0% found this document useful (0 votes)
106 views52 pages

The SIGABA ECM Cipher Machine A Beautiful Idea3

The document discusses how Edward Hebern invented the first American electromechanical rotor cipher machine in 1912 while in prison, and how the U.S. Navy tested his machines. It also describes how cryptologist William Friedman was able to break codes encrypted using Hebern's machine by exploiting its regular rotor movement. This led Friedman to realize electromechanical cipher machines needed a way to generate pseudorandom rotor movement to be cryptographically secure.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
106 views52 pages

The SIGABA ECM Cipher Machine A Beautiful Idea3

The document discusses how Edward Hebern invented the first American electromechanical rotor cipher machine in 1912 while in prison, and how the U.S. Navy tested his machines. It also describes how cryptologist William Friedman was able to break codes encrypted using Hebern's machine by exploiting its regular rotor movement. This led Friedman to realize electromechanical cipher machines needed a way to generate pseudorandom rotor movement to be cryptographically secure.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 52

Center for Cryptologic History

The SIGABA / ECM II Cipher Machine :


“A Beautiful Idea”
This publication presents a historical perspective for informational and educational purposes, is the
result of independent research, and does not necessarily reflect a position of NSA/CSS or any other U.S.
government entity.
This publication is distributed free by the National Security Agency.
If you would like additional copies, please e-mail [email protected] or write to:
Center for Cryptologic History
National Security Agency
9800 Savage Road, Suite 6886
Fort George G. Meade, MD 20755

Timothy Mucklow was a senior historian on the staff of the Center for Cryptologic History fol-
lowing a long career in information assurance at NSA and with the Air Force. After a decade in aca-
demia, he served as a military historian at the wing, division, and major command levels and produced
a series of monographs and articles on such topics as information assurance, telecommunications and
computers, and national defense issues, and has offered IA seminars at military facilities around the
world. He received his Ph.D. in 1982. He retired from the CCH in 2013.

Acknowledgments. The Center for Cryptologic History is grateful to Dr. Craig Bauer, professor
of mathematics at York College of Pennsylvania and editor-in-chief of the journal Cryptologia, for his
assistance with this project, and to LeeAnn Tallman for earlier research.
The SIGABA / ECM II Cipher Machine:
“A Beautiful Idea”

Timothy J. Mucklow

National Security Agency


Center for Cryptologic History

2015
Contents

Introduction ....................................................................................... 1
And Hebern Saw the Wheel .................................................................... 2
Deus ex Machine ................................................................................. 5
Rowlett’s Epiphany .............................................................................. 7
Iacta Alea Est. [The Die Is Cast.] ........................................................... 10
Enter the Navy ..................................................................................12
A Disagreeable but Rewarding Surprise ....................................................14
SIGABA Is Built ...................................................................................16
An Impenetrable Machine .....................................................................22
The Big Machine That Did .....................................................................25
Appendix A: Technical Analysis of SIGABA’s Key Space ..................................29
Appendix B: The Mechanics of SIGABA......................................................32
Notes ..............................................................................................38
Selected Bibliography ..........................................................................42
SIGABA / ECM II Cipher Machine
Introduction

D uring World War I, both the Americans and the


Europeans had used manual code and cipher systems
the “design of the system should not require secrecy,
and compromise of the system should not inconve-
to secure their critical information. With the advent nience the correspondents.”1 His argument was that
of the machine age, however, it was only a matter of neither the elements of an encryption algorithm nor
time before cryptography became mechanized. Two the workings of an encryption machine should con-
decades later, as the world embroiled itself in another stitute the basis of a cipher’s security. Instead, an eas-
major conflict, the use of cipher machines became ily changeable key should be the critical component.
commonplace, and the old craft of cryptology at long
In the 1930s, the U.S. Army cryptologist Wil-
last became mathematically based. With the dawn of
liam Friedman and his assistant Frank Rowlett drew
this new cryptologic age, a worldwide scramble for
on this simple precept to conceive a cipher machine
even stronger encryption began.
that was easy to use, simple to rekey, and ostensibly
Well before the First World War, much work had impossible to break. Then, in a fit of collaboration
been done to formalize the cryptographic discipline. with the Navy, as unprecedented as it was peculiar,
One of those who made a major contribution to the two services went on to perfect and jointly field
this process was the 19th-century polyglot Auguste their device. To the Army it was known as SIGABA,
Kerckhoffs, who, besides teaching languages, dabbled to the Navy, ECM (Electric Cipher Machine) II.
in cryptography. In 1883 he penned a two-essay series, (The technical aspects of SIGABA’s operation can
La Cryptographie Militaire, for the Journal des Sciences be found in Appendixes A and B.) Not only was
Militaire, which earned him a niche in cryptographic SIGABA the most secure cipher machine of World
history. Kerckhoffs in his writings had articulated six War II, but it went on to provide yeoman service for
basic principles of cryptography, but he is most com- decades thereafter. The story of its development is
monly remembered for his second, which states that improbable. Its impact was incalculable.

1 1
And Hebern Saw the Wheel

M r. Edward Hebern of Madera, California, is


generally credited with inventing the first American
posed, but using a set of cipher wheels whose
wiring was to be unknown to Friedman …
electromechanical rotor cipher machine.2 Accord- Friedman was successful in his attack on the
ing to lore, he came up with the idea around 1912 system.4
while serving a term in the state penitentiary for
With little effort, Friedman was able to decipher
horse thievery. If there were any connection between
every Hebern-enciphered message sent his way. To
breaking rocks and his cryptographic epiphany,
do this, Friedman, who had already married crypt-
Hebern never offered one. By the end of 1915 he
analysis and mathematics, used a statistical approach
had built a working model, filed a patent, and was
along with a “divide-and-conquer” strategy to break
knocking on the doors of potential buyers.3 It was
the Hebern machine. As Friedman saw it, there were
not long before the U.S. Navy began showing inter-
only two unknowns, the cipher wheel (with twenty-
est in Hebern’s device, and in the years that followed
six characters) and the rest of the machine. Since the
the Navy would buy and test a number of his suc-
cipher wheel stepped regularly and predictably, the
cessive machines. As something of a historical twist,
rest of the machine could be considered a constant.
one of Hebern’s early employees, Agnes (Meyer)
It was then a simple matter of sequentially analyzing
Driscoll, went on to become a luminary of American
the individual components.
cryptology in her own right.
Friedman realized from experience that the regu-
Half a century after Hebern set out to peddle
lar advancement of the Hebern cipher rotors, like all
his original device, the Army cryptologist Frank
machines of that era, posed an intrinsic cryptologic
Rowlett acknowledged in his memoirs that the
vulnerability that was mathematically exploitable. He
Hebern cipher machine was an important inven-
conjectured that to make a cryptographically sound
tion in pre-World War II American cryptography.
machine, one would need a countermeasure to address
Rowlett, nevertheless, went on to add,
predictable rotor movements. But just how such pseu-
Navy cryptographers asked [William] dorandom movement could be achieved consumed
Friedman for his assessment of the security Friedman’s thoughts for years. Friedman’s solution,
it afforded … [T]he Navy [supplied] a set when it finally came to him around 1926, was concep-
of messages prepared exactly in accordance tually simple and the technology proven.5 It occurred to
with the procedures which they had pro- him that paper tape similar to that used by telegraphers

2
And Hebern Saw the Wheel

Edward Hebern and his electromechanical rotor cipher machine


(courtesy of Ralph Simpson, CipherMachines.com)

could also be used to dictate the move-


ments of the cipher wheels. Friedman
reasoned that it would be relatively easy to
generate long sequences of random five-
unit keys and that the tape itself was eas-
ily replaceable and relatively inexpensive.
Holes punched in a tape would permit
feeler contacts to turn an electrical cur-
rent on or off, causing a cipher machine’s
rotors to step. Thus, with each key stroke,
randomly placed holes in a five-group tape
would produce an apparent random step-
ping for one or more of the cipher rotors.
The cipher text is printed on a small tape.
The decryption process simply works in
reverse. Key tape could be manufactured
in secure facilities, distributed to the
appropriate recipients, and inserted into
the machine according to rigidly imposed, William Friedman with AT&T printing telegraph, 1920 (courtesy
pre-established schedules.6 of the George C. Marshall Foundation, Lexington, Virginia)

3
SIGABA / ECM II Cipher Machine

The crypto rotor wheels William Friedman envisioned for his device
were flattened cylinders with an alphabet around the circumference.
One face of the cylinder had twenty-six spring-loaded copper pins
protruding from it; the other face had twenty-six flush copper contacts.
Inside each cylinder was a wire maze connecting the electrical contacts
on one side to the pins on the other.
Thus, an electrical impulse beginning with, say, the letter A on one
side might connect to H on the other side and so on around the wheel in
random fashion. Several cylinders serially juxtaposed on a spindle side-
by-side could further scramble impulses. (Image courtesy the online
Crypto Museum, www.cryptomuseum.com)

Five-group punched hole paper telegraphic tape

4
Deus ex Machine

O n April 23, 1932, Friedman revealed his idea of


randomly stepping rotors to his three junior cryptan-
not—recognize the unique contribution to cryptol-
ogy presented by randomly stepping rotors. There
alysts: Frank Rowlett, Solomon Kullback, and Abra- remain today bundles of letters between Friedman
ham Sinkov.7 This foursome at the time constituted and the patent office, often in duplicate and tripli-
the bulk of the U.S. Army’s cryptanalytic organiza- cate copies, meticulously detailing the merits of his
tion, the Signal Intelligence Service (SIS). At the respective patents years after the initial patent appli-
meeting Friedman enthusiastically explained that cations were filed.10 In spite of his repeated petitions,
“the inherent weakness of all such devices” (where patent office officials sitting only blocks away from
the devices were cipher machines that determined Friedman obdurately continued to presume that any
their own rotor stepping pattern) is that “the keying cipher rotor machine was just an unimaginative deri-
mechanism is a part of the device itself.” 8 He then vation of the Hebern machine.
went on to draw for them several rough sketches of a
In a letter from Friedman to the U.S. Patent
cipher machine with a “key tape transmitter” that, in
Office dated December 5, 1934, regarding Patent
conformity with Kerckhoffs’ second principle, would
Serial No. 682,096, he stated, “the cipher key here
separate keying material from the cipher machine.
serves as the physical embodiment of the ‘keying
His colleagues were duly impressed with his vision-
principle’ … and that its sole purpose is to serve as
ary countermeasure; certainly Friedman had good
the controlling element in effecting the displace-
reason to be pleased with his own inventiveness.
ments of the cipher wheels in a variable manner.
William Friedman, over the course of his life, Contrast this situation with that in Hebern.”11 In
demonstrated a penchant for collecting cryptograph- the Hebern cipher machine, the complexity is the
ically related patents. Ultimately he was granted machine’s wiring, not the initial setting of the cipher
thirty of them.9 Like inventors before and since, he wheels or the placement of the rotors. Friedman
discovered that dealing with the U.S. Patent Office continued, “Referring now only to the mechanism
was neither a pleasant nor an alacritous proposition. for displacing the cipher wheels, in Hebern there is
In fact Friedman’s own experiences with govern- embodied no such thing as a cipher key which cor-
ment red tape proved to be an ongoing source of responds to a ‘keying principle which is variable in
exasperation that consumed his time and sapped his character’ because the mechanism for displacing the
energies. As many times as he explained it to them, cipher wheels is absolutely fixed.”12 In the Hebern
patent office officials simply could not—or would cipher machine, the cipher wheels advanced in the

5 5
SIGABA / ECM II Cipher Machine

these movements are regular or


periodic in character, and con-
trolled by ratchet mechanisms
internal to the device itself. In the
present invention [Friedman’s],
these movements are controlled
by the cipher key transmit-
ter in an aperiodic manner …
It will be recognized that the
Hebern structure has the inher-
ent weakness of all such devices
where the keying mechanism is
a part of the device itself. Peri-
odical recurrence of movements
is a natural characteristic of all
such mechanisms and the pre-
dictable factor thus introduced
defeats the essential purpose.14
As the months turned into years,
it probably occurred to Friedman that
his purposes might have better served
by addressing his letters to Santa Claus.
From the time Friedman had iden-
tified the weakness in contemporary
rotor cipher machines and had concep-
tualized an effective countermeasure,
the subsequent refinements to his pseu-
dorandom stepping rotors were entirely
evolutionary. As Friedman’s design for
his new devices became more sophis-
Friedman’s M-134 converter. Note plugboard ticated, they manifested themselves
at upper left.
under a succession of names, includ-
ing M-134, M-134A, M-134-T1, and
M-134-T2. When Friedman’s M-134
same regular manner for every message. Friedman’s
Converter went into production, it “consisted of the
complaint was in response to a letter from the Patent
chassis, the machine itself, an assembly of wheels,”
Office, posted June 6, 1934, which states that Fried-
a plugboard, and a punched tape transmitter (read-
man’s claims were “rejected on Heburn [sic].”13
er).15 Drawing on a pool of ten different rotor wheels,
In still another letter from Friedman to the Pat- the M-134 used five wheels at a time, but it was
ent Office, dated April 3, 1934, he again contended, the paper key tape that provided movement to the
respective rotors and gave it cryptographic strength.
In Hebern the movements or displacements
of the code wheels are purely mechanical;

6
Rowlett’s Epiphany

F rank Rowlett, who had received several years of


cryptanalytic training under Friedman, was given
with vulnerabilities. Too, Rowlett really doubted
the overall practicality of Friedman’s key tape trans-
the unenviable task of creating the quantities of key mitter. He surmised that in stressful environments
tape necessary to drive Friedman’s electric cipher operators might easily tear or misuse key tapes. Just
machines. Rowlett recorded why and how in 1934 as disturbing to him was the prospect of their suc-
he had been assigned the job of making key tape. He cumbing to the temptation to reuse the key tapes
said, “Now Friedman’s thought was that we ought and their starting places. Rowlett was also concerned
to get cracking on the construction of these tapes so about numerous issues associated with Friedman’s
that when the machines came off the assembly line” machine. Not only were Friedman’s key tapes lengthy,
they could be sent out to field stations right away. but they required all parties on a particular network
Unfortunately, the machine Friedman had designed to maintain the same large inventories of keying
for the process of generating key tapes was beset with material. The distribution of and the accountability
problems, and Rowlett spent much of his time trying for tape alone would entail untold manpower and
to make it work properly. Rowlett continued with a resources, not to mention acquiring secure facilities
touch of irony: “I got stuck with the job of making for its storage.
the tapes because I had a little bit more practice in
Besides his reservations about Friedman’s key
mechanical things than the rest of the group and I
tape transmitter mechanism, Rowlett just didn’t like
think they were smarter than I because they didn’t
making the paper key tapes for the M-134 and would
let it be known.”16 Being the dedicated mathemati-
do almost anything to return to the more excit-
cian that he was, Rowlett set out to make the key
ing work of breaking into Japanese ciphers. Given
as random as possible. Since theory and practice are
Rowlett’s affinity for machine solutions, he sought
frequently at odds with one another, he discovered
a means of automating this impossible assignment.
that producing key tape proved to be a more difficult
With a burst of inspiration, he decided that it might
undertaking than Friedman had envisioned.
be possible to use one set of rotor wheels to gener-
Rowlett struggled with the problem of key gen- ate the random stepping movement for the M-134
eration until he became completely frustrated. It cipher rotor wheels. If he could find a way to do that,
was obvious to him that producing loops of paper the entire process would become much easier and
tape with holes randomly punched throughout their undoubtedly more secure. He would later admit in
lengths was too labor intensive and probably rife an interview,

7 7
SIGABA / ECM II Cipher Machine

Key tape for the M-134 converter

I don’t know that I ever was confronted with instead of the tapes. Well, I thought this
a more hopeless task than making these was a pretty powerful thing and I just was so
[Friedman’s] devices work and do what enthusiastic about it because it looked like
was needed and I soon became desperate. I was getting out of this impossible task, I
It didn’t take more than a month for me to went to tell Friedman about it.18
realize that I was fighting a real losing battle
Armed with enthusiastic conviction, Rowlett
here, and as you are apt to do in the case
approached his boss about scrapping the whole key
of where necessity becomes very evident you
tape transmitter concept and moving on to some-
try to figure out some better way of doing
thing more cryptographically sound. After permit-
things and I was dreaming about how rotors
ting Rowlett only briefly to present his discovery,
could be made to do other [things] and
Friedman would hear no more. He categorically
decipher [messages].17
refused to believe it possible to create a machine
As the drudgery of developing key tapes became capable of generating its own random stepping.19
more tedious, the solution came to him. Friedman, who was every bit as hard-headed as his
nemeses over at the Patent Office, remained wed-
I thought it would be a helluva good idea if
ded to using replaceable paper key tape. After all, his
we replaced these key tapes with a second set
salient argument for pursuing a patent for his key
of rotors which in effect would generate five
tape transmitter—which caused the cipher rotors
screens of impulses equivalent to the holes
to advance irregularly—was that any machine that
and no holes in the five levels of the tape and
determined its own stepping would be in violation
use this assembly of five additional rotors
of Kerckhoffs’ second principle and, therefore, vul-

8
Rowlett’s Epiphany

nerable to cryptanalysis. To what degree the loss of a


potential patent clouded the respectable Friedman’s
thinking is not known. Frank Rowlett, besides being
Frank Rowlett, besides being intellectually bril-
liant, was a tenacious man and not the sort to aban- intellectually brilliant, was a
don a cause in which he genuinely believed. Time
and again, he requested of his employer the oppor- tenacious man and not the
tunity to fully explain his concept. Each attempt
made Friedman all the more formal—to the point of sort to abandon a cause in
being brusque—and their strained working relation-
ship slowly approached a breaking point. Years later, which he genuinely believed.
Rowlett recounted,
This went on for, oh, I guess six, eight, ten
months, I confronted Friedman … Friedman
was still reluctant and finally out of a real fit
of desperation I said, either Mr. Friedman I
don’t know what I’m talking about or I know
what I’m talking about and you don’t under-
stand me … I think we’ve got to clear this up
because I’m going to have to quit that job. I
just can’t meet your requirements.20
Then being on something of a roll, Rowlett
went on to inform the unsmiling Friedman that
he would have to go above Friedman’s head to the
Army’s chief signal officer if Friedman did not give
his proposal a fair hearing. This of course was a
risky gamble for a family man in the depths of the
Great Depression. Friedman was now faced with
an ultimatum and had to weigh the respective mer-
its of retaining a capable and loyal assistant or los-
ing face to his subordinate. Reluctantly, he offered
Rowlett an opportunity to fully lay out his concept.21 Frank Rowlett

9
Iacta Alea Est.*

R owlett knew this would be his sole opportunity to


sell his stepping maze concept and came to the meet-
concept marked a gigantic step in cryptography, and
his prudent acquiescence to Rowlett changed the
ing well prepared. One after the next he responded course of history.
to Friedman’s sometimes insightful, sometimes petty
Surprisingly, once Friedman fully accepted
questions with rock-solid, unambiguous answers.
that he had been wrong, his attitude toward his
At the end of the tension-filled session, Friedman
junior changed entirely, almost. Rowlett went on to
was forced to pause and reconsider some of his own
recollect,
cherished cryptographic perceptions. Not only did
Rowlett’s invention run counter to Friedman’s previ- . . . I had never before found him so friend-
ously published opinions against the incorporation ly and so agreeable to work with. He still
of keying logic into a cipher machine, but it made retained his “boss-employee” attitude, but I
Friedman’s beloved pet project, the M-134, obsolete. could see that as he reached a more compre-
Following Rowlett’s presentation, there was a long hensive grasp of the principles I had discov-
and pregnant silence, after which Rowlett was coolly ered, he was accepting me as a professional
dismissed from the room. Ominously, Rowlett heard cryptanalyst rather than as a student.
nothing from his boss for the rest of the day. As the
Before the day was out, Friedman told Rowlett,
afternoon slowly wore on, Rowlett mentally cleaned
“I want you to start immediately on drafting pat-
out his desk and began to consider his prospects for
ent specifications, and I will work directly with you
future employment.22
in developing these new principles into their most
Years later Rowlett recalled the incident: “Well, advantageous form.” The odious work of key tape
next morning he [Friedman] came in, eyes shining, generation was put on hold for the time being.24 Pat-
just all excited and he says, we’re going to do this.” ents come. Patents go. Patents come again.
He called it a “beautiful idea … and he went up to
By summer’s end in 1935 Friedman and Rowlett
see the Chief Signal Officer with stars in his eyes to
had refined the details for a wholly new cipher
try to sell this new idea.” 23 A night of due reflection
machine, the M-134-C, which synthesized the
had forced Friedman to recognize that Rowlett’s
principles of Rowlett’s stepping maze and Fried-
man’s earlier M-134s. They also managed to com-
* The die is cast. plete a draft patent (shown opposite) for what was to

10
Iacta Alea Est.

become SIGABA. (Army cryptographic


equipment of that period was given code
names beginning with SIG [for signals]
followed by randomly chosen letters.)
Unfortunately for the cryptographers, the
nadir of the Depression had been reached,
and the Army’s budget for cryptograph-
ic research and development had been
entirely depleted by Friedman’s M-134
device. Thus, any fielding of SIGABA
was out of the question. Still, the Army
did have a small sum which could be allo-
cated for retrofitting Friedman’s existing
M-134s. This they used to construct a
number of “add-on” devices to be used
with the existing M-134s.25 The so-called
SIGGOO (M-229) component replaced
Friedman’s “key tape transmitter” with a
less robust version of Rowlett’s stepping
maze.
The SIGGOO assembly consisted of
a three-rotor setup in which five of the
keyboard inputs were live, as if someone
had pressed five keys at the same time on
an ENIGMA. These outputs were “gath-
ered up” into five groups as though all the
letters from “A” to “E” were wired together.
In that way the five signals on the input
side would be randomized through the
rotors and come out the far side, ensuring
power in one of five lines. The SIGGOO
rotors, therefore, could be controlled with
a day code, or key, eliminating the need William Friedman and Frank Rowlett’s draft patent of
for paper tape keying material. Because August 1935 for what was to become SIGABA
of cost constraints, though, not all of the
model M-134s could be retrofitted with
Goldberg device. It was, nonetheless, the strongest
SIGGOOs, and large numbers of them continued to
device in the shadowy world of cryptology and easi-
rely on Friedman’s original paper tape transmitters.26
ly surpassed anything the British or the Axis Powers
The M-134 Converter modification with its five had at their command. Since one cannot know what
rotor cipher wheels, the three rotor wheel SIGGOO he doesn’t know, neither Friedman nor Rowlett at
add-on, and a 26-wire plugboard was hardly elegant the time truly appreciated just how advanced their
in appearance and resembled a comic-strip Rube new cipher machine really was.

11
Enter the Navy

A s the torrid summer in 1935 lost its grip on


pre-air conditioned Washington, DC, it was appar-
“had been unsatisfactory [as well and] they now had
a lot of development money but didn’t have any ideas
ent to no one that the sister services were on a cryp- to invest the money in and for goodness sakes did
tographic collision course. While Friedman and the Army have something … any good ideas at all.”29
Rowlett occupied themselves with fine tuning a Friedman, who was initially taken aback by this
patent request for SIGABA, the Navy’s OP-20-G admission, sat up in his chair and brightly suggest-
organization was considering the acquisition of new ed that he might indeed have something to share.
cipher machines. The Navy cryptographers for sev- But first he would have to gain permission from his
eral years had relied on Hebern machines to meet senior leadership.
their communication security needs, but by the mid-
Later that month on the 21st, Lieutenant
1930s they began to doubt the machines’ integrity
Wenger and a colleague, John W. McClaran, met
and balked at a contract to procure more of them.
with Friedman and Rowlett at the Munitions Build-
Quietly, Lieutenant Joseph N. Wenger, acting chief
ing. There, the Army proudly shared its plans for
of the Navy’s cryptologic organization, OP-20-G,
SIGABA with the Navy. Lieutenant Wenger, who
and his colleagues began looking farther afield for
among his colleagues was hardly known for animated
something more secure.27
expression, remained true to form and demonstrat-
Early in October Wenger made the fateful deci- ed little indication that he was impressed. Ten days
sion to seek out Friedman’s advice and walked over later, on Halloween, another OP-20-G complement
the little-used pedestrian bridge that linked the Navy composed of Wenger, McClaran, and John Harper
and Munitions buildings on the Mall in Washing- ambled across the bridge for another meeting with
ton. Clearly, the Navy was in dire straits if it admit- Friedman. Again, Friedman explained Rowlett’s
ted to needing any advice from the Army. As Frank plan for the stepping maze and how it could con-
Rowlett remembered it, “in one of the rare periods trol cipher rotor movements.30 The following day
of consultation between [the] navy organization and a still larger contingent of naval officers assembled
the army organization I think it was Joe [Wenger]… in Friedman’s office for a briefing on the SIGABA
[who] told Friedman that the navy would been [sic] concept. Throughout this last presentation, Fried-
real disappointed with the [existing] contract.”28 man’s guests exhibited an uneasy institutional polite-
Lieutenant Wenger went on to tell Friedman that ness and left with Wenger’s same general lack of
the new machine Hebern was trying to sell the Navy enthusiasm.

12
Enter the Navy

Wenger told Friedman that


“they now had a lot of
development money but
didn’t have any ideas
to invest the money in
and for goodness sakes
did the Army have
something … any
good ideas at all.”
Lieutenant Joseph N. Wenger, acting chief of
the Navy’s cryptologic organization OP-20-G
(pictured later in his career)

Rowlett, after several anxious months, had tated. He knew it was a good idea, a very good idea,
received no feedback from OP-20-G and began to and became all the more convinced not only that it
pester Friedman for word about the Navy’s plans for would work but that it offered secure communica-
SIGABA. Friedman finally tired of these queries tions beyond anything thought possible at the time.
and confessed that Wenger had reported that “there With SIGABA’s future on indefinite hold, Fried-
were certain … operational difficulties and that man and his colleagues returned to solving Japanese
they just weren’t sure the idea would work.” To this, ciphers and producing paper key tapes for those
Rowlett mused, “well, I thought the doggone thing M-134s not fitted with SIGGOOs (since the stan-
would work and maybe they weren’t as smart as they dard M-134s and those with SIGGOOs were not
thought they were.” 31 Rowlett had reason to be irri- compatable, they were used on separate networks).

13
A Disagreeable but
Rewarding Surprise

F our years later, in 1939, with one war spreading


across Europe, another was looming on the Pacific
machine was. Wright responded that he believed
it was Wenger.33
horizon. In anticipation of the latter conflict, Rowlett
Frank Rowlett was no little upset by this news.
and Navy Lieutenant Commander Wesley (Ham)
Not only had he been excluded from SIGABA’s
A. Wright were working on the Japanese PURPLE
development, but he saw the credit for his inven-
analog cipher machine (used to decrypt the Japanese
tion going to another man. At best Ham Wright
diplomatic cipher designated as PURPLE) when a
had been misinformed, and at worst Wenger was
shocking revelation came to light. Rowlett recounted
hoping to pass off the rotor maze as his own idea. In
this incident in his memoirs:
either case, Rowlett decided to get to the bottom of
… we were speculating about what type the matter as quickly as possible. As soon as Wright
of cryptographic mechanism the Japa- left, Rowlett hurried down the hall to tell Major
nese might have used to produce such an William Reeder what had just occurred.34 Reeder,
unconventional substitution system. “May- who, as Friedman’s superior, was the head of the
be they’re using something like what we’re Signal Intelligence Service, needed to be informed
planning to use in our new Navy cipher of this turn of events. That Reeder already seemed
machine”, Wright remarked. “Let me to be aware of the Navy’s work on SIGABA came
explain it to you, and you can give me your as the second disagreeable surprise of the day. Just
opinion of it”. He then started to sketch out how long Reeder had been keeping this to him-
for me the cryptographic circuitry of for [sic] self is unclear. While he had not gone to Rowlett
the new Navy cipher machine.32 with the news, Reeder was perfectly straightfor-
ward about it when asked. In spite of this omis-
Rowlett with an increasingly sickening real-
sion, Rowlett still considered Reeder to be a good
ization recognized the new Navy cipher machine
supervisor and a friend. Several months later, Major
as his own invention. He was particularly stunned
Reeder, in trying to make amends, went to Rowlett
since back in 1935 he had learned through Fried-
with something that he wanted to hear. The Navy
man that the Navy had rejected the idea as being
would soon be taking delivery of a working model
impractical. Without revealing that he had con-
of their new cipher machine and Rowlett was invit-
ceived of the cryptographic principles involved,
ed to help test it.35
Rowlett asked Wright whose idea the cipher

14
A Disagreeable but Rewarding Surprise

On 3 February 1940 the Army’s primary SIS the stepping magnets of both mazes.” 38 He was
cryptologists—Friedman, Rowlett, Sinkov, and also fascinated by the additional set of rotor wheels
Kullback—trooped over to the Navy Building to see that the Navy had included. When he and Fried-
the prototype of the Electric Cipher Machine Mark man shared the stepping maze concept with the
II (SIGABA). Following an informal demonstration Navy, the design incorporated a plugboard similar
of the device, during which Friedman and Rowlett to that on the German ENIGMA machine. At first
could hardly contain their excitement, they were the Army cryptographers did not perceive “… the
finally given the opportunity to put it through its advantage of the extra set of rotors that the Navy
paces. It was a case of love at first sight. Rowlett later had introduced … We [Friedman and Rowlett]
said, “… it was the most beautiful thing to look at preferred the plugboard. The Navy for some rea-
from where I stood and I couldn’t keep my hands off son didn’t like plugboards but this was not a point
it and of course the Navy was delighted to find some- … to quibble about.” 39 With either the plugboard
body as enthusiastic about it as I appeared to be.” 36 or the index rotors, SIGABA represented the pin-
The Navy had exceeded any expectations Rowlett nacle of cryptography, and both the Army and Navy
had entertained, and he wanted to explore every were confident in its ability to resist all assaults.40
facet of the Navy’s engineering triumph. He and
Those who attended the demonstration at
Friedman were all the more heartened by the Navy’s
the Navy Building had every reason to bask in the
willing acknowledgment of the Army’s visionary role
warmth of self-congratulation. It had been a long
behind SIGABA. (The Army designated the combi-
and unlikely path from Friedman’s and Rowlett’s
nation of machines as the M-134-C, which was also
collaboration to the Navy’s manifestation of their
applied to the later, more mature SIGABAs. For the
concept, and Rowlett couldn’t help but reflect with
purpose of this paper, the M-134-C machine will be
a little nostalgia on the old ways of doing business:
referred to by its more popularly used Army short
“… before, you had to have a different set of code-
title, SIGABA, whenever possible.) 37
books and these became onerous, but with the new
Rowlett, besides being taken with the purely secure cipher machines, all the cipher clerk needed
mechanical aspects of the machine, was also quite was a box of rotors and a little pamphlet that told
interested in SIGABA’s wiring scheme. He later him how to use the rotors, which made the concept
said, “I was very curious about the circuitry that … very practical.” 41 As elated as all of them were,
they’d decided on in terms of the association of the none present could then conceive the extent to which
contacts on the in-plates of the control maze with their accomplishment would alter the future.

15
SIGABA Is Built

F riedman and Rowlett’s idea behind SIGABA


had been elegantly simple, but engineering it into
the Army mandated their inclusion for SIGABAs
deployed outside of the continental United States.42
practical reality proved to be a formidable undertak-
The quest to further enhance ECM/SIGABA
ing for the Navy. From late 1936 until January 1941,
would likely have continued for some time had not the
Navy cryptographers and their prime contractor, the
Battle for France, then raging in Europe, convinced
Teletype Corporation, developed a series of proto-
Navy officials that the time to field their new cipher
type machines, each building on the lessons gleaned
machine had arrived.43 As the Navy prepared to let
from its predecessor. Early prototypes were prone to
contracts for full-scale production of SIGABA, the
failure due to environmental conditions such as heat,
cryptologists at OP-20-G and their counterparts in
humidity, and vibration, but as the SIGABA design
the War Department were fully confident that SIG-
matured, these problems and others were resolved.
ABA/ECM II could deliver the degree of security they
Successive models were smaller, lighter, faster, and
sought. This optimism was founded in the combina-
more reliable.
tion of six factors identified by Navy cryptographer
By the time the ECM Mark II (SIGABA) Lieutenant Laurance Safford and paraphrased below:
neared its birthing stage, the resulting device proved
1. SIGABA/ECM II included a sufficient
to be both electromechanically robust and crypto-
number of cipher wheels in the maze to gen-
graphically strong. Variant models could operate
erate an astronomical number of “alphabets”
on 115-volt alternating electricity or 24-volt direct
and starting points. The five-wheel cipher
current/battery power to suit the respective needs
maze of the ECM and Combined Cipher
of the sister services. Production-run SIGABAs
Machine (CCM) provided 11,881,376
achieved an impressive 60 words per minute capa-
alphabets for each arrangement of code
bility. To allay any Army reservations about SIG-
wheels. (The ENIGMA, by contrast, offered
ABA’s reliability under austere conditions, Navy
only 17,576.) 44 [CCM was a specially modi-
engineers offered emergency, manually cranked
fied, less sophisticated SIGABA that was
attachments. These saw almost no actual use in
interoperable with the British mainstay
field operations. Another option was a forty-pound
TYPEX machine.]
thermite bomb for SIGABAs destined to go into
harm’s way. While the Navy steadfastly prohibited 2. The SIGABA/ECM II included ten
these emergency destruction devices aboard ships, reversible cipher wheels in a set. This pro-

16
SIGABA Is Built

A sketch of the wheel control unit, the “control rotor bank,” separate from the cipher unit,
called the “cipher rotor bank.” Note at bottom reads, “Wheel control unit and cipher wheel
unit are identical in design and interchangeable. Wiring to be arranged as indicated.”
It is similar to the diagrams Friedman showed the Navy during 1935.
(Friedman Collection, NSA/CSS accession #47270, box 10, folder 5)

17
SIGABA / ECM II Cipher Machine

SIGABA rotor maze. Note the five small ten-pin wheels that replaced the
Army’s original plugboard.

vided 9,667,680 possible wheel orders, mak- 4. The stepping of the alphabet maze was
ing an adversary’s “trial and error” solutions controlled by an independent source—in this
impractical, if not impossible.45 (The CCM case by both the five index rotor set and the
also offered ten reversible cipher wheels per five control rotor set.47
set.) [There is a common misconception
that SIGABA rotors could rotate either 5. The use of a multiplicity of stepping actions
forward or backward. While the direction of (5,855), dependent solely on the key, instead of
travel for SIGABA rotors was forward only, only one in other cipher machines.48
all fifteen rotors could be flipped in either 6. The replication of code wheel sets—both
direction to maximize their cryptographic “effective” and “reserve”—with prompt
potential.] change of code wheels in case of known com-
3. The use of aperiodic stepping of cipher promise, and a periodic change as an added
wheels, instead of regular or modified- security measure. The inclusion of a back-up
regular stepping motion, precluded all known set of fifteen wheels (five index rotors, five
analytical solutions and prevented “short-cut” control rotors, and five cipher rotors) was
solutions with captured cipher wheels.46 taken to dispel any lingering doubts as to the
absolute security of SIGABA.49

18
SIGABA Is Built

Women assembling SIGABA crypto rotor wheels

With America’s involvement in the war a near ships of the 14th Fleet, with another 100 in stor-
certainty, the Army identified the necessary funds to age awaiting distribution to capital ships and other
assist the Navy in procuring SIGABAs in numbers. units. Twenty-five more were ashore undergoing
The Army’s first lot of 459 machines was fielded in depot-level maintenance. Fortunately, the machines
June 1941. With a due sense of urgency, it distribut- destined for the doomed vessels were still in the
ed these SIGABAs to upper echelon headquarters in warehouse on 7 December 1941 and thus escaped
the CONUS and to selected organizations in Amer- destruction. A hundred of these Pearl Harbor-
ican possessions in the Far East. Naval units in the surplus machines were hastily transferred to the
Atlantic theater were given priority because forces in Army and later used in North Africa. To meet the
the Pacific had already been equipped with the older demand of producing more than fifty devices per
ECM Is.50 Of course some sensitive sites like Cor- month, contracts were let with additional manu-
regidor, Guam, and Pearl Harbor necessitated the facturers. By 1943 5,730 ECM Mark II/SIGABAs
highest degree of security for the signals intelligence were in service and more than 300 per month were
information they processed and thus received the being delivered. The only feature distinguishing the
new cipher machines as well. The Navy was in the Army and the Navy machines was the service-unique
process of outfitting its capital ships of the Pacific designation on the name plates. For the first time
fleet with ECM Mark IIs (SIGABAs) when Pearl in the nation’s history, the Army and Navy enjoyed
Harbor was bombed. Ninety-six had been issued to cryptographic interoperability.51

19
SIGABA / ECM II Cipher Machine

was between six months and two years. In 1943 the


The most security-sensitive average WAVE managed to solder the connections
for fourteen wheels per day. One actually assembled
work of wiring the rotors was a record twenty-two wheels on her shift. The average
for male shipyard electricians had been seven wheels
performed by more than per day before the decision was made to assign them
to other duties. Midway through the war, the Navy
200 WAVES (Women women alone had wired more than 150,000 rotor
wheels. Remarkably, there was not a single configu-
Accepted for Volunteer ration error and only one instance where a wheel had
been mislabeled! 52 Before the rotors were shipped
Emergency Service). WACs to the field, each one underwent more than 2,800
operations before they were certified for use. When
(members of the Women SIGABA production ceased after World War II,
some 10,060 machines were in the inventory along
Army Corps) ... performed the with over 450,000 crypto wheels to support them.
According to the U.S. Army’s account of the
same task for the Army. SIGABA’s development, History of Converter
M-134-C, “each of these experimental and adopted
models is in the direct line of cryptographic devel-
opment which culminated in Converter M-134C.”53
In keeping with Army procedures, the major con-
While trusted civilian contractors such as the
stituent subcomponents of the SIGABA encryp-
Teletype Corporation manufactured and assembled
tion system were given their own alphanumeric
the SIGABA chassis, the all-important crypto rotor
designations.54 Below are examples of SIGABA
wheels remained a military in-house activity. Uni-
crypto-nomenclature.
formed Navy personnel and civilian workers fabri-
cated rotors from stock materials at the Washington SIGABA Converter M-134-C chassis, less the
(DC) Navy Yard (later at the Nebraska Avenue Naval rotors
Station). The most security-sensitive work of actual- SIGKKK Maintenance instructions for
ly wiring the rotors was performed by more than 200 converter M-134-C
WAVES (Women Accepted for Volunteer Emergen-
SIGQZF Crypto-operating instructions for
cy Service). WACs (members of the Women Army
converter M-134-C
Corps) at Arlington Hall performed the same task
for the Army. Since each cipher machine required SIGBRE General instructions for converter
a minimum of two complete sets of fifteen crypto- M-134-C
graphic rotors, these military women were pressed to SIGIVI Cipher basket unit for converter
meet growing production quotas. On occasion when M-134-C
the Army’s requirements overwhelmed its capac- Set of 10 rotors as designated in the
ity to produce them, the Navy lent a willing hand current key list
in providing additional rotors. Because SIGABA/
ECM IIs saw heavy use, the life expectancy of a rotor Keying data in the current key list

20
SIGABA Is Built

As though SIGABA were not already suf- cipher machine rather than humans. Army and Navy
ficiently sound cryptographically, the Army and key generators produced each month’s daily settings
Navy took additional measures to ensure the sys- printed on a single sheet of paper. Thus, the key was
tem’s robustness. Unlike the cipher systems used by easy to transport, easy to use, easy to destroy, and, if
the British and the Axis Powers, SIGABA’s daily necessary, easy to replace.
settings were generated by yet another rotor-based

IBM SIGABA key generator used at headquarters to produce daily key settings 55

21
An Impenetrable Machine

T hroughout World War II, SIGABA offered


America an inestimable advantage. While decrypted
second was the CCM, a SIGABA variant made at
the ECM Repair Shop exclusively for joint Allied
Axis communications channeled a continual flood communications. Because of cost, only 631 of these
of actionable intelligence to U.S. decision makers, models were made. The third, most common and
SIGABA denied the enemy a similar resource. The most cost-effective, was the “X” Adapter manu-
collaborative nature of SIGABA’s design enabled factured by the Teletype Corporation in Chicago.
Army and Navy forces to coordinate their complex, Forty-five hundred of these were sent to depot-level
joint operations in complete secrecy. That the services maintenance facilities for installation. All three of
were operating identical machines turned out to be these options used Typex-configured cipher rotors.
particularly important in the early hours of the war By the war’s end, nearly all U.S. military communi-
when the distribution of machines and rotors was cations facilities could process joint-Allied secure
not yet complete. During critical engagements in the communications traffic.57
Philippines, Java, Australia, and North Africa, Army
In spite of SIGABA’s enormous cryptographic
and Navy crypto-maintenance personnel shared com-
strength, the sister services harbored a slight but
ponents back and forth as necessity demanded. There
healthy anxiety about the enemy’s cryptanalytic
were also emergency situations when the Navy and
capabilities and continued to upgrade SIGABA.
Army used each other’s equipment to ensure the
What if the Japanese or Germans were enjoying
unimpeded flow of secure wartime communications.56
success against SIGABA similar to the Allied suc-
The need for Anglo-American interoperabil- cess against ENIGMA? The notion could not be
ity similarly hastened Allied cooperation in field- blithely dismissed. Thus, during the latter part of
ing cryptographic solutions. Later modifications the war, the Allies sought means to validate the
permitted specialized, less sophisticated ECM integrity of their cryptographic efforts. Intercepted
Mark IIs [SIGABAs] to interoperate with Brit- enemy messages were closely scrutinized, and pris-
ish Typex cipher machines in support of joint U.S./ oners of war were examined for any hint that U.S.
British/Canadian operations. This functionality was cryptography had been compromised. Even before
achieved through three different means. The first of Berlin fell, it had become clear to the Allies that the
these, the ECM Adapter (CSP 1600), was produced Nazis had exploited some lower-level U.S. cipher
at the Washington Naval Yard ECM Repair Shop systems when there had been lapses in COMSEC
where 3,500 were made available for retrofit. The discipline. On the other hand, no evidence emerged

22
An Impenetrable Machine

that the Germans had made any headway against that was an interesting item in itself. They
SIGABA.58 did have some success with the Hagelin.
Deciphered Japanese traffic also indicated that Rowlett continued,
they had not broken into Allied ciphers. An inter-
…we had truckload after truckload of Ger-
cepted JN-A-20 message, dated 24 January 1942,
man cryptographic equipment. … We had
from the naval attaché in Berlin to the Vice Chief
some of the technical reports right up to the
of Naval General Staff Tokyo afforded a comforting
solution of code books and ciphers of other
revelation. In it the naval attaché said he considered
countries, photographic copies of second
“joint Jap[anese]-German cryptanalytical efforts” to
story jobs that they had performed on safes
be “highly satisfactory,” since the “German[s] have
and embassy code rooms …This was gone
exhibited commendable ingenuity and recently
over, carefully evaluated and assessed and a
experienced some success on English Navy systems,”
series of reports produced which you might
but are “encountering difficulty in establishing suc-
find under the term TICOM [Target Intel-
cessful techniques of attack on ‘enemy’ code set-
ligence Committee] Reports.63
up.” 59 In another decrypted JN-A-20 message, the
naval attaché wrote home that he had “…discovered With the return of peace in late 1945, the victo-
that Heine [German] CI [Cryptographic] organi- rious Allies began compiling exhaustive studies on
zation totals 800 persons and is … receiving unsat- Axis technologies and capabilities. One of these, the
isfactory results on American Communications.” Seabourne Report, was a series of technical treatises
He went on to report, “Since last year when Italy drafted by German subject matter experts. Volume
capitulated, English and American countermeasures XIII of this report detailed the Nazis’ successes
have become more vigilant due to interpreting the against Allied cryptographic systems. According
CI situation.” 60 The Japanese in their own internal to interviews with senior officials of the Luftwaffe
communications confessed that they had made no Signal Intelligence Service contained in the report,
real progress against American cipher systems and the Germans revealed they had made no headway
that the Americans were becoming even savvier against the British Typex cipher machine, which
about the security of their cryptographic operations. was greatly inferior to SIGABA. The Luftwaffe
cryptologists interestingly did not address SIGABA
Following V-E Day, Friedman and his asso-
specifically, and their American counterparts were
ciates were anxious to discover just what the Axis
reluctant to press for answers lest they raise unwant-
cryptanalysts had known. According to Rowlett, 61
ed questions from their former enemies. Consider-
We also were very much interested in what ing the immense disparity between SIGABA and
results … have been achieved by the Ger- Typex, it is a certainty that the Nazis made no
mans on the ECM or the SIGABA. And inroads into the “American Big Machine.” 64
we got the answer to that. … they talked to
The official War Diary of the German Signal
the fellow who was in charge of what they
Intelligence Group again seems to validate the find-
called the American Big Machine.62 See,
ings of the other inquiries. Entries made between
they’d identified the Big Machine as the one
February and November 1944 again strongly sug-
jointly used by the Army and Navy, and they
gest that the Axis made no inroads into SIGABA.
couldn’t tell we were using different rotors
While their cryptologists were reading “un-Steck-
or other things because their cryptanalytic
ered” Croatian ENIGMA machines (early com-
understanding was just not at that level; and
mercial model ENIGMA machines without a plug-

23
SIGABA / ECM II Cipher Machine

board or Steckerbrett), British transposition ciphers, exploit only a lower-level Hagelin machine on occa-
Yugoslavian ciphers, and a variety of others, they sion, there was no possibility that they could even
undertook efforts against an American five-letter begin to unravel SIGABA. Dr. Huettenhain, when
system. A translated notation in March 1944 reads, pressed about other Allied cipher machines, said, “I
“A study was begun of a group of 5-letter messag- know of no other type of American machine, but
es from U.S.A. links presumably were enciphered the British Typex is known. It was not broken, and
with a machine of unknown type.” 65 These doubt- so far as we know cannot be solved unless the wheel
lessly were SIGABA messages. With each succes- positions are known.” 69 Dr. Huettenhain continued,
sive month the War Diary reflected that no prog-
We have the ENIGMA [sic] which is simi-
ress was made on the five-letter American cipher
lar to the Typex, and as we believe that the
system. Then in September 1944, with the Allied
ENIGMA cannot be solved, no great effort
forces steadily advancing against the Germans, the
was made to solve Typex. Typex has seven
War Diary includes, “U.S. 5-letter traffic: Work dis-
wheels and we therefore believe it to be
continued as unprofitable at this time.” The Third
more secure than our ENIGMA. ENIG-
Reich’s cryptologists ostensibly decided to focus on
MA when used according to instruction is
ciphers they could exploit.66
unbreakable. It might be broken if a vast
During his detention by the Allies, the Ger- Hollerith complex is used but this is only
man cryptographic mathematician and POW Dr. slightly possible.70
Erich Huettenhain was asked what work was done
After months of interviews, none of the cryp-
on British and American ciphers. He answered that
tologic POWs could offer any information about
most of the successes were diplomatic and “Most
SIGABA. They spoke candidly, even proudly, about
of the American strip cipher was read.”67 When
their successes against British and American ciphers,
he was asked what type of cipher machines were
and why not? Had they broken SIGABA, surely they
broken, he responded, “The main machine bro-
would have been all the more delighted to regale
ken was the American Hagelin which was broken
their captors with their cryptanalytic prowess. Little
only when [an] error occurred.” He continued,
did they suspect that their own cherished ENIGMA
“[c]ommon and regular solutions [were] impossible”
machines had been entirely compromised. Exten-
against the Hagelin machine. Only when a soldier
sive evidence gleaned from the Japanese after the
or sailor grew lazy and neglected to change the daily
war indicated that they had made even less progress
key were the Germans able to leverage their way
against SIGABA than the Germans had.71
into a Hagelin-based cipher.68 If the Germans could

24
The Big Machine That Did

S IGABA, besides merely securing critical infor-


mation, made possible a major shift in the way
tion of intelligence activities, which enabled econo-
mies of scale and a synergy of efforts. Thus, by late
America’s intelligence organizations conducted their 1942 the sister services were able to streamline the
business. Before the summer of 1941, the flow of production and dissemination of processed intel-
information from far-flung radio collections sites ligence information. Whether Safford and Wenger
back to Washington took days to several weeks. had envisioned the new, centralized operational
The sensitive nature of intercepted data, analysis, model when they championed the development of
and reports necessitated that stringent methods be SIGABA back in 1935 is not known, but it is con-
employed to ensure their confidentiality during tran- sistent with Wenger’s overall grand designs for the
sit. Since the United States in the 1920s and 1930s Navy’s radio intelligence function. The Army simi-
did not have absolute faith in its own cryptographic larly benefitted from SIGABA’s implementation at
devices, the services had to rely on physical security its central cryptologic facility at Arlington Hall Sta-
to coordinate their classified communications. This tion near Washington, DC.
meant that messages had to be laboriously typed
SIGABA also supported Allied intelligence
onto onion-skin paper and then forwarded via costly
and military operations. The most striking example
registered air mail to their destinations.72 Given the
of this was its role in the Battle of the Atlantic.
paucity of air service in those days, however, most
From collection platforms in Britain and at sea,
mail was couriered to awaiting naval ships which
radio signals collected from German U-boats were
carried it back stateside. Army elements stationed in
enciphered by American-operated SIGABAs and
the continental United States relied on couriers as
routed chiefly by undersea cable to Washington.
well as the postal system. For those abroad in such
There, the Navy processed the four-rotor ENIG-
places as China, the Philippines, and Panama, the
MA traffic on the cryptoanalytic “bombes” at the
Army process mirrored that of the Navy. As war
Nebraska Avenue station. Hours later, critical,
became imminent, the demand for a rapid, secure
actionable wartime intelligence relating to U-boat
means of transmission became paramount, and
operations was again encrypted on SIGABAs and
SIGABAs were distributed to higher priority cus-
sent back across the Atlantic to Allied forces. While
tomers as quickly as they could be produced.
the United States and the United Kingdom enthu-
One of the concomitant benefits of secure radio siastically shared most of their cryptologic secrets
and telegraph communications was the centraliza- with one another, this did not apply to SIGABA.73

25 25
SIGABA / ECM II Cipher Machine

The United States for various reasons regarded be guarded twenty-four hours a day by armed mili-
any information about the SIGABA machine as so tary personnel.75 This had been the usual practice
sensitive that it did not share any of its principles during the war in overseas locations, and the services
or details with the British. SIGABAs deployed to did not want a relaxation of security with the return
British military facilities were operated and stored of peace.
in secure enclaves to which host-nation personnel
Army and Navy cryptographers, still concerned
were not permitted. Joint U.S.-UK tactical com-
that SIGABA’s principles might be compromised,
munications in the Pacific Theater were passed
undertook to remove SIGABAs from geographi-
along circuits using the Enigma-like British Typex
cal areas where they might be compromised and
machine; during the latter part of the war the
replaced them with SIGRODs.76 SIGROD was a
Allies used the CCM to coordinate joint activities.
transportable, electromechanical, keyboard-oper-
Throughout World War II high-level communica-
ated cryptographic machine capable of encipher-
tions to and from Roosevelt and Churchill passed
ing and deciphering message traffic at the rate of
through SIGABA-based circuits. Messages from
forty to fifty words per minute. It was nearly identi-
Downing Street were forwarded to the American
cal cryptographically to the joint U.S./UK CCM as
embassy in London where they were encrypted and
well as the British Typex machine. Smaller, lighter,
sent to Washington, DC, where they were rendered
and much cheaper to maintain than SIGABA, the
into plaintext and directed to the White House.
five-rotor SIGRODs were capable of processing top
When peace returned in the summer of 1945, secret information and, importantly, if compromised
more than 16 million Americans were wearing the would not disclose the sensitive cryptographic prin-
uniform of their country; two years later those num- ciples embodied in SIGABA.77 SIGABAs were
bers had dwindled to slightly more than 1.5 mil- phased out incrementally and replaced with the less
lion, and the nation was awash in surplus military powerful machines. Nevertheless, SIGABAs con-
materiel. Cryptographic equipment such as M-209 tinued to be used at higher level headquarters for
cipher machines and M-90 devices could be pur- processing the nation’s most closely held secrets and
chased for a nominal sum. Not for sale, however, were stockpiled in heavily secured facilities against
were the 10,060 SIGABAs which had successfully emergency situations when they might be needed
defied the best efforts of the Axis Powers. Postwar again. When the Korean War broke out, SIGABAs
technical analysis of German and Japanese crypto- were used extensively at higher echelons because of
logic capabilities put SIGABA’s principles in a per- their dependability and high degree of security. For
spective that the Americans themselves were only the remainder of the 1950s, the brainchild of Fried-
just then coming to appreciate. Studies suggested man and Rowlett could be found in military higher
that SIGABA was so much more technologically headquarters and critical message centers around
advanced than had been thought that its principles the world.78 During the course of twenty years,
needed even more protection after the war than dur- SIGABA had processed millions of classified mes-
ing it.74 Army and Navy cryptologists were not so sages, contributed to the saving of countless lives,
much concerned that an adversary might be able shortened the agony of war, and helped to advance
to exploit SIGABA if he were privy to its design; the cause of freedom. That it altered the course of
rather, their chief concern was that an enemy could history goes unquestioned.
use its cryptographic principles to protect his own
In 1956 a grateful Congress awarded $100,000
communications. To this end the sister services pro-
to William Friedman for his contribution to SIGA-
mulgated policies that mandated that all SIGABAs
BA and for other cryptologic achievements. Two

26
The Big Machine That Did

The SIGABA/ECM II

years later a similar sum was granted to Laurance cratic parody of itself, the Patent Office on 16 Janu-
Safford for his World War II cryptographic work; it ary 2001 granted a patent for SIGABA—some six-
would be another eight years before Frank Rowlett ty-six years after its conception and thirty-two years
received his reward. Then, in something of a bureau- after William Friedman’s death.

27
SIGABA / ECM II Cipher Machine

28
Appendix A: Technical
Analysis of SIGABA’s Key Space

T he SIGABA machine (U.S. Patent 6175625)


had five cipher rotors, five control rotors, and five
letters/numbers on the rotors could be “on top” for
each of the 10 wheels, giving another 2610 possibili-
index rotors for a total of fifteen rotors. All fifteen ties. Configuring the control and cipher rotors in the
rotors, in banks of five, were encased in a remov- machine in their correct forward/reverse orientations
able module or rotor maze on top of the 100-pound and the correct letter “on top” positions would take
machine. The cipher and control rotors were inter- an exhaust of (210)(10!)(2610) ≈ 5.2 × 1023 for their
changeable and greatly added to the overall security initial settings. SIGABA code clerks and operators
against a brute force attack. Each rotor had twenty- used key charts published monthly to determine the
six contacts on one side wired to twenty-six con- wheels’ forward/reverse orientation, placement posi-
tacts on the other side. These cipher rotors could be tion in the rotor maze, and the “on top” settings of
placed in any order in the five slots. They could also each rotor for each message.
be inserted forward facing or reverse. The initial set-
The third bank of cryptographic wheels con-
tings for the rotor placement were dictated by a code
tained in the rotor maze was the five index rotors
book, and changed daily.
which sat next to the control rotors and was the clos-
Adjacent to the bank of five cipher rotors were est to the keyboard. The index rotors were smaller
the five control rotors. Not only were the cipher than the other rotors and had only ten contacts on
and control rotors interchangeable, but it took only each side. Unlike the other ten rotors, the index rotors
seconds to rearrange them in a new configuration. did not “rotate/advance” during operation but were
In Friedman’s own words, “Wheel control unit and set daily to initial values dictated by the published
cipher wheel unit were identical in design and inter- key chart. While it was physically possible to insert
changeable.” Even if the enemy had captured the these rotors in either the forward or reverse posi-
rotors and knew the wirings of each rotor, they still tion, the index rotors were placed forward through-
had to exhaust over 10! = 3,628,800 different com- out World War II. Like the other rotors, each index
binations of cipher/control rotor locations. Further- rotor could be placed in any of the five slots in the
more, the enemy would have to determine if each index rotor bank. This capability afforded an exhaust
rotor was placed forward or backward. This increased of 5! = 120 to find the correct index rotor permu-
the key space by a factor of 210, since there were two tation. From June 1945 onward, however, the index
choices for each of the ten rotors. Also, any of the 26 rotor configuration did not change, and the setting for

29
SIGABA / ECM II Cipher Machine

each index rotor was provided by the daily key list. An each of the 10 and 10! possibilities for each of the 5.
enemy without knowledge of how it was implemented So the total number of configurations for the system
would have to consider reversals (another factor of 25) would be (26!)10(10!)5 ≈ 2992.8. One need not count
and which value was “on top” (another factor of 105). rotor reversed orientation settings or consider which
Therefore, finding the correct setting for each of the value is “on top,” because such changes are already
fifteen rotors would take (210)(10!)(2610)(5!)(25)(105) ≈ represented by one of the wirings considered. Some
2.0 × 1032 attempts, which is approximately 2107.3 pos- wirings would produce identical encryptions, but the
sible key combinations. Once the index rotors were effect on the already massive key space is minimal.
no longer permuted daily in 1945, this dropped to
Even with this enormous theoretical key space,
(210)(10!)(2610)(25)(105) ≈ 1.7 × 1030, which is 2100.4
the Army and Navy took no chances when they sus-
trial decryptions. This was surprisingly good by the
pected (later proved false) a physical compromise of
modern Advanced Encryption Standard (2128, 2192, or
SIGABA in the last months of the war in Europe. In
2256) and Data Encryption Standard (256) key sizes.
a surge effort, the services produced an entirely new
Of course, as long as the enemy did not recov- set of fifteen rotors for each of the 10,060 machines
er the machine, the wirings in each rotor added to in the inventory. In retrospect, given the state of Axis
the security. For each of the ten cipher and control cryptologic prowess, this was not warranted. All the
rotors there were 26! theoretical ways for them to be same, General Eisenhower, who personally ordered the
wired. A large number of these wirings would have fielding of the new rotors, was not willing to imperil the
been avoided since it does not seem random to have success of military operations or the lives of his troops.
rotors that mapped A  B, B  C, C  D, etc. But Thus, at the war’s end, each SIGABA was equipped
these “nonrandom” wirings should be included in the with two sets of rotors.
theoretical possibilities, since they are valid wiring
The SIGABA’s wiring scheme is provided on
combinations. If an enemy did not know any of the
the next page.
rotors’ wiring, there would be 26! possibilities for

30
Appendix A: Key Space

SIGABA wiring scheme


31
Appendix B: The
Mechanics of SIGABA

R efer to the wiring diagram in Appendix A for


more information about SIGABA’s cryptographic
The fact that each message used a differ-
ent message indicator (initial setting) meant that
wiring. each message for that day would start in a differ-
ent place along the scrambling maze. While reus-
The SIGABA machine works in a relatively
ing the scrambling maze does not mean the enemy
simple manner. Four input signals are activated with
could immediately read the messages, it did weaken
each keystroke. These four signals go through the
the cipher system. Certainly, overuse of the high-
five control rotors. These rotors (see SIGABA wir-
level SIGABA system could have led to German
ing scheme on previous page) are designated as S1
or Japanese cryptanalytic capabilities against SIG-
through S5 in the diagram. After each keystroke, the
ABA, but the Americans avoided this by relying on
center, or third, control rotor (S3) will step forward
the tactical Hagelin machines to conduct much of
one letter. After every twenty-six keystrokes, both
their chatter. Furthermore, to conduct an exhaus-
the third (S3) and the fourth (S4) control rotor will
tive attack against the SIGABA machine would
step. After every 676 keystrokes, the second (S2),
take 248.4 trial decryptions to get the rotors in the
third (S3), and fourth (S4) rotors will step. The first
right order. Overuse would weaken the encryption
(S1) and fifth (S5) rotors are stationary. This ensures
system, not break it.
that the impulse signals from every keystroke are
scrambled in a different fashion. Rotors S3, S4, and The twenty-six outputs from the control rotors
S2 comprise a 17,576-long counter. That is, after S3 are bundled together into nine different input sig-
does a full cycle, rotor S4 steps. Then, after S4 fin- nals that are then connected to the index rotor bank.
ishes a full cycle (and S3 has completed twenty-six At most, there could be four live wires going into
full cycles), rotor S2 will step. By the time S2 has gone the index rotors. At the minimum, there could be a
through its full cycle, the whole machine has enci- single live wire. In the diagram pictured in Appendix
phered 17,576 letters, which is only 214.1, making this A, there are three live wires going from the control
the “cycle length” for the control rotor scrambling rotor bank to the index rotor bank.
maze. After 17,576 enciphered letters, the machine
The index rotors, denoted I1 through I5, further
will start over, essentially reusing the same scram-
scramble the inputs from the control rotors. Once the
bling system. This was why the rotor placements had
signals travel through the five index rotors, they are
to be changed daily.
bundled by adjacent pairs into five wires that connect
to each one of the five cipher rotors. If the wire com-

32
Appendix B: Mechanics

ing from the index rotor bank is live, then the cipher Each letter of the message travels through only the
rotor it is connected to will step forward one letter. five cipher rotors, C1 through C5 in the diagram. The
Otherwise, it will not move. For each keystroke, there control rotors and index rotors exist only to dictate the
can be between one and four live wires coming from stepping of the five cipher rotors. The plaintext letter
the index rotor bank and going into the cipher rotor from the keyboard travels via wire to the left-hand side
bank. In the diagram, there are exactly two wires that of the cipher rotor bank. The signal wire is then scram-
are active: wire number two and wire number five. bled through the five cipher rotors and emerges on the
This means that both the second (C2) and fifth (C5) other side of the cipher rotor bank as an encrypted let-
cipher rotor will step, but the other three will remain ter. The cipher text is then printed on a small tape. The
stationary for this particular keystroke. decryption process simply works in reverse.

Five pages from the SIGABA user’s manual Crypto-Operating Instructions for Converter M-134-C

33
SIGABA / ECM II Cipher Machine

(2) Plain-text Position ("P").- All keys of the keyboard and the space bar can be
operated, and the converter will print plain, unenciphered text exactly as typed.
The rotors remain motionless during typing. ·
(3) Reaet Position ("R").-Only the numeral keys 1 to 5, inclusive, and the "Blank"
and "Repeat" keys can be operated. The rotors may be zeroized with the
controller in this position and the zeroize-operate key in the "Zeroize" position.
(See paragraph lOb.) The tape will not feed while the controller is at "R."
When the controller is moved to or through the "R" position, the tape may
advance as many as five spaces. This is caused by the tape feed ratchet re-
setting so that printing will begin on the first letter of a five-letter cipher group.
(4) Encipher Position ("E").- The alphabet, "Blank," and "Repeat" keys and the
space bar can be operated. Numeral and "Dash" keys cannot be operated.
The converter enciphers the letters struck on the keyboard and prints the
resulting cipher text.
(6) Decipher Position ("D").-The alphabet, "Blank," and "Repeat" keys can be
operated. Numeral and "Dash" keys and the space bar cannot be operated.
The converter deciphers the letters struck on the keyboard and prints the
resulting plain text.
e. For a more detailed explanation of component parts of the converter, consult the
maintenance instructions for Converter M- 134-C.
7. Classification of Parts.
a. The converter, exclusive of rotors, is classified CONFIDENTIAL.
b. The cipher unit, exclusive of rotors, is classified CONFIDENTIAL.
c. The index rotors are classified CONFIDENTIAL.
d. The alphabet and stepping control rotors are classified SECRET.

34
Appendix B: Mechanics

SECTION III
KEYI NG INSTRUCTIONS
Paragraph
Keying Elements...................................... . .............. . ... . 8
Rotor Arrangement and Alignment of Index Rotors. . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Alignment or Stepping Control and Alphabet Rotors. ......... ..... ..... ....... 10
The 26-80 Check.... ....... ..... ... .. ........................... . ......... 11

8. Keying Elements.-Converter M- 134-C employs two keying elements:


a. The daily keying element consists of the daily rotor arrangement (assembly) and the
alignment of the index rotors. The alignment of the index rotors is different for
each security classification.
b. The message keying element consists of the alignment of the stepping control and
alphabet rotors used at the beginning of the encipherment or decipherment of a
message.
9. Rotor Arrangement and Alignment of Index Rotors.
a. Each converter is provided with five small rotors to be used in the index (front)
position and ten large rotors to be used in the stepping control (middle) and alphabet
(rear) positions.
(1) Index Rotors.- Each of the index rotors bears a sequence of 2-digit numbers:
one rotor is marked with the sequence 10 to 19 inclusive; another, the sequence
20 to 29 inclusive, etc. The complete set of five index rotors is numbered from
10 to 59 inclusive. The index rotors are always used in a fixed order in the five
rotor positions (10-19, 20- 29, 30-39, etc.).
(2) Stepping Control or Alphabet Rotors.- Eacb of the stepping control or alphabet
rotors bears an identifying number, usually opposite the letter "O." Most sets of
rotors will, in addition to the numbers, bear an identifying letter or letters, usu-
ally associated with the identifying number. A set of ten rotors is numbered
from 1 to 10 inclusive, 11 to 20 inclusive, or 21 to 30, etc. These rotors are all
interchangeable and reversible within any set of ten.
b. Rotors are inserted and aligned according to instructions published in a key list
which is included in each edition of a Converter M- 134-C system.* A sample extract
from a key list is shown below.
ROTOR ARRANGEMENT SECRET
DAY (FOR ALL CLASSIFICATIONS) 26-30
OF STEPPING CONTROL ALPHABET INDEX (FRONT) CHECK
MONTH (MIDDLE) (REAR) ALIGNMENT GROUP
:t fJR 4 6 2R 7 1 8 5 9 8R lfJ 28 31 49 5fJ R N H v c
2 2 8R 9R 1 5 6 4R 8 7 s 14 25 33 46 59 s E M N 0

•If old-style key lists are still effective after the effective date or this document, ignore the columns headed
"INITIAL ALIGNMENT (CONTROL AND CIPHER)."
7

35
SIGABA / ECM II Cipher Machine

CONFIDENTIAL RESTRICTED
DAY 26-30 26-30
OF INDEX (FRONT) CHECK INDEX (FRONT) CHECK
MONTH ALIGNMENT GROUP ALIGNMENT GROUP
1 12 28 31 44 63 p w v M T 17 25 36 43 58 M c s D T
2 15 2jl 32 48 56 E H E w B li! 27 84 42 66 R s T H H

c. The Key List.- The key list contains the arrangement of the stepping control and
alphabet rotors for each day of the month and the alignments of the index rotors
for each of the several security classifications for every day of the month. The
arrangement of the stepping control and alphabet rotors remains the same through-
out the cryptographic period for all security classifications. The alignment of the
index rotors differs for each security classification.
(1) Arrangement of Rotors.- Figures in the column marked ROTOR ARRANGE-
MENT (FOR ALL CLASSIFICATIONS) specify which stepping control
and alphabet rotors are to be used on a specific day of the month and the
positions of these rotors in the converter. Numbers in the table refer to the
second digit of the rotor number. A set of rotors bearing the numbers 21 to30
inclusive, for example, will be regarded as being marked 1, 2, 3, ..... 0. "R"
in the table indicates that the rotor so designated is to be inserted in the reversed
position, i. e., the characters on the periphery will appear upside down to the
operator. The rotors will be inserted in their respective positions in order,
from left to right as the operator faces the converter. Example: On the second
day of the month, the sample extract from a key list in paragraph 9b designates
2-3R- 9R- l - 5 for the stepping control rotors and 6-4R- 8- 7- 0 for the alphabet
rotors. Rotors marked 2, 3, 9, 1 and 5 (disregarding the tens digits) will be
inserted in t he control position in that order, from left to right as the operator
faces the converter, with rotors number 3 and 9 reversed. The remaining five
rotors marked 6, 4, 8, 7 and 0, will be inserted in the alphabet position in that
order from left to right with rotor number 4 reversed.
CAUTION: Do not touch rotor contacts when arranging the rotors.
(2) Alignment of Index Rotors.-The sets of numbers under INDEX (FRONT)
ALIGNMENT designate the alignment of the index rotors used for enciphering
and deciphering messages on a specific day of the month. In three separate
colwnns, each headed INDEX (FRONT) ALIGNMENT, the key list gives
the daily alignment of the index rotors for each classification. The alignment
of the index rotors is determined by the classification of the message and the
day of the month. Example: According to the key list above, on the first day
of the month the numbers of the index rotors should be aligned from left to
right on the white reference mark at 10 23 31 49 50 for SECRET messages; at
12 28 31 44 53 for CONFIDENTIAL messages; and at 17 25 36 43 58 for
RESTRICTED messages.
10. Alignment of St epping Control a nd Alpha bet Rotors.-The alignment of the
stepping control and alphabet rotors at the beginning of encipberment or decipherment
of a message constitutes the message keying element. The letters to which the stepping

36
Appendix B: Mechanics

control and alphabet rotors are aligned constitute the message rotor alignment (internal
message indicator). T he message rotor alignment is derived and aligned on the stepping
control and alphabet rotors in the following manner:
a. Select a group of any five letters at random (message indicator). All 26 letters of
the alphabet, including the letters "0" and "Z," may be selected. Letters of the
alphabet in proximity to the letter "O," i. e., P, Q, R, or L, M, N, will not be deliber-
ately or consistently selected in the message indicator merely to reduce the number of
steps required to align the letters of the message indicator on the stepping control
rotors as explained below. Bona-fide words must not be used except as they occur
by chance.
b. Zeroize the converter. This is accomplished by switching the zeroize-operate key
to "zeroize," turning the controller to "R," and then pressing down the "Blank"
and "Repeat" keys simultaneously until the letter "0" on the stepping control and
alphabet rotors comes to rest at the reference mark.
c. Leave the controller at "R" and switch the zeroize-operate key to "Operate."
d. Strike the numeral "1" key the number of times required to align the first stepping
control rotor (next to the left end plate) to the first letter of the message indicator.
The first stepping control rotor will step one letter each time the "1" key is depressed.
e. Align the second stepping control rotor by striking the numeral "2" key, the third
by striking the numeral "3" key, etc., until all five stepping control rotors are aligned
to the five letters of the message indicator. With each step of the stepping control
rotors, the alphabet rotors will step in an irregular manner.
N OTE: If the letter "O" is to be aligned on any of the five stepping control rotors,
it will be necessary to step that rotor 26 times when setting up the message
indicator.
f. If any rotor is stepped past the correct letter or if the rotors are not aligned in proper
sequence, the entire process must be repeated from the zeroize position (subparagraph
lOb). Do not use the "Repeat" key with the numeral keys in aligning the m~ge
indicator and avoid a sharp, quick touch of the numeral keys. It is p<>s.5ible to press
the numeral keys and release them too quickly so that the stepping control rotors
will step but the alphabet rotors will not, thus resulting in an incorrect alignment.
g. After the stepping control rotors have been aligned, check the alignment of the
alphabet rotors to insure that all five are not aligned to the letter "O." The alphabet
rotors should step in an irregular manner while the stepping control rotors are being
aligned. If for any reason all of the alphabet rotors do not step, they will remain
aligned to the letter "O." This is an indication that the converter is not functioning
properly or that the procedure outlined herein has not been followed correctly.

11. The 26-30 Check.


a. The 26- 30 check groups provided in the key list are used to check the correctness of
the daily rotor arrangement and index alignment and t he stepping of the stepping
control and alphabet rotors. The 26-30 check is performed as follows:
(1) Arrange the stepping control and alphabet rotors and align the index rotors in
accordance with the key list and security classification to be checked.

37
Notes

1. David Kahn, The Codebreakers: The Story of Secret 18. Ibid., 39.
Writing. New York: MacMillan, 1967, 235. 19. Ibid., 39j.
2. Kahn, Codebreakers, 415. 20. Ibid., 40.
3. Jack Levine, United States Cryptographic Patents, 21. Ibid., 39-40.
1861-1989 (Terre Haute, IN: Cryptologia, 1991), 22. Ibid., 40; Rowlett, Story of Magic, 97.
85. 23. Rowlett, Oral History Interview 1974, OH-1974-
4. Frank B. Rowlett, The Story of Magic: Memoirs of 01, 40, Ft. Meade, MD: Center for Cryptologic
an American Cryptologic Pioneer (Laguna Hills, History.
CA: Aegean Park Press, 1998), 69. Also Cap- 24. Ibid.; Rowlett, Story of Magic, 97-98.
tain Laurance Safford, United States Navy OP- 25. Heather Ellie Kwong, thesis: Cryptanalysis of the
20-S-5, History of Invention and Development of SIGABA Cipher, San Jose State University, 2008.
the Mark II ECM, SRH-360, 19-21 (National 26. http://www.sccs.swarthmore.edu/users/08/ajb/tmve/
Archives and Records Administration [NARA]: wiki100k/docs/SIGABA.html.
RG 457, box 1124). 27. Laurance Safford, History of Invention and Devel-
5. Army Security Agency, History of Converter opment of the Mark II ECM, SRH-360, NARA
M-134-C, Vol. 1, or SRH-359 (Washington, DC: RG 457, Box 1124. The ECM Mark I was the
Army Security Agency, n.d.), 18. Navy’s primary cipher machine from May 1936 to
6. Ibid., 51-60. January 1942. It was also the Navy’s first attempt
7. Friedman Collection, NSA/CSS Archives acces- at an “in-house” cipher machine. In spite of ECM
sion #47270, box 14, folder 1. Mark I’s formidable cryptographic strength, the
8. Ibid. Navy remained uneasy about its vulnerability to
9. Levine, United States Cryptographic Patents, 115. exploitation. The Mark I arguably was the most
10. Friedman Collection, 1920-1960, NSA/CSS powerful cryptologic device extant, but at the time
Archives accession #47270, box 14, folders 1-2. the cryptographers at OP-20-G had little appre-
11. Ibid., folder 1. ciation of this fact. The ECM Mark I embodied
12. Ibid. elements of contemporary Hebern machines and
13. Ibid. then was heavily modified by Donald Seiler and
14. Ibid. Laurance Safford. These novel cryptographic
15. Frank B. Rowlett, Oral History Interview 1974, enhancements placed the Mark I in a league of
OH-1974-01, Part B, 45c, Ft. Meade, MD: Cen- its own, well above the Hebern machine. Seiler
ter for Cryptologic History (CCH). also went on to perform much valuable engineer-
16. Ibid., 45c-45d. ing work on the ECM Mark II’s development.
17. Ibid., 38. According to Safford, the Mark I was beset with

38
Notes

numerous mechanical deficiencies that required them because user errors in plugging had directly
almost constant maintenance to keep it opera- led to compromise of the key, not to mention the
tional. This fact alone was probably the largest, effect on the reliability of communications. The
immediate driving factor in the Navy’s quest for Navy’s greatest argument in favor of rotors over
a new machine. When he returned from sea duty plugboards was that rotors offered 49 times more
and was briefed by Wenger on his interview with stepping combinations than the latter (SRH
Friedman and Rowlett, Safford readily embraced 360, 28). German and American cryptographers
the SIGABA principles and pressed for its independently of each other conceived of using
development. plugboards as enhancements to their respective
28. Rowlett, Oral History Interview, 1974, OH-1974- cipher machines.
01, 41, Ft. Meade, MD: Center for Cryptologic 40. Rowlett, Oral History Interview 1974, OH-1974-
History. 01, 45, Center for Cryptologic History, Ft. Meade,
29. Ibid. MD.
30. Safford, History of Invention and Development of 41. Ibid., 46.
the Mark II ECM, 24-25, 29. 42. Safford, History of Invention and Development
31. Rowlett, Oral History Interview 1974, OH-1974- of the Mark II ECM, 40; The Navy deemed the
01, 45n, Ft. Meade, MD: Center for Cryptologic thermite emergency destruction devices too dan-
History; Army Security Agency, History of Con- gerous to use aboard ships where fire is the great-
verter M-134-C, vol. 1, SRH-359, Washington, est threat. Navy tests of thermite bombs reduced
DC: Army Security Agency, n.d., 139. the ECM’s critical components to molten metal
32. Rowlett, Story of Magic, 143. in 97 seconds. Safford, History of Invention and
33. Army Security Agency, History of Converter Development of the Mark II ECM, 61. This not-
M-134-C, 138-139; Rowlett, Story of Magic, 143. withstanding, the Army had few qualms about
34. Rowlett, Story of Magic, 143-144. fielding both thermite and TNT charges atop
35. Ibid., 144. SIGABA’s 800 lb. security cabinets aboard tacti-
36. Rowlett, Oral History Interview 1974, OH-1974- cal communications vehicles. There is no record
01, 46, Ft. Meade, MD: Center for Cryptologic of an unintended mishap associated with these
History. devices; even so, the Navy’s concerns were partly
37. On 1 August 1941 the sister services formally vindicated by more than one Army incident of a
adopted the SIGABA/ECM II as their joint near discharge.
cipher machine. Each agreed not to share any 43. Safford, History of Invention and Development of
information whatever about the machine to any the Mark II ECM, 42; 37-52 passim.
outside person or organization. 44. Ibid., 6; the ECM Mark I, despite the Navy’s
38. Rowlett, Oral History Interview 1974, OH-1974- reservations about it, was still more powerful
01, 46, Ft. Meade, MD: Center for Cryptologic than any cipher device fielded by either the Brit-
History. ish or the Axis Powers.
39. Ibid., 47. The Navy’s Donald Seiler had given due 45. Ibid., 6-8.
consideration to plugboards and rejected them in 46. Ibid.
favor of the index rotor wheels even before the 47. Ibid.
first ECM Mark II had reached the blueprint 48. Ibid.
stage. Safford, History of Invention and Develop- 49. Ibid.
ment of the Mark II ECM, 39. Safford, Chief OP- 50. Safford, History of Invention and Development of
20-G, relates that besides being inconvenient to the Mark II ECM, 54.
the user, plugboards in the past had not pre- 51. Ibid., 55-56.
vented the initial solutions of six different cipher 52. R. A. Ratcliff, Delusions of Intelligence: Enigma,
machines the Navy had tested. He also eschewed Ultra, and the End of Secure Ciphers (New York:

39
SIGABA / ECM II Cipher Machine

Cambridge University Press, 2006), 81. Also Saf- War II as Revealed by “TICOM” Investigations and
ford, History of Invention and Development of the Other Prisoner of War Interrogations and Captured
Mark II ECM, 52. Material, Principally German, vol. 5, The Ger-
53. Army Security Agency, History of Converter man Air Force Signal Intelligence Service. Army
M-134-C, 7. Security Agency, 10 October 1946.
54. Ibid., 147-158, 247-248. 65. War Diary of the Signal Intelligence Group, Feb-
55. The Navy generated its SIGABA key at the ruary–November 1944 (NSA/CSS Archives,
Nebraska Avenue Naval Station where it oper- accession #5411, box G22-0303-3), 64.
ated its ECM Mark II maintenance facility. 66. Ibid.
Army cryptographers produced SIGABA keys 67. Erich Huttenhain, interview, 10 July 1945
at Arlington Hall. Later in the war, the Army (NARA: RG 457, box 1006).
moved its SIGABA maintenance and training 68. Ibid.
functions to Vint Hill Farms near Warrenton, 69. Ibid.
VA. 70. Ibid.
56. Safford, History of Invention and Development of 71. Ibid. The Japanese Signal Intelligence Service, 17
the Mark II ECM, 31. October 1952 (NARA: RG 457, box 1129).
57. Ratcliff, Delusions of Intelligence, 105, 164, 168, 72. Ratcliff, Delusions of Intelligence, 178; Pre-War
177. By the war’s end, more than 8,000 SIG- Radio Intelligence Activities in the Philippines,
ABAs had the capability to interoperate with the 63, CCH Files; Timothy Mucklow, Federal His-
British Typex machine. tory Journal, 2011, 59.
58. J. G. Seabourne, The Signal Intelligence Service 73. Army-Navy Joint Policy Concerning Distribu-
of the German Luftwaffe, vol. 13. November 24, tion and Disclosure of Cryptographic Design of the
1945 (National Archives and Records Admin- ECM-M134C, 26 June 1942; Safford, History of
istration: RG 457, box 976). The Navy’s anxi- Invention and Development of the Mark II ECM,
ety about its own ECM Mark I led Wenger to 57-58.
approach Friedman about the Army’s crypto- 74. Safford, History of Invention and Development of
graphic research—Rowlett, Oral History Inter- the Mark II ECM, 110-111, 115-116; Ratcliff,
view 1974, OH-1974-01, 45m—in the first Delusions of Intelligence, 168.
place. The sister services’ continued development 75. Letter, Col. George A. Bircher, SC, to CG, AAF,
of, and their absolutely rigorous application of, Washington, DC, Attn: Ch, Sec, Air Communi-
COMSEC doctrine demonstrate their concern cations Office, 9 May 46, sub: Policy on Storage
over enemy attempts to break into U.S. crypto of Converter M-134-C:
systems. . . . 2. This Agency does not consider that
59. JN-A-20 messages, 1942 (National Archives the cessation of hostilities justifies any relax-
and Records Administration: RG 457, box 1006, ation in security regulations pertaining to
temporary folder). cryptographic material and has noted with
60. Ibid.; Ratcliff, Delusions of Intelligence, 201-203. alarm a tendency towards such relaxation
61. Rowlett, Oral History Interview 1974, OH-1974- during recent months as has been evidenced
01, 105. by an increase in the number of physical and
62. Ratcliff, Delusions of Intelligence, 202. Several dif- cryptographic compromises.
ferent German cryptologic organizations named 3. Since the cryptographic principle and
SIGABA as the “Big Machine.” design of Converter M-134-C is in the
63. Rowlett, Oral History Interview 1974, OH-1974- sole possession of the United States and it
01, 158-159. is considered to be the best cryptographic
64. Seabourne, Signal Intelligence Service. Also see device of its type, it is not advisable to jeop-
report: European Axis Signal Intelligence in World ardize its use or storage under unsatisfac-

40
Notes

tory security conditions. It is considered because it simply could not keep pace with mod-
that eliminating the necessity for a 24-hour ern high-speed telecommunications. Research
armed guard, regardless of geographical based on newly acquired information suggests
location, even if the converter were stored that SIGABA was taken out of the invento-
in a CH-76 in a locked code room, would ry to keep the advanced technology out of the
result in such a condition, and that no justi- hands of the Soviets. Because every SIGABA
fication exists for undergoing such a risk. . . . and rotor had been accounted for and because
Ratcliff, Delusions of Intelligence, 168. Army and Navy cryptologists had much con-
76. History of Converter M-134-C, 5 May 1950, fidence in the integrity of SIGABA principles,
included in the CCH copy of SRH 360, 1-10. they were confident that SIGABA’s secrets
77. Ibid., 7; Extracts: Annual Report, Security Divi- could be retained; History of Converter M-134-C,
sion, CSAS-80, Fiscal Year 1948 (included in 5 May 1950 included in CCH copy of SRH
CCH copy of SRH 360), 1-5; Excerpt: Tab A10, 360, 7-10 (included in CCH copy of SRH 360).
“Staff Study Concerning the Issue of SIGROD SIGABA and its temporary successor SIG-
Machines to all Military Attachés,” Annual ROD were slowly replaced in the 1950s by the
Report, Sec Div, Tech Staff, FY 48 (23 April TSEC/KL-7 (ADONIS/POLLUX). The new
1947) 1948 (included in CCH copy of SRH cipher machine was an electronic-mechanical
360), 1-4; Extract: Tab A7, “Staff Study on the hybrid that employed a programmable cipher
Introduction of the SIGROD,” Annual Report, rotors/bezel assembly (eight rotors/thirty-six
Security Division, Technical Staff, FY 48 (3 Feb- pins), cams, and vacuum tube technology along
ruary 1950), 19 (included in CCH copy of SRH with a novel re-flexing principle. It was phased
360). out of the U.S. military inventory in the early
78. Until recently, cryptologic historians were under 1980s.
the impression that the end for SIGABA came
not because of any cryptographic weakness but

41
Selected Bibliography

Published Pekelney, Rich. “Electronic Cipher Machine


(ECM) Mark II,” San Francisco: USS Pam-
Budiansky, Stephen. Battle of Wits: The Complete
panito, San Francisco Maritime National Park
Story of Codebreaking in World War II. New York:
Association, July 2008. http://www.maritime.
The Free Press, 2000.
org/ecm2.htm.
Chan, Wing On. “Cryptanalysis of SIGA-
Ratcliff, R. A. Delusions of Intelligence: ENIGMA,
BA.” Master’s thesis, San Jose State Univer-
Ultra, and the End of Secure Ciphers. New York:
sity, 2007. http://cs.sjsu.edu/faculty/stamp/
Cambridge University Press, 2006.
students/Sigaba298report.pdf.
Rowlett, Frank B. The Story of Magic: Memoirs of an
Clark, Ronald William. The Man Who Broke PUR-
American Cryptologic Pioneer. Laguna Hills, CA:
PLE: The Life of the World’s Greatest Cryptolo-
Aegean Park Press, 1998.
gist, Colonel William F. Friedman. Boston: Little,
Savard, J.J.G. and R. S. Pekelney. “The ECM Mark
Brown, 1977.
II: Design, History and Cryptology.” Cryptologia,
Kahn, David. The Codebreakers: The Story of Secret
23, no. 3 (1999): 211-228.
Writing. New York: MacMillan, 1967.
Stinson, Douglas R. Cryptography: Theory and Prac-
Kwong, Heather Ellie. “Cryptanalysis of the SIG-
tice, 3rd ed. Boca Raton, FL: Taylor & Francis
ABA Cipher.” Master’s thesis: San Jose State
Group, 2006.
University, 2008.
Tucker, Alan. Applied Combinatorics, 4th ed. Singa-
Lee, Michael. “Cryptanalysis of the SIGABA.”
pore: Wiley, n.d.
Master’s thesis, University of California, Santa
Werner, Herbert A. Iron Coffins: A Personal Account
Barbara, 2003. http://ucsb.curby.net/broadcast/
of the German U-Boat Battles of World War II. New
thesis/thesis.pdf.
York: Holt, Rinehart and Winston, 1969.
Levine, Jack. United States Cryptographic Patents,
1861-1989. Terre Haute, IN: Cryptologia, 1991.
Mucklow, Timothy and LeeAnn Tallman. “SIG-
ABA/ECM II: A Beautiful Idea.” Cryptologic
Quarterly, 30 (2011): 3-25.

42
Selected Bibliography

From the Center for Cryptologic Hurt, John B. A Version of the Japanese Problem in the
History and National Security Agency/ Signal Intelligence Service (Later Signal Security
Agency) 1930-1945, SRH-252. Washington, DC:
Central Security Service Archives Army Security Agency, n.d.
Miller, A. Ray. The Cryptographic Mathematics of
ENIGMA. Ft. Meade, MD: Center for Crypto-
SRH = Special Research Histories
logic History, National Security Agency, 2006.
SRMN = U.S. Navy Discrete Records of Historical Mowry, David. “William F. Friedman” (Ft. Meade,
Cryptologic Import MD: Center for Cryptologic History, National
Security Agency, 2002).
Photographic Equipment – IBM Sigaba Unit manual,
Army Security Agency. History of Converter n.d. (NSA/CSS Archives Accession #41220, Box
M-134-C, vol. 1, SRH-359. Washington, DC: G03-0701-6).
Army Security Agency, n.d. Reeder, William. Center for Cryptologic History,
Cryptographic Security Section. HCM (Hebern biography files.
Cipher Machine) Cipher No. 21, SRMN-063. Rowlett, Frank B. Oral history interviews #OH-
Washington, DC: Navy Department, Office of 1974-01 through OH-1974-12. Center for
Chief of Naval Operations, 25 August 1935. Cryptologic History, 1974.
“Crypto-Operating Instructions for M-134-C,” Safford, Captain Laurance. Officer Biographies,
or “SIGQZF-3,” November 1946. NSA/CSS n.d. (NSA/CSS Archives Accession #47403, Box
Archives, accession #13943, box H18-0506-6. H16-0203-3).
Friedman Collection, 1920-1960, NSA/CSS ———. History of Invention and Development of
Archives, accession #47270, boxes 1-15. the Mark II ECM, SRH-360. Washington, DC:
Friedman, Elizebeth. Oral History Interview #OH- United States Navy OP-20-S-5, Office of the
1973, Center for Cryptologic History, May Chief of Naval Operations, 30 October 1943.
16-17, 1973. ———. Center for Cryptologic History, biography
The Friedman Legacy: A Tribute to William and Eli- files.
zebeth Friedman, 3rd ed. Ft. Meade, MD: Cen- Small, Albert W. Arlington Hall Station, “Letter to
ter for Cryptologic History, National Security Lieutenant Colonel Charles G. Renfre, Armed
Agency, 2006. Forces Security Agency,” 12 October 1950. NSA/
Friedman, William. Elementary Course in Cryptanal- CSS Archives accession #5515, box #G22-0303-
ysis, SRH-214. Washington, DC: Army Security 3, folder #19.
Agency, 1930. War Department. “Converter M-209, M-209-A,
———. Elementary Course in Cryptanalysis, SRH- M-209-B, Technical Manual.” Washington,
216. Washington, DC: Army Security Agency, DC, 17 March 1944. Reprint. Washington, DC:
1940. National Cryptologic Museum.
———. Elementary Course in Cryptanalysis, SRH- “War Diary of the Signal Intelligence Group,” Feb-
218. Washington DC: Army Security Agency, ruary–November 1944. NSA/CSS Archives,
1946. accession #5411, box G22-0303-3.
———. Expansion of the Signal Intelligence Service from Wilcox, Jennifer. Sharing the Burden: Women in
1930 to 7 December 1941, SRH-134. Washington, Cryptology during World War II. Ft. Meade, MD:
DC: Army Security Agency, 4 December 1945. Center for Cryptologic History, National Secu-
Friedman, William and Lambros Callimahos. Mili- rity Agency, 1998.
tary Cryptanalytics, Part I & II. Washington, DC: ———. Solving the ENIGMA: History of the Crypt-
National Security Agency, 1956. analytic Bombe. Ft. Meade, MD: Center for Cryp-
tologic History, National Security Agency, 2006.

43
SIGABA / ECM II Cipher Machine

From the National Archives and Muentz, Lieutenant D. R. Interview, 10 July 1945.
Records Administration, RG 457, box 1006.
College Park, MD Rentschler, R. R. Interview, 10 July 1945. RG 457,
box 1006.
“CRYPTOSYSTEMS 742 (SIGFKE) & 604 Rosen, Leo. M-134-C. RG 457, box 1124.
(SIGFHK),” April 12, 1948. NSA/CSS Archives Safford, Laurance. United States Navy OP-20-S-5,
accession #13943, box H18-0506-5. History of Invention and Development of the Mark
Friedman, William F. “Analysis of a Mechanico- II ECM, SRH-360. RG 457, box 1124.
Electrical Cryptograph,” 1934-1935. RG 457, Seabourne, J. G. “The Signal Intelligence Service of
box 745. the German Luftwaffe,” vol. 13. November 24,
Huttenhain, Erich. Interview, 10 July 1945. RG 457, 1945. RG 457, box 976. The Seabourne Report is
box 1006. named after Colonel J. G. Seabourne, the project
“The Japanese Signal Intelligence Service,” 17 leader.
October 1952. RG 457, box 1129.
JN-A-20 messages, January 24, 1942. RG 457, box
1006, temporary folder.

44
Related Publications

The Cryptographic Mathematics of ENIGMA

The Friedman Legacy: A Tribute to William and Elizebeth Friedman

German Cipher Machines of World War II

Solving the Enigma: History of the Cryptanalytic Bombe

You might also like