


n d I — 


— I 


n n 


Ing 


— I 


n s 


I — 


— I 


n t 


I — 


— I 


n V 


I — 


— I 


d 


I — 


curl 


s 


Ity 


— I 


pi 


I — 


— I 


p p 


Ing 


-s c r I 


p t 


I — 


— I 


q 11 


I — 


ski 


r m 


Ish 


— I 


r r 


I — 


— I 


s c 


I — 


s a 1 1 


s f 


led 


— I 


s h 


Ing 


dl 


s 1 


Ike 



LISTS AND TABLES 129 



dl 


S 111 I s s 


c N V e N e 


— I 


s s I — 


— N V i N c e 


— I 


s t I — 


b 1 Ong 


BrI 


tain 


b s 1 e t e 


K h a K i 


M r c c 


K 


i c K 


atrO c i Ous 


booK 


111 a K e r 


— c t — 


s i u g 11 L 


a r Ly 


d i Ous 


c a L 


c u Late 


10 g w Oo d 


CG L 


1 u L a r 


p i s On 


M 


a d Man 


fO 1 i 


M 


a i M 


—0 1 1 Ow 


coM 


me ]\I orate 


c m f r t 


M 


11 e M n i c 


— m m — 


M 


u m M y 


c m p — 


iN 


c a N descent 


— n d — 


— N 


c e N — 


e r r n e Ous 


— N 


c i N — 


c n f — 


— N 


c N — 


— n i — 


— N 


d a N — 


c n V — 


— N 


d e N — 


— p h — 


— N 


d i N g 


t p m s t 


— N 


d N 


— p p — 


iN 


f a Nt 


f r b r e 


— N 


g e N — 


f r g Ot 


eN 


g i N e 


— r i us 


— N 


i N 


fO r 1 Orn 


— N 


i u Notion 


enOrmOus 


— N 


k i Ng 


— r p — 


— N 


1 a Nd 


— r r — 


— N 


m e N t 


f r s k 


ca N 


n N 


— s c — 


aN 


i Nt 


e X p 1 s i On 


N 


N 


b 1 s s m 


N 


u N 


— pO s t — 


reN 


w N 


— t i On 


— N 


s e N — 


— t t — 


— N 


t a N — 


b u d i r 


— N 


t e N — 


n X i Ous 


— N 


t i N — 


b X w d 
9 



130 CRYPTOGRAPHY 



bO 


y h d 


diS 


c u 


Ss 


P 


a 1 


Pable 


S 


e a 


Son 


P 


a m 


Per 


s 


e n 


S — 


P 


a u 


Per 


a s S 


e t 


s 


P 


e 


Pie 


diS 


g u 


St - 


P 


e r 


P — 


— She 


s 


P 


r 


Poise 


e n t h u S 


i a 


s — 


P 


r e 


P — 


S 


m a 


Sh 


P 


r 


P — 


— S 


m i 


Ss — 


P 


u 1 


P 


— S 


n e 


Ss 


P 


u m 


P 


c nS 


1 


s 


P 


u r 


P — 


tr e S 


p a 


Ss 


fR 


a t 


R i c i d e 


deS 


P i 


Se 


ext R 


a 


R d i n a r V 


diS 


P 


Se 


— R 


d e 


R 


poS 


s e 


Ss 


R 


e a 


R 


aS 


s i 


St 


ca R 


e e 


R 


ch a S 


t i 


Se 


R 


e p 


R — 


S 


u b S — 


w a R 


f a 


Re 


T 


a c 


T 


p e R 


f 


Rm 


T 


a n 


T 


laR 


g e 


R 


sT 


a r 


T 


— R 


i e 


R 


T 


a s 


Te 


pR 


i m 


Rose 


T 


a u 


T 


— R 


i 


R 


— T 


e c 


T 


— R 


k e 


R 


— T 


e n 


T 


— R 


m e 


R 


T 


e s 


T 


CO R 


n e 


R 


T 


e X 


T 


R 


o a 


R 


ouT 


f i 


T 


— R 


g 


R — 


T 


h a 


T 


— R 


p a 


Rt 


paT h e 


Tic 


— R 


p e 


R 


wiT 


h s 


Tand 


p u R 


p o 


Rt 


— T 


i a 


Te 


— R 


r 


R 


— T 


i e 


Th 


hoR 


S 


Race 


T 


i 1 


T 


c u R 


S 


Ry 


T 


i n 


T 


— R 


t e 


R — 


— T 


i R 


T 


d p a R 


t n 


Re 


o u T 


1 


T 


Fo])R 


u a 


Ry 


d i s T 


o r 


T 


f o R 


w a 


Rd 


T 


n 


T 


a S 


b e 


St OS 


u T 


V u 


T 



LISTS AND TABLES 131 



— U ra 


Ur 


U n h U r t 


U n 1 


Ucky 


U n s 


U i t a b 1 e 


rU p t 


Ure 


tU r b 


Ulent 


p U r s 


Ue 


m U s c 


Ular 


MU s e 


Urn 


br U s q 


Ue 


gU t t 


Ural 


V a 1 


Ve 


V e 1 


Vet 


— V 1 


Ve 


Way 


Ward 


Z i g 


Zag 



— T r a T e 

— T r e T — 
— T r i Tion 

b e T r o T h 
T r u T h 
ouT s 6 T 
aT t i Tude 
s i T u a Tion 
T u f T 
g r a T u i T — 
sU b d Ue 
s U c c U m b 

— U c t U — 
s U f f Use 

— U 1 o Us 

— U 1 t u — 
h U m b U g 



ENGLISH.— VL 

Ld^e Letters at Equal Intervals (separated hy 
Three Letters). 

A b e y A n c e ' d A h 1 i A 

h A b i t A b 1 e r A i 1 w A y 

— lAborAt^ AcquAintAnce 

AbreAst stAircAse 

A b r o A d c h A i r m A n 

AbstAin mAlefActor 

— A b ul A — 
c o A c h m A n 
b 1 A c k m Ail 

bAckwArd 

A c t u A 1 
erAdicAte 

A d o r A b 1 
heAdquArters 

A e r i Al 

— Agin A — 
coAgulAte 



-A 1 g 


i A 


— A 1 i 


s A t i n 


— A 1 1 


i A — 


Al p 


h Abet 


A 1 r 


e A d y 


s t A m b 


A t 


n A m e 


s A k e 


— A m i 


n At — 


f i n A n c 


i Al 


1 A n d 


m A r k 


c h A n g 


e A b 1 e 



132 CRYPTOGEAPHY 



1 x\ n g u A g e 


technical 


g Ang w Ay 


C hur Ch 


mecliAnicAl 


— C i e n C — 


A n i m A — 


— C i f i C 


orgAnisAtion 


n e C k 1 a C e 


A n n u A 1 


C 1 i n C h 


A n m A 1 y 


Clutch 


niA n s 1 Aughter 


C a 1 C e 1 1 a r 


trAnspArent 


C m i Cal 


substAntiAl 


C n s C — 


mAnufActure 


Council 


d i 1 A p i d A t e d 


Crutch 


c A p i t A — 


elect r i C 


A p p e A — 


Cubic 


A p p 1 A u d 


— C u r a C y 


c h A r c o Al 


D e c i D e 


chArgeAble 


D e 1 u De 


chAritAble 


D e s i Deration 


pArliAment 


D i V i De 


A r m A 


WeDnesDay 


A r r e Ar s 


D r e a D 


— A r r i Age 


D w i n D 1 e 


— A r t i Al 


— E a b 1 E 


reAsonAble 


— E a c h E — 


A s p h A 1 1 


E a g 1 E 


A s s u Age 


mE a g r E 


A s t r A — 


1 E a g u E 


lA t e r Al 


c 1 E a n s E 


— A t i c A 


f E a r 1 E s s 


coA t f Arms 


— E a r n E 


— A t u r A — 


rehEarsE 


A V G r A ge 


E a s t E r 


n A V i g A t e 


— E a t h E — 


AvocAtion 


— E b a t B 


A V w A 1 


d E c i d E 


A w k w Ar d 


s p l^j c i m E n 


a B s r B 


p r E c i s E 


V ol C a n i C 


corrEctnEss 


C h a n C e 


R E c u r E 


C h a r C o a 1 


— E c u t E 



LISTS AND TABLES 



133 



m E d i a E V a 1 

— E due E 
schE d u 1 E 

dE f a c E 

— E f o r E 
— IE g a t E 

b E h a V E 
dE i f i Ed 
f or E i g n Er 
E i t h Er 

— E 1 a t E 

d E 1 i 11 E a t e 

dE 1 i V Er 
en V E 1 p E 

bE 1 o V Ed 
E 1 u d E 
E m b 1 E m 
E m i n E 11 1 

— E III i s E — 

r E 111 u n E r a t — 
p a r E 11 t li E s i s 
sevEntiEth 

gE n t I E 

c E n t r E 
E n u m E r a t — 

pE o p 1 E 

rE p 1 i Ed 

— E p o s E 
scE 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 

expErimEnt 

dE r i V E 
nevErthEless 
i n t E r V i E w 
w h o 1 E s a 1 E 
r E s c u E 



r E s i d E 
dE s i r E 

— E s o 111 E 

— E s q u E 
invEstmEiit 

r E s u ra E 

— E t c h E — 
r E t i c E 11 1 

1 i f l^J t i m E 
b E t o k E n 
EvanEscent 

— E V i c E 

E V i d E n t 

— E V i s E 
r E V i V E 

E vo k E 
benEvolEnt 
d E V o t E 
b E w a r E 
ExchEquer 
E X p r Ess 
E X t r Erne 
F e a r F u 1 
G o i n G 
G r u d G e 
H a r s H 
H a t c H 
arcH bi s Hop 
H e a t H 
H e i g Ht 
H i t c H 
t H o u g H 

— I e n t I — 

— I g h t Iiig 
CO 111 pi 1 a t loii 

pilgrim 

villain 
I in p 1 I c i t 
I 111 p r I — 
I 11 q u I — 



134 



CKYPTOGEAPHY 



Inspire 
Instinct 
Intuition 
diminution 
Invoice 
circuit 
mischief 
d I s c r I — 
biscuit 
dl s t r I — 
capitalist 
— It at Ion 
stocKbroKer 
K n a c K 
Kno cK 
K n u c K 1 e 
L a b e L 
L a n d L — 
c L e a r L y 

L e g a L 
s i L e n t L y 
compLeteLy 
■ L i b e L 
symboL i c a L 
L i k e L y 
L i V e L y 
L o c a L 
L o V e L y 
L o y a L 
LuckLess 
absoLuteLy 

M a d a M 
c o m M a n d M e n t 
M a X i M u m 
aMazeMent 
aMendMent 
s o M e t i M e s 
M i n i M u m 
M n u M e n t 
M o V M n t 



— M p 1 e M e n t 
c o M p 1 i M e n t 
s y M p t o M 

aMuseMent 

— r N a m e N t 
c N c e r N 

e N c h a N t 
a N c i e N t 

— N d e m N 
c o N d i g N 

— N d me N t 

— N e m e N t 
coN f r o Nt 

EN g 1 a Nd 
s a N g u i N e 
cogNisaNce 
pheNomeNon 
N o m i N — ■ 
k N w i Ng 
deli N q u e N t 

— N s h i N e 

— N s i g N 

— N s i o N 

— N s t a Nt 

— N t a i N 
iN t e r N 

— N t i o N 

b V i u s 

— c c t i On 

— r g a t r y 
t i 1 s me 

w h 1 e s m e 
s 1 i ] q uy 
sOmebOdy 
s m e h w 
cOmprOmiso 
c n g 1 meration 
c n t r — 

a n n y ni u s 
p" t i n 



LISTS AND TABLES 135 

f r g — S e 1 e S s 

h r i z n — S e n e S s 

pOrtfOlio diSguiSe 

Orthodox boSideS 

— OrtiOn buSineSs 
tOrtuOus suSpenSo 
pOstpOne diSperSe 

— OtatO — reSponS — 
throughout — S t i c S 

neighbourhood ChriStmaS 

Out d S t r e Ss 

Outgoing abStruSe 

ParaPet TainT 

PersP— — TeenTh 

PhosPhate TempT 

PlumP — TeraT — 

Postpone TheaTre 

P r o m P t T h e f T 

ProsPer au Then Tic 
Q u i n Quennial h i T h e r T o 

— R a t o K — T h i r T y 
tRaveRse aThleTic 

suKchaRgo wiThouT 

a R d o u R a m e T h y s T 

RecoRd — TiciTy 

R e V e R s e — T i e n T 

buR g 1 a R — T i g a Te 

bRibcRy TighT 

auRifcRous esTimaTe 

w R i t e R — T i n c T 

coRkscRew culTivaTe 

aRmouR — pTiviTy 

bRokeR — TmenT 

p R o p e R u T m o s T 

f u R t h e R T o a s T 

ARthuR — TracT 

paR t n e R — T r a i T 

moRtuaRy conTrasT 

disburse TroaT 

d i S c 1 o S e s T r e e T 



336 CKYPTOGPtAPHY 

sTricT sUmptUous 

paTrioT pUnctU — 

— sTrucT Undoubted 

TrusT conUndrUm 

congraTulaTe UniqUe 

— orTunaTe — UrioUs 

T w e n T y U s e f U 1 

trusTworThy — U t h f U 1 

q U a d r U — b e a U t i f U 1 

chaUffeUr caUtioUs 

rUinoUs — VatiVe 

frU i t f U] V t i Ve 

sU 1 p h Ur W e s t Ward 

scUlptUre WindWard 



ENGLISH.— VII. 

Three Like Letters with Intervals of One. 



e X 



p A 1 A t Able 


— E V E r E 


MA 1 Ay A 


— Ibl 1 Ity 


C A n A d A 


— hibl t Ion 


c A r A V A n 


rigl d Ity 


cAtArAct 


diml n Ish 


.trAvAgAnt 


I n I t I a — 


— E c E d Ent 


— I s I t I — 


p i E c E m E a 1 


criti c Ism 


p r E d E c E s s r 


c I V I 1 Ian 


r E f E r E n c e 


d I V I s Ion 


— EgEiiErate 


10 c OmOtive 


vE h EmEnt 


chrOn 1 Ogy 


E 1 EmEnt 


m n p I y 


E 1 E V E n 


m n t n — 


c EmE t Ery 


chlO r f Orm 


\v h E 11 K V E r 


— sT i T u T — 


w h E r E V E r 


UnU s Ual 



LIST8 AND TABLES 137 



Bifjrams Bcpeated. 



CO Oa 


TH i THer 


— dE : 


EEE 


mUE m UE 


ICICle 


oUT p UT 


— ININ 


g 


h A B i t A B 1 e 


bAG 


g 


AGe 


CHur CH 


bAE 


b 


A E o u s 


p i C K p C K e t 


BA 


r 


B A r u s 


DE c i DE 


CA 


1 


C A r e u s 


DE 1 u DE 


CA 


s 


C Ade 


IN cl INe 


CA 


u 


C As — 


IN f r I N g e 


DA 


r 


D An ell es 


I N s t I N c t 


p E A 


c 


EAble 


perlTonlTis 


— EN 


d 


ENt 


OE a t OE 


— EN 


t 


EN 


P H s P H a t e 


r e V E E 


b 


EEate 


PO s t POne 


— EE 


g 


E E 


Q U i n (^ Uennial 


pEE 


V 


EEse 


remissness 


— IN 


g 


INg 


d i S T r u S T 


— IN 


k 


INg 


f r T H w i T H 


MA 


d 


MAn 


sENtimENt 


MU 


r 


MUr 


M AtheM A tics 


— NG 


i 


NG 


seNTimeNT 


— NT 


e 


NT 


coNTineNT 


ON 


i 


ON 


cOUrteOUs 


PA 


1 


PAble 


plENipotENtiary 


PO 


r 


POise 


intEEpretEE 


EE 


d 


EEss 


coNTraveNTion 


EE 


P 


E E sen t 


cOUrageOUs 


SE 


n 


SE_ 


i n T E r p r e T E r 


aSS 


a 


S S i n 


uNDerstaND 


— SS 


e 


SS 


etc., etc. 



138 



CRYPTOGEAPHY 



Words of Ten Different Letters, ivhicli may he used in Suh- 
stitution of the Figures 0-9 or 1-0, and thereby form 
Numeral Key-Words. 



AUTHORISED 

BACKGROUND 

BANKRUPTCY 

BUCKINGHAM 

CHIVALROUS 

COMPATIBLE 

COMPLAINTS 

DESOLATING 

DESTROYING 

EXHAUSTION 

FLOURISHED 

FORMIDABLE 

GELATINOUS 

HYDRAULICS 



HYPNOTISED 

HYSTERICAL 

ILPRACOMBE 

IMPERSONAL 

IMPORTANCE 

JOURNALIST 

LACHRYMOSE 

MACKINTOSH 

MENDACIOUS 

METAPHYSIC 

MINERALOGY 

MISFORTUNE 

MODERATING 

PATRONISED 



PATRONYMIC 

PLAYWRIGHT 

PRESUMABLY 

PREVIOUSLY 

PROCLAIMED 

PROFLIGATE 

PROMULGATE 

PURCHASING 

REGULATION 

REPUBLICAN 

SUBJECTION 

SYMPATHISE 

UNSOCIABLE 

WORKINGDAY 



Surnames such as Tichhourne, or short sentences such 
as Fair Custom, may also be used. 



ENGLISH.— VIIL 





Projjortion of 


Words if 


. Webster's Dictionary 






classified 


accordin(j 


to tlieii 


- Initials. 






Per Cent. 


Total. 




Per Cent. 


Total. 


A . 


G-43 


6-43 


N . 


1-61 


58-71 


B . 


5-35 


11-78 


. 


2-38 


61-09 


C . 


9-82 


21 -GO 


P . 


8-51 


69-60 


D . 


5-95 


27-55 


Q . 


0-59 


70-19 


E . 


4-22 


31-77 


R . 


4-94 


75-13 


F . 


4-22 


35-99 


S . 


. 12-02 


87-15 


G . 


3-27 


39-26 


T . 


5-83 


92-98 


H . 


3-G9 


42-95 


U . 


1-55 


94-53 


I . 


4-22 


47-17 


V . 


1-84 


96-37 


J 


0-83 


48-00 


w . 


2-92 


99-29 


K . 


0-77 


48-77 


X . 


0-12 


99-41 


L . 


3-39 


52- IG 


Y . 


0-30 


99-71 


M . 


4-94 


57-10 


Z . 


0-30 


100-01 



LISTS AND TABLES 131) 

Tho extra O'Ol per coiit. in tlio total is duo to the 
approximate nature of the calculations. The above pro- 
portions vary from one dictionary to another. 



FEENCH.— I. 
Order of Letter Frequency, 

According to Valerio; ENAIESTUOLDCMPVP, 

etc. 
According to Langie (in the works of Bossuet, Voltaire, 

Maupassant, and France) : E S A T I N, etc. 
According to Kasiski: ESRIANTOUL, etc. 

Order of Frequency of Final Letters {Valerio). 
E S T K A N L I U D, etc. 

Tlie Comvwnest Bigrams {Valerio). 

ES, EN, LE, DE, ON, OU, NT, EE, NE, ED, TE, EM, 
SE, EE, AE, ME, AN, IT, ET, IE, TI, EL, NS, UE. 

Frequency of Double Letters. 

According to Valerio: SS, LL, EE, NN, TT, EE, CC, EE, 

MM, PP. 
According to Kasiski: SS, EE, NN, TT, EF, CC, EE. 

Double Letters at the End of Words. 
EE. 



140 CEYPTOGRAPHY 

FRENCH.— II. 
The Commonest Trigrams. 

According to Valerio: ENT, EDE, LES, LLE, QUE, AIT, 
EME, ION, EUR, ELL, SSE, EST, DAN, DEL, 
MEN, DES, TIO, ESE, ANS. 

According to Kasiski: ENT, QUE, ION, QUI, TIO, ONT, 
AIT, ANT, OUR, ANS, LES, AIS, OUS. 

The Commonest Two-Letter Words. 
AN, AU, CE, CI, DE, DU, EN, ET, IL, JE, LA, LE, MA, 
ME, NE, NI, NU, ON, OU, SA, SE, SI, TA, TE, TU, UN. 

Words of One Letter. 
A, 0, Y. 

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.— III. 

Order of Letter Frequency in Relation to Position 
of Letter in Word. 

Initial letters (Valerio) :DLEPACSMRIF, etc. 

Second letters (Langie) : E A U N R I T, etc. 

Third letters (Langie): S E U N T I R, etc. (order in- 
different). 

Antepenultimate letters (Langie): E, followed by A, 
I L (no order), etc. 

Pciiultimato letters (Langie) :EUNILORS, etc. 

Final letters (Valerio) : E S T R A N L I U D C X, etc. 



LISTS AND TABLES 141 

Initial Consonant Bigrams (Valerio). 

BL, BE, PL, PE, FL, FE, VE, CL, CE, GL, GE, TE, 
DE, CH, PH, TH, SC, SP, ST. 

Final Consonant Bigrams (Valerio). 

NT, NS, ET, NC, CT, EC, SC, ND, ED, NG, EG, 
MP, NQ, ST, GT (doigt, vingt), SS (express). 



FEENCH.— IV. 
Final Bigrams (Langie). 

A French text of 1,000 letters contains, on an average 

25 words ending ES. 
23 words ending NT. 
10 words ending EE. 

9 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. 

Final Trigrams (Langie). 

A French text of 1 ,000 letters contains, on an average 
9 words each ending LES, ENT. 
7 words ending ONT. 
6 words ending EES. 
5 words each ending GES, INE. 
4 words ending TEE. 



142 CRYPTOGKAPHY 

FRENCH.— V. 

Five-Letter Group repeated in Succession. 
FAIRE FAIRE. 

Bepeated Groups separated hy a Single Letter. 
VIS A VIS, 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 hy Three Like Letters. 

CREEES. 

Q is always followed by U in the body of a word. 

X is preceded by U, except in the words six, dix, fixe, 

prolixe, mixture, etc., axe, sexe, boxe, etc. 
H is preceded by : 

C, as in chemin, cheval, cher, etc. 

P, as in photographie, etc. 

T, as in theatre, etc. 

Word of Tivelve Different Letters, irhich may be used in 
Substitution of the Figures 1-12, and thereby form 
a Numerical Key-Word. 

IMPREVOYANTS.i 

* In English "considerably" might be used.— Translatok. 



LISTS AND TABLES 



143 



FRENCH.— VL 

Like LeUers at Equal Intervals {sejjarated hy Two 
Letters). 



A f f 


Aire 


p r E m i 


Er 


c A ] c 


Aire 


e V i d E m ro 


Ent 


c A m p 


A g n e 


f E m m 


E 


f r A n ? 


A i s 


E m p 


E r e u r 


sc A n d A 1 e 


— E n c 


E 


e br A n 1 


A 


— E n d E — 


— A p p 


A — 


gE n i 


E 


bA r b 


Are 


— E n n 


E 


— A r r 


A — 


dE n r 


Ee 


p A r t 


Age 


gE n r 


E 


— A s s 


A — 


d ef E n s 


E 


A t 1 


As 


E n s 


Eigner 


— A t t 


A — 


E n s 


E m bl e 


Bar 


Bare 


p E n s 


Er 


Bom 


Be ^ 


— E n t 


E — 


C a 1 


C a i'r e 


inE p t 


E — 


C i r 


C — 


— E q u 


E — 


— C n 


C — 


hE r b 


E 


— C r 


C — 


— E r c 


E — 


s p e C t a 


Cle 


— E r i 


E — 


— D a r 


D — 


c a s E r n 


E 


D i n 


De 


g 11 V E r n 


E m e n t 


pE c h Er 


— E r t 


E 


siE c 1 


E 


— E r V 


E — 


sE c r 


Et 


dE s c 


E n d r e 


r E c u 


Eillir 


r E s p 


Ect 


E f f 


Et 


— E s s 


E — 


rE f 1 


E c h i r 


— E s t 


E — 


rE g i 


E 


— E t r 


E — 


s E g m 


Ent 


— E t t 


E — 


a 1 1 E g r 


]^] s s e 


— E u 1 


E 


c h a n c E 1 i 


Er 


— Eur 


E — 


— E 1 1 


E 


— E u s 


E 



144 CEYPTOGEAPHY 



— E u 


V E — 


expl s i On 


rE V 


u E 


— t i — 


E X 


p Edition 


t u j Ours 


F 


r Fait 


cO u r Onne 


G 


n G 


P e u Pie 


H a 


c He 


P r e P — 


vie 


time 


P r P — 


dl f 


f Icile 


extrE a o Edinaire 


si g 


n I f i e r 


E e p Eesentant 


al g 


u Ille 


ompeE e u E 


r ecu el 1 


1 Ir 


— E i e E 


ml 1 


1 Ion 


c E i E e 


I m in I — 


— E r u E — 


a I n 


s I 


mo E s u E e 


— cr I p 


t Ion 


foE t e Eesse 


mis 


s Ion 


— E t i E 


c a p I t 


a In e 


— S a i S — 


mill t 


aire 


S a n S 


— L 1 


u L — 


S e n S 


iN f 


a N t er i e 


S o u S 


uN i 


N 


a S s i S t er 


a N n 


N c er 


deS s u S 


— N s 


e N — 


S u i S 


— N t 


a N — 


— T a n T — 


— N t 


i N — 


— T e n T — 


— N t 


N — 


T r T 


pO i 


s On 


— T u T 


fO 1 


i 


a T t i T u d e 


e m in t i n 


beaUcoUp 


— n 


t — 


rU p t Ure 


C) p p s e 


po U r q U oi 


— r 


d — 


poUr s Uite 


hO r 


1 Oge 


br U s q U e 


-0 r 


t — 


etc., etc. 



LISTS AND TABLES 14.' 

Like Letters at Equal Intervals {separated hy 
Three Letters). 



— 11 A i s s A — 


Lege 


,L 


c A p i t A i 11 e 


L c a 


. L 


C h a 11 C e li e r 


si M p 1 e 


Ment 


pr E c i s E r 


e N c h a 


N t er 


s E c u E r 


c i N q u a 


,Nte 


— Ecut Er 


i N s t a 


N t 


— E i 1 1 E — 


iN s t i 


Net 


E 1 g E 


h r i z 


On 


— E m b 1 E r 


pr p r t i 


On 


— E n d r E 


P u r 


Pre 


— E n t i E r 


f a R c e u 


R 


E n t r E r 


p a R c u 


Rir 


— E r i t E — 


p R e n d R e 


p r E s q u E 


1 a R g e u 


n 


m E s u r E r 


p R p 


t R t i n 


EtagE 


m a R q u e 


sR 


E t a 1 E r 


deS s a i 


Sir 


pE t i t E 


Tr ai 


T 


m E t t r E 


eT r i 


T 


E X t r E m e 
Instinct 


etc., etc. 


Bigrams 


Repeated. 


- 


cAN c AN 


g u V E R n 


E R 


b A R b A R e 


vER s 


E R 


B A r B Ar e 


— EU r 


E U — 


CA 1 CAire 


MU r 


MUre 


CA n CAn 


— NS e 


NS 


CA s CAde 


— NT a 


NT 


— EN d ENt 


— NT e 


NT — 


—EN s ENt 


— ON t 


ON — 


EN t EN — 


tou j 


OUrs 


— BR c ER 


PR 


PRe 


— ER g ER 


SE n 


SE 


f E R m E R 


poSS e 


SSion 
10 



146 



CEYPTOGEAPHY 



TE n TEr 


loINt aIN 


' fUE e U E 


QU e 1 QUe 


mUE m UEe 


— EEndEE 


CHer CHer 


TE a i TEe 


chEEchE E 


qUEl q UE 


cHEr c HEr 


etc., etc. 



EEENCH.— VII. 

Proportion of Words in Littre's Dictionary classified 
according to their Initials. 





Per Cent. 


Total. 




Per Cent. 


Total. 


A . 


6-00 


6-00 


N . 


1-90 


61-55 


B . 


3-80 


9-80 


. 


2-60 


64-15 


C 


10-80 


20-60 


P . 


10-60 


74-75 


D 


6-75 


27-35 


Q .. 


0-80 


75-55 


E . 


7-00 


34-35 


E . 


7-50 


83-05 


F 


4.85 


39-20 


S . 


7-10 


90-15 


G 


3-30 


42-50 


T 


5-60 


95-57 


H 


2-50 


45-00 


U . 


0-40 


96-15 


I 


3-50 


48-50 


V . 


3-20 


99-35 


J 


1-30 


49-80 


w . 


0-02 


99-37 


K . 


0-05 


49-85 


X . 


0-03 


99-40 


L 


3-00 


52-85 


Y . 


0-01 


99-41 


M . 


6-80 


59-65 


Z 


0-11 


99-52 



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. 

According to Valerio: E I A E L N T S C D P, etc. 
According to Vesin do Eomanini: Fj I A 0, followed by 
L N E S, etc. 



LISTS AND TABLES 147 

Order of Frequencij of Final Letters {Langie). 
I A E N L E D U. 
The same letter frequently ends two, three, four, or 
five consecutive words. 

The Commonest Bigrams {Valerio). 

EE, ES, ON, EE, EL, EN, DE, DI, TI, SI, LA, AL, AN, 
EA, NT, TA, CO, IN, LE, TO, 10, AE, NE, OE. 

Frequency of Double Letters (Valerio). 
LL, SS, TT, EE, PP, NN, BB, GG, CO. 
All the consonants may be doubled except H, J, and Q. 

Words of One Letter. 
A, E, I, 0. 

ITALIAN.— 11. 

The Commonest Trigrams {Valerio). 

CHE, EEE, ZIO, DEL, ECO, QUE, AEI, ATO, EDI, 
IDE, ESI, IDI, EEO, PAE, NTE, STA. 

The letters J and H are always followed by a vowel. 
The letter H is used only in the groups CH and GH, 

and in four forms of the verb avere (to have): HO, 

HAI, HA, HANNO. 
The letter Q is always followed by U. 
Proportion of E (Valerio): 12-G per cent. 
Proportion of vowels (Valerio) : 4G per cent. 

SPANISH.— I. 

Order of Letter Frequency (Valerio). 

EAOSIENLDTCUP, etc. 



148 CEYPTOGEAPHY 

Order of Frequency of Final Letters (Valeria). 
AESONLEYIDZU, etc. 

The Commonest Bigrams [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 Letters. 
CC, LL, EE, infrequently AA, EE, 00, NN. 
According to Valerio: EE, LL, EE, SS, DD.^ 

Words of One Letter. 
A, E, 0, U, Y. 
Single-letter words that may occur in succession are 
A or Y A. 

SPANISH.— II. 

The Commonest Trigrams (Valerio). 

QUE, EST, AEA, ADO, AQU, DEL, CIO, NTE, OSA, 
EDE, PEE, 1ST, NEI, EES, SDE. 

Douhlcd Letter beginning a Word. 

LL. 

The letters Z, J, H, and V are always followed hy a vowel. 
Q is always followed by U. 
Proportion of E (Valerio): 14 per cent. 
Proportion of vowels (Valerio) : 48 per cent. 

* Neither S nor D can be doubled in the same word, but they 
occur consecutively as tlie final of one word and tlie initial of the 
next. — TiiA Ns i ,A'r< ) R. 



LISTS AND TABLES 149 

GERMAN.— I. 

Order oj Letter Frequency. 

According to Kasiski: E N I R S T U D A H, etc. 
According to Valerio : E N R I T S D U A H, etc. 
According to Vesin do Romanini: E, then N I R S U Uj 
etc., the rarest being Q X Y J C. 

Order of Frequency of Final Letters (Valerio). 
N E R T S D H U Z F, etc. 

The Commonest Bigrams (Valerio). 

EN, ER, CH, ND, DE, IE, TE, RE, EI, UN, GE, DI, 
ES, BE, IN, IT, HE, etc. 

The Commonest Final Bigrams, 
EN, ER, then the letters S, T, and E. 

Frequency of Double Letters. 
EE, TT, LL, SS, DD. 

Double Letters at the End of Words. 
NN, SS, less frequently LL, EE. 

GERMAN. II. 
The Commonest Trigranis (Kasiski). 

EIN, ICH, DEN, DER, TEN, CHT, SCH, CHE, DIE, 
UNG, GEN, UND, NEN, DES. BEN, RCH. 



150 



CEYPTOGRAPHY 



The Commonest Two- Letter Words. 

AB, AM, AN DA, DU, ER, ES, IM, IN, OB, SO, UM, 
WO, ZU, then JA, NU, etc. 

The bigram UN frequently commences a word. 

Q is always followed by U. 

C is always followed by H or K, except in " foreign " 

words. 
Proportion of E (Valerio): 18 per cent. 
Proportion of vowels (Valerio) : 35 per cent. ; (Kaedingj : 

42-12 per cent.^ 



GERMAN.— III. 

Provortion of Words in Sachs's Dictionary classified . 
according to their Initials. 

Per Cent. Total. 





Per Cent. 


Total. 




A 


. 11-50 


11-50 


N 


B 


7-70 


19-20 





G 


0-66 


19-86 


P 


D 


3-92 


23-78 


Q 


E . 


5-15 


28-93 


R 


F 


4-67 


33-60 


S 


G . 


6-44 


40-04 


T 


H . 


6-56 


46-59 


U 


I 


0-96 


47-55 


V 


J 


0-77 


48-32 


w 


K . 


6-40 


54-72 


X 


L 


3-45 


58-17 


Y 


M 


3-30 


61-47 


Z 



1-92 


63-39 


0-88 


64-27 


3-08 


67-35 


0-26 


67-61 


2-46 


70-07 


10-67 


80-74 


2-80 


83-54 


3-96 


87-50 


4-90 


92-40 


3-85 


96-25 


0-03 


96-28 


0-03 


96-31 


3-08 


99-39 



^ The German authority, F. F. W. Kaoding, based liis calcu- 
lations on a total of 00,558,018 letters (I) ; he established the 
presence of 9,200,044 E's, 0,303,537 N's, etc. It may be noted 
that one volume of the large dictionary of Larousse contains about 
20,000,000 characters. 



LISTS AND TABLES 151 

The shortage of U-Ol per cent, in the total is due to 
blanks and the approximate natiu'e of the calculations. 

Note. — These proportions vary from one dictionary to 
another. In this case, the middle of the dictionary occurs 
at K; in Feller's pocket dictionary it occurs at M; in 
Niethe's numbered dictionary at L, etc. 



EUSSIAN.— I. 

Order of Letter Frequency.^ 

According to Langie: OANISETVKLKM, etc: 

The letter I predominates in French transcription. 
Texts in Russian characters : A I L E N, hard sign, 

D T M V R U K P, etc. 
English transhterations : OYAIELNHDTSMUV 

E Z K P, etc. 
French transliterations: OIAELNTDCHMUVR 

K P, etc. 

Final Letters.^ 

According to Langie : Hard sign, then 0, E, (I)A, I, (K)H 

(ignoring final hard sign), A, I, etc. 
Russian texts : Hard sign, U, I, soft sign, (Y)A, E, Y, A, 

M, V (ignoring final hard sign), etc. 
English transliterations : A, U, E, 0, I, soft sign, Y, M, V, 

etc. 

The Commonest Bigrams (Langie). 

ST, NO, EN, GO, TO, KA, KO, NA, ER, RA, LI, SK, 
OS, M', RO, PO, ZA. 

^ Including results of supplementary investigations by trans- 
lator. 



152 CEYPTOGEAPHY 

The Commo7iest Trigrams (Langie). 

AGO, STV, ENI, OST, YKH (bigram in Eussian 

characters), TOE, STA, IKH (bigram), BNN, NOV, 

OEO, STO, EGO, LIS, NI(IU, 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, EE, (I)U(I)U, SS, 00, ZZ. 

Words of One Letter {Langie). 
I, (I)A, 0, U, A. 
Note by Translator. — To these may be added, if 
final hard sign is ignored, V, K', S'. 

EUSSIAN.— II. 
The Commonest Tetragrams {Langie). 
NY(K)H', PEAV, TSTV, VENN, UET:, VSTV. 
Note. — ^The colon in the above represents the soft sign. 

The Commonest Pentagrams {Langie). 
SKAGO, STVIE, L:STV. 

The Commonest Hexagrams {Langie). 
STVENN, NNOSTI. 

Like Letters separated hy One Letter {Langie). 

ILI, KAK, OBO, OVO, OGO, ODO, OKO, OLO, ONO, 
OSO, POP, TOT, TUT, etc. 

Proportion of (Langie): 10-7 per cent. 
Proportion of vowels (Langie): 43-5 per cent. 



LISTS AND TABLES 



153 



PORTA'S TABLE. 

This table was composed by Giovanni Battista da 
Porta, a Neapolitan physician, author of a work on 
cryptography entitled De furtivis litterarum noiis, vulgo 
de ziferis, Naples, 1563. 



A 


B 


a 


b 


c 


d 


e 


f 


g 


h 




J 


k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


C 


D 


a 


b 


c 


d 


e 


f 


g 


h 




j 


k 


1 


m 


z 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


E 


F 


a 


b 


c 


d 


e 


f 


g 


h 




j 


k 


1 


m 


y 


z 


n 


o 


P 


q 


p 


s 


t 


u 


V 


w 


X 


G 


H 


a 


b 


c 


d 


e 


f 


g 


h 




j 


k 


1 


m 


X 


y 


z 


n 


o 


p 


q 


r 


s 


t 


u 


V 


w 


1 J 


a 


b 


c 


d 


e 


f 


g 


h 




j 


k 


1 


m 


w 


X 


y 


z 


n 


o 


p 


q 


r 


s 


t 


u 


V 


K 


L 


a 


b 


c 


d 


e 


f 


g 


h 




J 


k 


1 


tn 


V 


w 


X 


y 


z 


n 


o 


p 


q 


n 


s 


t 


u 


M 


N 


a 


b 


c 


d 


e 


f 


g 


h 




j 


k 


1 


m 


u 


V 


w 


X 


y 


z 


n 


o 


p 


q 


r 


s 


t 





P 


a 


b 


c 


d 


e 


f 


g 


h 




j 


k 


1 


m 


t 


u 


V 


w 


X 


y 


z 


n 


o 


P 


q 


r 


s 


Ql 


R 


a 


b 


c 


d 


e 


f 


g 


h 




j 


k 


1 


m 


s 


t 


u 


V 


w 


X 


y 


z 


n 


o 


P 


q 


r 


S 


T 


a 


b 


c 


d 


e 


f 


o 


h 




j 


k 


1 


m 


r 


s 


t 


u 


V 


w 


X 


y 


z 


n 


o 


p 


q 


U 


V 


a 


b 


c 


d 


e 


f 


g 


h 




J 


k 


1 


m 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


n 


o 


P 


W 


X 


a. 


b 


c 


d 


e 


f 


g 


h 




j 


k 


1 


m 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


n 


o 


Y 


z 


a 


b 


c 


d 


e 


f 


g 


h 




J 


k 


1 


m 


o 


p 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


n 



The capital letters on the left serve to form the key 
or word agreed upon, the letters of which, in succession, 
indicate the alphabets selected. Each pair of capitals 
jointly control the alphabet ranged in two lines to their 
right. 

Let us suppose that the capital letter G is used to cipher 
the plain letter i. It will be noted that in the double 
line to the right of G the letter s occurs immediately 



154 CEYPTOGEAPHY 

below i ; 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 
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 word " red " by means of the key-word 
CAE, we first look for r in the double line to the right 
of C, and find immediately above it the letter /. Proceed- 
ing in like manner with the second letter e (key-letter A), 
and the third letter d (key E), we obtain the result: 

red 
CAE 
=f r V 

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 table was established by Blaise de Vigenore, 
translator and French diplomat, author of a work 
entitled Traite des chijjres ou secretes juanieres d'ecrire, 
Paris, 158G. 

The upper horizontal line of capitals represents the 
plain-text alphabet ; the column of capitals to the left is 
used to form the key-word. 

Supposing the first letter of the key-word is E, and the 
first letter of the plain text i, we descend from I in the 
top lino of capitals until wo reach the line of small letters 



LISTS AND TABLES 



155 



beginning from K in the column to the left. At the point 
of intersection we find z, which becomes the first letter 
in the ciphered text (see p. 28). 

To decipher the word " kik " by the aid of the key- word 
REX, we first look for k in the horizontal line beginning 

























Letters 


of Pl 


oin text 
























A 


B 


C 


D 


E 


F 


G 


H 


1 


J 


K 


L 


M 


N 





P 


q|r 


S 


T 


U 


V 


W 


X 


Y 


z 




A 


a 


b 


c 


d 


e 


f 


S 


h 


i 


J 


k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


z 




B 


b 


c 


d 


e 


f 


fi 


h 


i 


i 


k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


a 


Y 




C 


c 


d 


e 


f 


^ 


h 


i 


J 


k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 
a 


a 


b 


X 




D 


d 


e 


f 


g 


h 


i 


,i 


k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


b 


c 


W 


'cB> 


E 


e 


f 


g 


h 


i 


,j 


k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 
e 
f 


V 
U 

T 


-~j 


F 


f 


S 


h 


i 


,i 


k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 




G 


fi 


h 


i 


,i 


k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


C 


H 


h 


i 


i 


k 


i 


m 


n 


o 


p 


q 


r 


s 


t 


JjJ V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


g, 
h 
i 

J 
k 


S 
R 
Q 
P 





1 


i 


,i 


k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


g 
h 


^ 


J 


,i 


k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


g 


"? 


K 


k 
1 


1 


m 


n 


o 


P 


q 


p 


s 


t 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


g 


h 


^ 


m 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


ft 


h 


i 


,i 


:3i 


M 
N 


m 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


g 


h 


i 


J 


k 


1 


N 
M 


^ 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


g 


h 


i 


j 


k 


1 


m 





P 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


ft 


h 


i 


J 


k 


1 


m 


n 


L 

k 

J 




P 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


S 


h 


1 


J 


k 


1 


m 


n 


o 
P 


s; 


Q 


q 


r 


s 


t 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


g 


h 


i 


,i 


k 


1 


m 


n 


o 


^ 


R 


r 


s 


t 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


S 


h 


i 


,j 


k 


1 


m 


n 


o 


P 


q 


1 


< 


S 


s 


t 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


S 


h 


i 


i 


k 


1 


m 


n 


o 


P 


q 


r 


H 


Si 


T 


t 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


S 


h 


i 


,j 


k 


1 


m 


n 


o 


P 


q 


r 


s 


G 

F" 




U 


u 


V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


g 


h 


i 


,i 


k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


~j 


V 


V 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


fi 


h 


i 


,i 


k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


u 


E 




w 


w 


X 


y 


z 


a 


b 


c 


d 


e 


f 


g 


h 


i 


j 


k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


u 


V 


D 




X 


X 


y 


z 


a 


b 


c 


d 


e 


f 


g 


h 


i 


,i 


k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 







Y 


y 


z 


a 


b 


c 


d 


e 


f 


9, 


h 


i 


,j 


k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


u 


V 


w 


X 


B 




z 


z 


a 


b 


c 


d 


e 


f 




h 


i 




k 


1 


m 


n 


o 


P 


q 


r 


s 


t 


U V 


w 


X 


y 


A 




z 


Y 


X 


WV 


U 


T 


SiR 


Q 


PiO 


N ML 


K 


J 


1 


HG 


F 


EiDC 


B 


A 



"> 



Letters of plain text. Reserve 

at R, and at the top of the column in which the k occurs 
we find T, which is the first letter of the plain word. 
Proceeding in like mamier with the others, we obtain: 

k i k 
REX 

=t e n 



156 CRYPTOGEAPHY 



NUMBEE OF POSSIBLE COMBINATIONS. 

With — Number of Combinations. 

3 letters^ ...._. 6 

4 „ - - - . . -24 



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 



BEITISH SUENAMES. 

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, EE, ING(S), LEY, FOED, STON(E), MAN, 
OCK, BY, HAM, LAND, ICK, ETT, WELL, FIELD, 
KIN(S), LOW(S), WOOD, MOEE, BUEN, HUEST, 
WOETII, DALE, SHAW, BOEOUGH, STOWE, 
EIGHT, WAY, STEAD. 

* 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 lER. 

38 end in ARD. 

21 end in EAU. 

19 end in AUD. 

15 each end in LET, LLE. 

13 each end in AND, NET. 

12 each end in AUX, ERE, ERT. 

11 each end in LOT, RON, SON. 

10 each end in OUX, TTE, ULT. 

9 each end in CHE, GER. 

8 end in LIN. 

7 each end in RIN, UET. 



CRYPTOGRAPHIC MATERIAL. 

One or two manuals of cryptography (see Bibhography). 
Dictionaries in several languages. 
English, French, German, etc., rhyming dictionaries. 
Two Saint Cyr slides (see p. 110). 
Two graduated rules, one numbered from 1 to 50, the 
other from 51 to 100. 
Paper ruled in squares. 
Slates ruled in squares. 
Tracing paper. 
Coloured pencils. 

A T-square (useful for consulting Vigent're's Table). 
A ready reckoner for rapidly calculating proportions. 



158 CKYPTOGEAPHY 

A 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 E's, 7 A's, 7 I's, 7 N's, 
etc. The use of counters from time to time rests the 
eyes, and enables one to try a number of combinations 
more rapidly than could be done with pen or pencil. 



BIBLIOGEAPHY 

WORKS RECOamENDED. 

La cryptographie devoilee, by C. F. Vesin de Romanini. Paris, 1857. 

Die Geheimschrifien und die Dechijfrir-Kunst, by F. W. Kasiski 
Berlin, 1863. 

Handbuch der Kryptographie, by Ed. B. Fleissner von Wostrowitz. 
Vienna, 1881. 

La cryptographie militaire ou des chijfres usites en temps de guerre, 
by Aug. Kerckhoffs. Paris, 1883. 

La cryptographie et ses applications a Vart militaire, by H. Josse. 
Paris, 1885. 

Essai sur les methodes de dechiffrement, by P. Valeric. Paris, 1893. 

La cryptographie pratique, by A. de Grandpre. Paris, 1905. 



PART IV 

THE PLAYFAIR CIPHER SYSTEM, ETC. 

By Translator 

It is surprising that there is no work in cryptography 
in Enghsh, 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 
States, for a book or manual on this fascinating subject, 
but without success. M. Langie defines cryptography 
as the art of communicating thoughts secretly, and this 
certainly appears to me to be a better definition than the 
stereotyped one of " secret \vriting," as it is perfectly 
obvious that writing is not the only medium by which 
secret communication can be effected. It used to be 
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 
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 froin one post to another. 

Cryptography, in some form or other, has a surprisingly 
great bearing upon the everyday events of ordinary life. 
You -will find upon your handkerchief a mysterious little 
symbol which to you is meaningless, but which in your 
laundry indicates your name and address, and many a 
fugitive criminal has been brought to justice by such a 

159 



160 



CEYPTOGEAPHY 



slender clue as a laundry mark upon some garment which 
he has had to leave behind at the scene of his crime. 

Eacegoers may have noticed individuals standing on 
the top of a cab or on some coign of vantage, semaphoring 
energetically, in the intervals between the races. This 
process is called " tictacking," and I understand that the 
men operating it have various codes, which they do their 
best to keep secret, for transmitting prices from Tatter- 
sails' ring to the outside bookmakers. 

The marking of cards by sharpers is a form of crypto- 
graphy in which an amomit of ingenuity is exliibited 
worthy of a better cause. Playing cards, ostensibly for 
conjuring pm'poses, are sold publicly in the United States, 
each card being marked in such a manner that any one 
with a little practice can as readily read the card from the 
back as its face. One of the commonest forms of indicat- 
ing the face of a card on its back is in the form of a clock, 
as shown in the following diagram: 





Fig. 1. 



FifJ. 2. 



Fig. 1 shows the back of the card. At the four corners 
of Fig. 1 will be observed four small rings, an enlarge- 
ment 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 PLAYFAIR CIPHER SYSTEM, ETC. K'.l 

an aco, and so on until eleven o'clock, which indicates 
a Jack, while twelve o'clock denotes a Queen, and a dot 
in the centre spot represents a King. The suit is indi- 
cated 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 si)ados. Even after you are in- 
formed that the cards are marked, it is surprisingly 
difficult for the uninitiated to detect these marks and, in 
l)assing, I cannot refrain from repeating the advice so 
often given that great care is essential when playing cards 
with strangers ! 

Business men find it necessary to make extensive use 
of some form of secret writing to indicate the price of the 
various goods they sell. This is most commonly done 
by means of some easily remembered word of ten 
dilferent letters which are used to denote figures. For 
instance, the word " bankruptcy " might be employed, 
B to represent 1, A 2, and so on. An extra letter is 
generally used in the case of a figure which is repeated. 
For instance, 11 shillings would be expressed as BX/- 
or BZ/-. Certain firms use a variety of signs such as 
circles, rectangles, triangles, etc., but obviously it would 
not be a difficult task for anyone to break this code. 
It is sometimes most important for manufactiu'ers and 
merchants that their prices should be strictly secret, 
and I 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 wull 
at the 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 

11 



162 CEYPTOGEAPHY 

adopted by spies in the various countries involved in the 
war to transmit information secretly. It would be im- 
possible 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 conj&ned to 
naval and military requirements or to diplomatic offices. 
Many important financial houses are w^ell aware that 
uui^rincipled rivals would stick at nothing in order to 
be able to tap their messages and to break their cipher. 
It is obvious that the transfer by telegraph of large sums 
of money must be done with very great care and secrecy, 
and all the great banks employ elaborate methods to 
insure that their secrets will not fall into dishonest hands. 

It is well known that tramps in all countries have their 
methods of 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. 
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 
to be detected. I do not altogether agree with this, as 
it frequently happens that it is of vital importance to the 
recipient of a secret message that he should be certain 
that his are the only eyes to see this message. When a 
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 loiter may be opened and photo- 



THE PLAYFAIR CIPHEE SYSTEM, ETC. 1G3 

graphed and tho cryptocrram solved witliout the rightful 
recipient being aware of I he fact, and he fondly imagines 
that he alone is tho 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 tho 
only eyes to read the concealed message. There are 
many varieties of these sympathetic inks, the most 
widely known being milk, orange or lemon juice, dilute 
sulphuric acid, etc., which are all revealed by the apphca- 
*tion of heat. A very simple, although not very well- 
known 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 lines of a letter or inside the wrapper of a 
newspaper, as may be arranged. The recipient then 
pours a little ink on the arranged section and promptly 
rubs it ott" with water. The scratching of the pen with 
moisture has removed the glaze on the paper in such a 
way that it is invisible even w^ith a powerful magnifying 
glass; but when ink is deposited on the surface it attacks 
those portions where the glaze has been removed, thereby 
making the words written stand out quite distinctly, 
while the surrounding glazed surface is merely slightly 
soiled by the application of the ink. 

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 
revealed by the application of heat. A solution of nitric 
acid may also be used, and this can only be revealed by 
wetting the paper. After it dries, however, it again 
becomes invisible, so that the above-mentioned objection 
renders it unsuitable. A 2 per cent, solution of acid of 



164 CEYPTOGKAPHY 

lead when used as an ink is quite invisible, and can only 
be made readable by immersion in hydrogen sulphide 
gas. This would appear to be a comparatively safe ink 
to use, but in the course of some experiments I made in 
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 
disappear again by slightly moistening it with peroxide 
of hvdrogen. Writing done with a solution of potassium 
ferrocyanide can only be made visible by the application 
of some ferric salt. Probably the safest form of secret 
ink is 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 
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 si)ot. I would recommend that anyone who 
is desirous of using this form of secret ink should first 
make some experiments to see that they get the exact 
strength of solution requu-ed and the right quality of 
paper. There is a counterpart to sympathetic inks in 
the form of disappearing inks which, hov/ever, are of very 
little practical value. The best known of these is a 
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. 105 

M. Langie states that the cipher invented by Francis 
Bacon is extremely easy to break, but I am of the opinion 
that this system used with certain variations could be 
made extremely difficult. A well-known Baconian en- 
thusiast, Colonel Eabyan of Chicago, believes that Bacon 
interpolated a great deal of secret information into the 
manuscripts of his various works by means of his cipher. 
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 I 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 cr^^ptography should make themselves 
conversant with the Morse Code or Alphabet. The 
following table 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 
to three dots, and a space between two elements or 
signals in a letter is equal in length to one dot. The 
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 if the foregoing rules are not strictly 
adhered to. Bad spacing will convert A into E T or N 
into T E, and a slight examination of the following table 
\\ill show many other telegraphic identicals, which all 
have to be borne in mind when endeavouring to decii)her 
a cryptogram which has been telegraphed. 



166 



CEYPTOGEAPHY 



International Morse Code Signals. 



Letters. 



Figures. 



A . - 


N - . 


1 . 


B -. . . 





2 . , 


C - . -. 


P .--. 


3 . . . -- 


I) -. . 


Q --.- 


4 .... - 


E . 


E .-. 


5 


F . . - . 


S ... 


6 - ... . 


G --. 


T - 


7 -- . . . 


H . . . . 


U . .- 


8 . . 


I . . 


V . . .- 


9 . 


J . 


w .-- 





K - . - 


X -. .- 




L . -. . 


Y - .-- 




M -- 


Z --. . 





M. Langie has omitted to give any reference to the 
" Playfair " cipher, which has been extensively used for 
miUtary purposes. This cipher is one of the substitution 
variety, and may be operated with one or more key- words, 
which may be located in the cipher square by pre-arrange- 
nient. This square is divided into twenty-five separate 
compartments, and the letter J is always represented by I. 

Suppose the key-word to be " BANKEUPTCY "—to 
be distributed between the first and fourth lines of the 
square. Fig. 1 will show their position: 



b 


a 


n 


k 


r 














u 


P 


t 


c 


y 









Fta. 1. 



THE PIjAYFAIR CIPHEH SYSTEM, ETC. 1G7 

The other letters of the alphabet 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: 




The rules fur enciphering by the Playfair method are 
as follows : 

1. Divide the plain text of the message to be sent into 
groups of two letters. When there is an odd number 
of letters, say 21, complete the last odd letter by the 
addition of X or Z. 

2. In the case of repeated letters, such as EE or 
LL, divide these by inserting X or Z. 

3. Each pair of letters in the square, wdien filled in by 
agreement, must be either in the same vertical line, the 
same horizontal line, or at the diagonally opposite corners 
of a rectangle formed by the smaller squares within the 
whole square. 

4. When the pair of letters to be enciphered occurs in 
a vertical colunm, substitute letters immediately below 
the letter of the plain text. When this letter is at the 
foot of a colunm, substitute for it the letter at the top 
of any colunm — e.g., to encipher U S, which occurs in 



1G8 CEYPTOGBAPHY 

the first vertical column of Fig. 2, the substitution would 
be SB. 

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 
left of that column — e.g., to encipher T Y, which are in 
the fourth horizontal column of Fig. 2, the substituted 
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 line with it — e.g., on Fig. 2, A C would 
be enciphered K P, D would be represented by G I, 
and E L by A Q. 

7. The enciphered message may be written in groups 
of three, five, or eight letters, and the letter agreed upon 
for the purpose of dividing repeated letters may be used 
to till up a group. Hhould more than one letter be re- 
quired to complete a group, as many letters as are required 
may be taken from a prearranged word, such as STOP, 
FINISH, etc. 

To decipher a message sent in this code, you simply 
divide the letters into pairs and reverse the writing 
according to the preceding rules. The decipherer 
should never neglect to write down the X or Z, as the 
case may be, when used as a divisory letter. This simple 
precaution will often save a lot of time in decoding a 
lengthy message. 

The following example will show the method of en- 
ci))h('ring in accordance with the foregoing rules, with 
" baiikru])tcy " iis the key-word, distributed in the first 
and lourth ccjlunnis of the square. 



THE PLAYFAm CIPHEB SYSTEM, ETC. 1G9 

Suppose the message required to be enciphered to be: 
" You may expect reUef in three days," and that X 
is to be used to divide du})hcated hitters. Divide the 
phiin text into groujjs of two letters each, as follows: 

YO UM AY EX PE CT KE LI EF IN TH 
C(^ TI EP GV VL YC AH ML EG MB YE 
HE XE DA YS 
AH VG EB UZ 

and then underneath each group of substituted letters, 
as shown above. This being done, the message should 
be divided into groups of live, and sent as follows : 

CQTIE PGYVL 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 to be used to divide repeated letters, and 
STOP to complete an unfinished group. The resulting 
cipher will be as follows: 

VDQVPEQV QPICERAL LOWQFELB PFQPWULC 

The Playfair system is one of the best forms of cipher- 
ing, for several reasons. It is very simple to commit to 
memory, after which all that is necessary for the sender 
and receiver to bear in mind are the key- words or sequence 
of key-words, and the position they are to occupy in 
the square. The key may consist of one or more words, 
provided they contain different letters — e.g., EAIE 
CUSTOM or A JUKY OE MEN. The key-word may be 
changed in alternate words or at certain intervals. Eor 
instance, a message might be sent using as key-words 
BANKEUPTCY in the first and liftli colunms of the 



170 CEYPTOGRAPHY 

square, CUMBERLAND in the second and fourth, and 
TICHBOURNE in the first and third, and many other 
variations will readily suggest themselves to the student. 
The message may also be sent in groups of three, four, 
five, or eight letters, so there are abundant opportunities 
for throwing obstacles in the way of a decipherer who is 
not in possession of the keys. 

An indication of the difiiculties 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 V H F H and G, and in the second example the four 
E's are transcribed D L F and F. 

That these difficulties 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, 
who has kindly given me the following account of 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 LicuL-Commander W. W. Smiih, 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 PLAYFAIU CiniEU SYSTEM, ETC. 171 

(c) When divided into groups of two letters each, no 
group contains a repetition of the same .letter, as NN 
or EE. 

(d) Unless the message is very short there will be 
recurrence of groups, and this recurrence will, in general, 
follow the order of normal frequency of digraphs. 

(c) In messages of length, unless encipherment has 
been made from several squares of different keys, whole 
words are hkely to recur in the form of repeated 
groups. 

In the solution of the Playfair, we need not consider 
the normal frequency of individual Enghsh letters, 
E, T, 0, A, N, etc. We are, however, very much con- 
cerned with the normal frequency of pairs or digraphs: 
th, 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 
page 167, it is seen that the cipher letters YE represent th of 
plain text, and so long as this same key is in use, tli plain 
can only be represented by YE in cipher. Likewise, on is 
always MK, and an NK. (Note: Throughout this dis- 
cussion we mil represent cipher letters by capitals and 
])lain text by small letters. Also, in referring to equations 
as above, we may designate the letters of the equation 
as 1, 2, 3, and 4. Thus 1, 2=3, 4, where 1 and 2 are 
letters of the cipher group, and 3, 4 are plain text 
letters.) 

Case 1. Letters at opposite corners of a rectangle: 

If YE=th 

then EY==ht 

TH=vf 

HT-:fy 



172 CEYPTOGEAPHY 

Case 2. — Two letters in the same line or column: 
In line 1 of the square, 

NK=an 
and KX=na 

But AN is not equal to nk, and NA is not equal to kn, 
and reciprocity is only partial. 
We may therefore note Rule I. as follows: 
Eule I. — Eegardless of the position of the letters in the 
square, if the assumption is made that 1, 2=3, 4, 
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 may be assumed: 

3, 4 (cipher) =1, 2 (plain), 
and 4, 3 (cipher) =2, 1 (plain). 

Now, as each letter of a line or column can be com- 
bined wdth but four other letters of its own line, and with 
four letters of its own column, and as each letter when 
employed at the corner of a rectangle can be combined 
with each of 16 letters to form a group, it would appear 
that Case 1 is twice as probable as Case 2. 

Now, in the square, note that : 

NK=an FA=en 

FK=gn FL=em 

MK=on also FP=et 

TK=cn FV=ew 

WK=xn FG=ef 

From this it is seen Ihat of the twenty- four equations 
that can bo formed when each letter of the square is 
employed cither as the initial or linal letter of the group, 



THE PLAYFAIR CIPHER SYSTEM, ETC. 173 

five will indicate a repetition of a corresponding letter of 
plain text. 

Hence, Rule II. — -After it has been determined, in the 
equation 1, 2=3, 4, that, say, FA=en, there is a proba- 
bility of one in five that any other grou]) beginning with 
¥ indicates e-, and that any group ending in A indi- 
cates -w. 

After such combinations as er, or, and en have been 
assumed or determined, the above rule may be of 
use in discovering additional digraphs and partial 
words. 

Rule III. — -In the equation 1, 2=3, 4, 1 can never 
equal 3, and 2 can never equal 4. Thus, KR could not 
possibly indicate er, or AY=an. This rule is of use in 
eliminating possible equations when the cipher is under 
investigation. 

Rule IV. — In the equation 1, 2=3, 4, if 1 and 4 are 
identical, the letters are all in the same line or column 
and in the relative order 3, 4, 2, — . In the square 

shown, NK=an, and the order is ANK , which is 

equivalent to -ANK-, or — ANK. This is a very useful 
rule. 

Rule V. — If 2=3, the letters of the equation are in the 
same line or column, and in the relative order 2, 1, — 4, 
which is equivalent to 4, 2, 1, --, or - 4, 2, 1, -. Thus it is 
seen that in the square, BR=rfc, and the order is RB--K, 
which is the same as KRB — , or B — KR. 

Some cryptographers claim that from an analysis of 
the cipher message, the letters which are found to com- 
bine in groups with the greatest variety of other letters 
will very likely be the letters of the key-word. This 
may be of some value provided the key were contained 
in the first two lines of the square, an4 if the key-letters 



174 



CRYPTOGKAPHY 



could positively be eliminated it would be possible to 
solve the message and reconstruct the square. Unfor- 
tunately, these letters cannot be positively eliminated, 
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. 

Rule VI. — Analyse the message for group recurrences. 
Select the groups of greatest recurrence and assume them 
to be high-frequency digraphs. Substitute the assumed 
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 
hastening the transfation of the cipher. Let us now 
take up the solution of the actual test message given 
below: 



APBNOH 

ANNSXR 

KBSNHL 

ENPFVB 

OAEYSC 

FRTACS 

VSBXOH 

RUXUUO 

NESCSl) 

RARGCI 



RNAORA 

OUUADT 

DYPYHS 

NVOBNX 

SOKTDN 

HGQROA 

RNRARB 

QENSXU 

VNNSGR 

YCNIVK 



GIOREB 

BN.OARP 

NYSIQC 

GNXROU 

KNXDTE 

BHNSRS 

TNTHXU 

GRGBTR 

ARGBIZ 

DADYPY 



WQGRUD 

NIYERB 

WRCSEQ 

OAFLIG 

CIRNOM 

ECOROT 

QNFLRN 

CNORLC 

RAREHN 

RXXUUY 



In working with the cipher, disregard thi^ above group- 
ing and rearrange the message in pairs of letters. 

We will first analyse the al)ove message by drawing 
a chart of gronj) recurrences (Fig. 1.) 



THE PLAYFAIR CIPHER SYSTEM, ETC. 
First Letters of Pairs. 



Hi 



» 

H 
H 

Q 

O 

o 

1 

i 


/A 
B 
C 
D 
E 
F 
G 
H 
I 

K 
L 
M 
N 

P 
Q 
R 
S 
T 
U 
V 

w 

X 

Y 

J 


A 

2 
1 
1 

1 

— 

A 


B 

1 
3 

B 


C 

2 
1 

2 
C 


D 

1 

1 

1 

2 
D 


E 

1 
1 

1 

1 
£ 


F 

2 

1 

1 

F 


G 
2 

1 
1 

G 


H 

1 
1 

1 
1 

H 


I 
1 

1 

1 

1 

I 


K 

1 

1 
1 

K 


L 
L 


M 
M 


N 

1 
1 

7 

1 

1 
1 

N 




i 
1 

2 

1 

3 

1 
2 




P 
1 

2 
P 


Q 

1 

1 

2 

Q 


R 
4 
2 
1 

1 

1 

4 

1 

1 

1 

1 

R 


S 

1 
1 

1 

1 

1 

s" 


T 
1 

1 
1 

1 

T 


U 
1 

1 

1 

_ 



1 

U 


V 

2 

1 
1 

1 
V 


w 

1 
1 

w 


X 

1 

2 
4 

X 


Y 

1 
1 

Y 


Z 

1 

_ 

z 


A 
B 
C 
D 
E 
F 
G 
H 
I 
K 
L 
M 
N 

P 

"q 

R 
S 
T 
U 
V 
W 
X 
Y 
Z 



Fig. 1, 



176 CEYPTOGEAPHY 

This chart shows that the following groups occur in the 
cipher four times each: 

OA EN GE 

EA NS XU 

No group occurs more than four times. This is un- 
usual. The groups BN, CI, and OE occur three times 
each. All of the above groups must represent common 
pairs of letters in English text. It is w^ell known that 
the order of frequency of common pairs of letters is as 
follows (from a count of '2,000 semi-military letters) : 



th 


50 


er 


40 


on 


39 


an 


38 


re 


36 


he 


33 


in 


31 


ed 


30 


nd 


30 


ha 


26 



at 


25 


en 


25 


es 


25 


of 


25 


or 


25 


nt 


24 


ea 


22 


ti 


22 


to 


22 


it 


20 



St 


20 


io 


18 


le 


18 


is 


17 


ou 


17 


ar 


16 


as 


16 


de 


16 


rt 


16 


ve 


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 commcfbly occurring groups of the ci])her, 
OA, EA, EN, NS, GE," and XU, wo nolo 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 an equation cannot equal 
the fourth. Nor can IIA or EN symbolise re or rt (Rule 
III.). Thus wo can cliiiiinalo a fow of the possible 



THE PLAYFAIE CIPHER. SYSTEM, ETC. 177 

meanings of the groups. ]jut any one of the six groups 
may represent th. The most common four-leiter group in 
Enghsh is known to bo THEE, while such groups as 
INTH, ENED, TION, etc.. 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 
the group occm'S, thus XU4. 

Note that in lines 1 and 7, the follow^ing combination 
recurs: BN3 OH2 RN4, and that in lines 2 and 4 we have 
XRg OU2, and in lines 3 and 10 we have DYg PYg. But 
it will be useless to attempt to guess the meaning of the 
two last mentioned groups, as the individual groups are 
not frequently used in the cipher, and occur only with 
each other. Thus XROU and DYPY may indicate 
four unusual letters that recur in the cipher, or they may 
be caused by the insertion of nulls between repeated 
letters. To guess their meaning w^ould 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 
begins and ends with unusual words. The repeated 
groups BNo OHo RN4, however, present opportunities. 

Also, note the combinations in the cipher of our most 
common groups: 

RN4RA4 
GR4OA4 
NS4GR4 
NS4XU4 

We must first assume each of these groups in turn to 
be ther, which is the most common four-letter combina- 
tion in English text. Failing to estabhsh this relation, 
other combinations of common digraphs will be assumed. 

NS4 may be fh, but XUj is probably not er, as it occurs. 

12 



178 CRYPTOGRAPHY 

at the end of the cipher, in XUUY. However, it must 
be considered as a possibility. NS4 6R4 cannot be tlier, 
as GR cannot be er (Rule III.). 

Suppose GR4 OA4 to be ther. This is an excellent 
assumption, as reciprocals of both groups occur in the 
message, and if GR=f/i and OX=er, RG=/if and 
AO=re. 

Now ht is an uncommon digraph, and can occur only 
in w-it-ht-he or in a combination of gM, as in " eight " or 
" thought." Pursuing this assumption, we assume in 
line 10, for the combination HX-RA-RG, -w-it-ht. 
Then RA=?i, and AR=/i. Substitute these values 
throughout the message and we get for AO-RA in line 1, 
reit, and for GR AR in hne 9, thti. These do not appear 
promising, and after carrjang the investigation farther 
it was decided to abandon the original assumption that 
GR4 0A4=//!er. 

Note: In all work of this nature false assumptions 
will be made, but as the investigation proceeds they will 
eventually be proved false. In this case a great many 
false starts were made due to unusual conditions, and 
ther was abandoned in favour of such combinations as 
Hon, ered, etc., before the investigation resulted in success. 
For the sake of brevity, these steps will be omitted. 

Assume RN4RA4 to be ther (line 7). 
Then RN=th 
and RA=er 

NR=ht 

AR=re. 

Unfortunately, we have no NR in the cipher, and 
but one AR. Make these substitutions throughout 
(Fig. 2). 



THE PLAYPAIE CIPHEK SYSTEM, ETC. 179 

AP BN OH KN AO RA GI OR EB WQ GR UD 

th er he 

AN NS XU OU UA DT BN OA RP NI YE RB 
r- 

KB SN HL DY PY HS NY SI QC WR CS FQ 

EN PF VB NV OB NX GN XR OU OA FL IG 
r- h- 

OA EY SC SO KT DN KN XD TF CI RN OM 

th 

FR TA CS HQ GR OA BH NS RS EC OR OT 
-r he 

VS BN OH RN RA RB IN TH' XU QN FL RN 
th er e- -r th 

RU XU UO QE NS XU GR GB TR CN OR RC 

he -t 

QN ZS SD VN NS GR AR GB IZ RA RE HN 

he re er e- 

RA RG CI YC NU VK DA DY PY RX XU UY 
er eh 

First stop not undcrscorod. Substitutions of '2n(l 
and 3rd steps, pages ISO and 1^1, are underscored — . 
(Message divided into pairs.) 

Fig. 2. 

There is also a possibility (Rule I.) that TH=r?t, and 
RE=ar, but these occur only once each, and the reci- 
procals ER and HT do not occur at all. 

Now, if RA=er, we know, from Rule IV., that the 
three letters are in the same line or column of the square 
and in the order ERA — . 

As RN is assumed to be ///, the partial square^ must be, 



180 CEYPTOGEAPHY 

depending whether this equation is formed from a rect- 
angle or a hne: 

(1) E E A - T 

H - - N 

or. (2) T 

E E A - - 
H 

N 

In either case, we have estabhshed the fact that H is 
somewhere in the same column with B, regardless of the 
position of T and N, and we have for certain: 

(3) 

E E A - - 

H 

Now note otJier groups containing E. We know from 
Eule II. that if EA=er, there is a chance of one in five 
that EB=e-, and EE=e- and GE=-e. 

Now GE is used four times in the cipher, and as there 
is a good chance that it may be -e, let us assume it to be 
the highest digraph ending in e in tho frequency list. 
This digraph is he. 

Then GE=he 
EG=eh 

Make these substitutions in the cipher (Fig. 2), at the 
same time adding to the square. 

(4) - 

E E A - - 
G H - 

Going back to (1) and (2), if we combine (2) with (4) 
wo have: 



THE PLAYFAIR CIPHEE SYSTEM, ETC. 181 

(5) - T - 

E E A 

G H 

- N - 

From the above we may obtain some partial equations, 
sucli as TH=-r, TA=-r, XN=-r, EN=r-, etc. Sub- 
stitute; these values in Fig. 2. 

As we have as yet no substitutions that can be extended, 
we must attempt to find common digraphs to be sub- 
stituted for more of the connnon groups. In line D we 
have the groups NS - GE - AE or NS - he-re. NS is 
used four times, and must be a common pair. Turning 
back to the list of normal frequency of digraphs, we find 
that 0)1 is the group of highest frequency next to tic 
and er. Assume NS to be on. 

If this assumption is correct, then by Eule IV., since 
1=4, these three letters are in the same line, and in the 
order ONS — . It is evident from our partial square 
that ONS — cannot form a column. We may therefore 
buikl up the square as follows: 

(6) T 

E E A - - 

G H - - - 

N S - - 

Note: In the partially completed square the hori- 
zontal lines are definitely fixed as shown, for EA=er, 
and NS=o?i. We also know that in the column TEHN, 
E follows T and N follows H, for EN=^/i. But we are not 
certain that EH= Tr, as there is a fifth letter to be placed 
in the cohunn, and it can only come below E or below N. 

Now substitute in Fig. B, on for NS and no for SN, 
and also the new groups that we get from the square 
above, namely: 



182 CEYPTOGKAPHY 

NG-oh 8K==na SE =oa SH=n- 

GN-=ho OH=ng EN=ro HS=-n 

ES-=an HO=gn NE=or GS=-o 

AN=rs OA=se KO=en SG=o- 
NA=sr AO=es OE=ne 
ES -=ao ON=-o 

AP BN OH EN AO EA GI OE EE WQ GE UD 

ri ng th es er -o ne -t he 

AN NS XE OU UA DT BN OA EP NI YE EB 

rs on sq su ri se o- e- 

KB SN HL BY PY HS NY SI QC WE CS ¥Q 

no -n n- rt 

EN PE VB NV OB NX GN XE OU OA EL IG 

ro ho se o- 

OA EY SC SO KT BN KN XD TE CI EN OM 

se n- th 

FE TA CS HQ GE OA BH NS ES EC OE OT 
-r se r- on 

VS BN OH EN EA EB IN TH XU QN EL EN 

ri th er e- -o nw th 

EU XU UO QE NS XU GE GB TE CN OE EC 

on he nt ne 

QN ZS SB VN NS GE AE GB IZ EA EE HN 

on he re er e- wh 

EA EG CI YC NU YK BA BY PY EX XU UY 
er eh 

Step on page 1S3 underhned — . 

Step on page 184 undorhncd twice =^. 

Fig. :}. 



THE PLAYFAIli CIPHER SYSTEM, ETC. 183 

In the last two lines of Fig. 3 we now have erle- HN er eh. 
HN cannot be tli, and if taken from square it would 
be rh. This would spell nothing, and as the word " where " 
suggests itself, we may assume HN=w/i, which would 
give us an addition to the square as follows : 

(7) 



E 


E A 




W 


G 


H - 





N S 



Substitute TR=wf, TR=niv, and R^^ivh in Fig. 3, 
also ^Nli=rt. 

Line 1 now shows a first word evidently ending in 
ing, but it cannot be the word " having " as the square 
does not permit lia to be represented by AP. 

BN occurs three times, and must be a common group. 
Ti w^as tried, but was soon found not to be satisfactory. 
After a few similar suppositions ri was decided upon, 
and BN=n substituted throughout. 

If BN=n, we see from the square that i cannot follow 
GH in line 3, and that B and i must be in the same 
separate column as below: 

(8) T 

ERA- B 

W - 
G H ~ 
N S - I 

It is evident now that linos 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. 



184 CEYPTOGEAPHY 

Substitute in Fig. 3 the equations taken from the new 
square: RB^r-; IG=o-; IN=-o; GI=-o; NI=o-; 
SI=w-. 

Note the first Hne. 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 Gl=om and I(J=mo. 

We have but one of our six commonly-used groups 
left undeciphered — ^namely, XU4. In line 7 we have 
-0 nw XU. Eeturning to the partial square, X and Z, 
being unconnnon letters, are probably not in the key. 
If not in the key, X probably follows W, and if U likewise 
follows T in the first line, XU=" as," a common digraph, 
and wXU=" was." 

Also, as E 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 




E 


E A - 
W X 


B 


G 


H K - 


M 





N 8 - 


I 



Many new groups now result. Fill in these substitu- 
tions in Fig. 4 and we see that more new groups are 
evident to complete the words. Such groups are under- 
scored in Fig. 4. 

In line 8, CN is evidently io, and this gives us the 
letter lacking in lino 5 of ihe square. If CN=io, line 5 
must read ONBIC, and the colujini BiM 1 must adjoin 
UAXKS. 



THE PLAYFAIE CIPHEE SYSTEM, ETC. 185 

AP BN OH RN AO EA GI Oil EB WQ GK UD 
du ri ng th es er mo no -t he jxi 

AN NS XE OU UA DT JJN OA EP NI YE EB 
rs on \va sq su rp ri se dt o- ea 

KB SN HL DY PY H8 NY SI QC WE CS EQ 
ma no -n n- rt 

EN PE VB NV OB NX GN XE OU OA EB IG 
ro io s.w ho wa sq se om 

OA EY SC SO KT UN KX XI) TE 01 EN OM 
se n- hu hs th ig 

FE TA CS HQ GE OA BH NS ES EC OE OT 
ur in g.t he se rm on. an ne nq 

VS BN OH EN EA EB IN TH XU QN EL EN 
ri ng th er ea so nw as to th 

EU XU UO QE NS XU GE GB TE CN OE EC 
at as qs oq on as he me nt io ne d- 

QN ZS SD YN NS GE AE GB IZ EA EE HN 
to on he re er e- \vh 

EA EG CI YC NU VK DA BY PY EX XU UY 
er e.h st ' a.w as 

Fig. 4. 

Thus, building up the square from the pairs under- 
scored in Eig. 4, we have: 

Q T U - P 

E E A B B 

- W X - - 

G H K ]\I - 

N S I C 

It is apparent thai eohinin 5 of tlie square should be 
transposed to left of column 1 (this does not at all 
affect the equations obtained), and the fifth and second 
lines are seen to ;yield the key- word " considerab." 



186 CKYPTOGEAPHY 

P Q T U - 

D E E A B 

- - W X - 

- G H K M 
C N S I 

The letters now remaining to be placed are V, ¥, Z, L, 
and Y. 

Carrying this process farther, we come to YE=eq, and 
may safely fill in the square: 

P Q T U V 

D E K A B 

L Y W X Z 

F G H K 1^1 

C N S I 

We now complete the solution (Fig. 5). The key- word 
is " 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 : 

F G H K M 

C N S I 

or P Q T U V 

D E K A B 

L Y W X Z 

All of the above squares are equivalent, and it is 
probable that the key was used in the second and fourth 
lines as in the last square. 

It is now interesting to note how the Rules held true 
in a message not prepared by the writer. All of the 
commonly used groups of the cipher are listed below, with 



c 





N S 


I 


p 


Q 


T U 


V 


D 


E 


R A 


B 


L 


Y 


W X 


Z 


F 


G 


H K 


I\I 



THE PLAYFAIK CIPHEK SYSTEM, ETC. 187 

their equivalent digraphs. It is seen that the digraphs 
follow normal froquoncy very closely, and that in no 
case does a repeated group indicate an uncommon pair: 

llN4=th BN3=ri 

KA4=er Clg =is 

NS4 =on 0E3=:ne 

GK4=he 
XU4 =as 

AP BN OH RN AO RA GI OR EB WQ GR UD 

\)u. ri ng til es er mo* ne* da* yt he pa 

AN NS XR OU UA DT BN OA RP NX YE RB 

rs on wa sq su rp ri se dt os eq ea 

KB SN HL DY PY H8 NY SI QC WR CS FQ 

ma no fw el ql kn ow ns po rt in gp 

EN PE VB NV OB NX GN XR OU OA EL IG 

ro cl iv it ie sw ho wa sq se Id om 

OA EY SC SO KT 1)N KN XI) TF CI RN OM 

se qe ni no hu re hs la ph is th ig 

FR TA CS HQ GR OA BH NS RS EC OR OT 

hd ur in gt he se rm on an do ne nq 

VS BN OH RN RA RB IN TH XU QN FL RN 

ui ri ng th er ea so nw as to Id th 

RU XU UO QE NS XU GR GB TR CN OR RC 

at as qs oq on as He me nt io ne dn 

QN ZS SD VN NS GR AR GB IZ RA RE HN 

to xi ca ti on He re me mb er ed wh 

RA RG CI YC NU VK DA DY PY RX XU UY 

er eh is lo st uiu br el ql aw as qx 

* Evidently a group left out, meant for 'One clay." 

Q was used as a null. 
Fig. 5. 



1S8 



CEYPTOGEAPHY 



One claim made in favom' of the Playfair is that common 
pairs, such as tli, er, on, etc., will not be enciphered in 
their normal frequency, due to their chance of being split 
up when the message is divided into two-letter groups 
prior to enciphering. This claim is well based, but when 
th is split up it will probably yield another common 
digraph, he. Furthermore, even though a large percen- 
tage of these digraphs is split, their frequency in the 
cipher is still relatively great. It is interesting to examine 
the following table prepared from the above problem: 



Digraph. 


Times Occurring 
in Message. 


Times Represented 
in Cipher. 


Times Split 


th 


6 


4 


2 


er 


6 


4 


2 


on 


8 


4 


4 


he 


7 


4 


3 


as 


7 


4 


3 


ri 


4 


'2 


2 


re 


4 


1 


3 


is 


3 


'2 


1 


in 


5 


2 


3 



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 frequentl}^ happens, 
however, that important hrms liave 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 PLAYFAIR CIPHER SYSTEM, ETC. ISO 

code word — so many forward or back as may be arranged. 
This system, however, wouhl present very little difficulty 
to anyone who wished to break the message, and a much 
safer method is to cipher the numbers wliicli 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 : 

UZYNRAEHOB 

A more elaborate system of ciphering is to have a 
series of tables for the conversion of the figures with 
letters, comjjiled as follows: 

00=AB 
01=AC 
02=AD 

03=AE 
and so on, up to 99. 

It is obvious that any pair of letters in the alphabet 
may be used to represent any pair of figures, so that the 
variations of this form of ciphering run into millions, 
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 
on this system is practically unbreakable. 



190 TEST MESSAGE IN CIPHEE 

BNHCfYKZJ ELKOCWVD ARBGEXD 

KLVEDEST ABLNYVSG VIWCOCR 

DZRKICXF YTANBECB ZBEWGBH 

GFIDAVTCO WCWDADXE RYNQGXY 

YVPJEGFA BLXUBNQC EZLGAPH 

HICKDROQ UZXHVEFG DQYNZNY* 

SBKEQHJK ORLVDICM GIWFYKN^ 

EBIZSOBQ SDATWRAL HVISTVA^ 

CAXSDPED RSBIXFEG YPTAHQP 

OPHYXHEJ ACBNUMOF VRDIRTL 

SOBLJCJET EDRXLAVB ICSZFED 

FXEZEKLJ UTKYBSEN DNJOCKL 

TJEHSQLO WVFRUGUT REHXZBI 

ABEDZFYT SJUJGADV XFOBGHI 

SJETVTYD VLHYZBUK BRAVENH 

CRDYFEZF ICJTFOVS ATLFINC 

SFTYZCBE DLTEMRGU KCSJAWE 

DRZLUSIX PEQCFAGF JEFGUDZ: 

EMNYCHCR IVTBIZBG MEQMYWL. 

NTEFTRYD XPMAQBML ONBPANC 

RJPARXEW DCYVSOSB MLIVCHU 

HLOCKT. JE RTCLOCHZ AMUKRXL 

Q U D G A K N S C Y L A B R F G J ]:> S E X V Y 

CUMKJOKS LASPBIQZ EVLJMAN 

W L F S () N F R (! I C L M D U (J H S A K A V Z 

T L X I N V H M E P Z L F U R S Q U G K I li N 

U G N Z J. Y D F A I. K I) F N K I S G L U R Z U 

O T R I W V A V Y C; A N T E I) P X J) U Z V N I 

R A B'M V IT f; r. V Y R T> (! E X V R O F U G A L 

V U K L X U I) I F Q R B O N G D F E G O R X E 



I 



() T D R I Q U 

I] Z X F N O K 

x I) w r M c (J 

1. W () C MTU 

LI M J D T 

B L N E K U 

E W Q V A V L 

N E Z V C P A 

X R N Y F L Z 

(! F H r; u V a 

R U L M X J 

7 F C G I F X 

T U D K V Y N 

X C I B S G E 

E P F L I C H 

; X T A 8 W E L 

" A R H I T V 

A X G P B Y N 

» D M B Y C X Z 

w U K R X J E P 

K L A Z HZ Y (} 

n S H N Y S P Y 

V M Y G H K B E 

Z C H L B Y V W 

'* I T V H C Y Z 

5 A T G L S E T 

in 1) E S M A C K 

3 D S Y F L O N 

I E R H B Y B 

N'T V J A X V W Y 



CRYPTOGRAPHY 

G 11 X A Z A J G 
K W I T I B M D 
K () W T F X E N 
Z E C W T G O S 
L B V C U S U.D 

V J G S I T V A 
D U N C D L X 
VFEBNJCU 
EQBUNCSI 
RS EWLFAJ 
ZTOBATRL 
Z C Z L E G L M 
PSOJLCDE 
N C D L I W D X 
CLATCNDU 
TZUKBPFA 
SRCWICHB 
CTREZCVJ 
D A R V K I N M 
L U C K M R I G 

VI R V D U X C 
NO HUVEQR 
Z B U T S U C R 
ONDEJ S QF 
N E F T O G K 
C L Y R T O r L 
VYOFLNOL 
LCXODI GF 
N H X D L J U 



191 

BUPBRfil D 
A S J Y Z N F E 
ZYHDRAJ X 
J HPUDFND 
QAWMWXEK 
HBARDTUZ 
D F B Y R G P G 
F G X A P M Q 
J K S U L V Q F 
AVJ BOKYM 
U B A C R T I M 
Y Z F I G L A S 
S F S W T K M 
OTLDNAJ T 
FRLJ YDAJ 
R Z L J O Q U V 
YRTEDOQU 
A K J M B V Q 
CMEZLTVY 
F E Z J F JI T 
J F A N F E S G 
J T O F T S A R 
L J A F Y B L S 
I L M E R J V U 
B M U R T L I S 
H D I V S A Q JI 
BUVLBNI G 
HROBETZC 
G II V D A N C D 



192 CEYPTOGKAPHY 

Conclusion 

I should hesitate to say that a cryptogram can 1 
invented that will defy solution, provided it is of reasr | 
able length and is not so im^olved and intricate as 
make its use inexpedient. For practical purposes 
cipher should be upon some system which can easily tl 
committed to memory, and it should not involve an 
great expenditure of time in coding messages. Th 
cryptogram on page 190 has been ciphered in accordanc 
with these rules on what I believe to be a novel principle 
and will, I venture to think, require a great deal of paini 
and patience to solve. In the event of failure to solvi 
the cipher, at a reasonable time after publication thil 
method of ciphering and the solution may be obtainecl 
from the Marconi International Code Co., Marconi House? 
London, W.C. 2, or the Marconi International Codf 
Corporation, 2230, Park Kow Building, New York, or 
receipt of a stamped and addressed envelope. 



rnlNTKIi IN ORKAT nniTATN nv 
lill.l.lSd ANU «11NS, LTD., OUlLliUlUIJ AND KSIIKR 



v>r 



>.>r 



>S^ 



.^ 



UNIVERSITY OF CALIFORNIA LIBRARY 

Los Angeles 

This book is DUE on the last date stamped below. 



INTER 
DIS 



OiSC 

OCT 

0' 



AUG ^ 1^^^ 

4 M my 1 .? \m 



"X- 



•0CT17 Jail 

MAY 1 8 1994 






41584 




%oi\miQ'^ 







^OFCAtlF0% ^ 



37 




"^Ji^iaoNvsoi^ 



>^OFCALIF0/i»^ ^. 



AWEUNIVERS/a 






^lOSANCElfj"^ 



v^lOSANCafj>^ 




13DNVS01^ %a3AINn3\\V 



^ 



^ 1 ir^ ^ 




^ 



^ "l^"'Yl58"06339'8b38 ^ 



i 



^''^ Urv'MfliTTv- 
UC SOUTHERN REGIONAL LIBRARY FACILITY 



\m\ 




my 



^<!/0dnv3J0'^ <rji]ONv.soi^ "^AaiAiNoawv 



001 091 151 



J^^0 



^.OFCAIIFO/?^ 




A\^EUNIVERS'/A ^,^lOSANCFlfXy> 




^ommii'^ '^mwiov^ ^^/smMnMS^ 



> 



o 



vvlOSANCElfr,. 




%i]3AINn3\\V 



^MIIBRARY^/: ^MIBRARYQr 





'^<i/0JllV3J0'^ ^<!/(hinV3J0'^ 



^^ 






o 
130NVS01^ 



s^vlOSANCElfX/ 




"^AaJAINfl-^WV 



^OFCAIIFO/?^ ^OFCAIIFO% 




^^^AavaaiH"^ ^^^Aavaaii-i^ 



>- 



k^ ^l-llBRARYQ^ 




\WE UNIVER^//, 




.>;lOSANCElfj> 
o 




^ojuvj-jo^" <rii30Nvsm^ ^/^aaAiNOJUV 



^ ^OFCAlIFOff^ ^^WEUNIVER% ^lOSANCElfj^ 



4^^