America’s First Code-Breakers – How the U.S. Military Helped Win the WW1 Intelligence War

America's military was entirely unprepared for the intelligence war that awaited it on the Western Front. It would have to race to catch up. (Image source: WikiCommons)

The U.S. military was entirely unprepared for the information war that awaited it on the Western Front. It would have to race to catch up. (Image source: Imperial War Museum via WikiCommons.)

“Despite a few setbacks, America’s code and code-breaking efforts enjoyed some remarkable successes.”

By John F. Dooley

WHEN AMERICA ENTERED the First World War on April 6, 1917, the United States Army had virtually no experience with codes and code-breaking. France and Britain had been solving German codes and ciphers for the past three years and the U.S. military raced to get its own operations up and running.They would be based partly in Washington, and partly in France.

For the next year, the United States would play catch-up in everything from codes and ciphers to traffic analysis and radio direction finding. By the spring of 1918, America was contributing its own expertise to the wireless war. Despite a few setbacks, America’s code and code-breaking efforts enjoyed some remarkable successes. Yet, the United States’ intelligence war against Imperial Germany has been all but forgotten.

First World War armies on the front lines communicated with their respective headquarters by way of what were called "trench codes." America had no such system when its armies began arriving in France.

Armies on the front lines during the First World War communicated with their respective HQs by way of “trench codes.” America had no such system when its troops began arriving in France in 1917… and it would need one fast.

An Inauspicious Start

Since as far back as 1914, armies on the Western Front relied on what were known as ‘trench codes’ to pass orders to regiments on the line. The first American trench code, a small book consisting of some 1,600 words and phrases, was developed in early 1918. At its heart was a one-part code that used a monoalphabetic substitution cipher as its superencipherment step to make messages more secure. It was intended for distribution down to the company level. But it was never used. Why? It turns out it wasn’t very secure. In early May 1918, as the code was being prepared for distribution to units on the line, the American Expeditionary Forces’ (AEF) assistant chief signal officer, Major Parker Hitt, set out to test the security of the new system. He asked one of the army’s cipher experts try to break a set of coded messages. Lieutenant J. Rives Childs, head of the Cipher Solutions Section, was given a copy of the trench code and a set of 44 cryptograms that were scrambled using the trench code. Despite having no details of how the encryption worked, Childs returned to his commanding officer in less than a day with all 44 messages solved. This was the death knell for the first American trench code.

Wireless radio, which was barely more than a decade old by 1914 enabled instantaneous communications for armies. It also allowed enemies to eavesdrop. (Image source: WikiCommons)

Wireless radio, which was barely more than a decade old by 1914, enabled instantaneous communications for armies in World War One. It also allowed enemies to eavesdrop. (Image source: WikiCommons)

Towards a Working Trench Code

After Childs brilliant cryptanalytic effort, the Americans turned to creating a new code. This time, it was to be a two-part code, which would have the advantage of being easier to both encode and decode in the field. Plus, it eliminated the need for the superencipherment. Its main disadvantage was that the more it was used, the easier it would be for the eavesdropping Germans to figure it out. An enemy offensive (or even just a trench raid) also increased the chances that the Germans would capture a copy of the codebook. Because of this, the Code Compilation Section committed itself to creating, printing, and distributing a new version of the code every 10 days to two-weeks — something the Allies had never done. The result was the spectacularly successful River series of codes, beginning with the delivery of the Potomac code on June 24, 1918. The first edition was 2,000 copies and contained code words for 1,800 words and phrases. The code was squeezed into just 47 pages and was designed to fit in a pocket. It was released down to the battalion level and was followed by half a dozen more codes before the end of the war in November 1918.

The aftermath of the Black Tom Island munitions explosion of 1916. The blast, which rocked New York Harbor and injured hundreds, was carried out by German agents operating in the U.S. Code breakers with military intelligence would soon ramp up efforts to intercept enemy messages and net German spies operating in the U.S. (Image source: WikiCommons Media)

The aftermath of the Black Tom Island munitions explosion of 1916. The blast, which rocked New York Harbor and injured hundreds, was carried out by German agents operating in the United States. Code breakers with military intelligence would ramp up efforts to intercept and crack enemy messages so they could net German spies operating on American soil. (Image source: WikiCommons Media)

John Manly and the Waberski Cipher

Lothar Witzke. (Image source: author)

Lothar Witzke. (Image source: author)

From the first day of America’s entry into the war, German spies planted in Mexico began slipping across the border to attempt missions of sabotage and espionage in the United States. On Feb. 1, 1918, one such operative, a young German naval lieutenant turned spy named Lothar Witzke, infiltrated the border using a fake Russian passport bearing the name Pablo Waberski. As soon as he set foot on American soil, agents from the Treasury Department’s Bureau of Investigation, the precursor to the FBI, were waiting for him. On searching Witzke, a cryptogram was found sewed into his jacket. This document was dispatched to military intelligence headquarters in Washington where Captain John Manly, second in command of the section, took it up and set to work deciphering the message.

John Manly

John Manly

Manly’s work on decrypting the Waberski cipher was masterful. He began by determining the language of the cipher message. He conducted a frequency analysis to give himself hints on the type of cipher system, which provided data on the letters that were and were not used. Manly guessed that he was looking at a columnar transposition cipher. Armed with knowledge of German, he began to organize the message into columns that would make sense for that language. In the end he came up with a brilliant solution. The Waberski message and Manly’s decipherment of it were used to convict Witzke of espionage and earn him a death sentence.

Army intelligence officers try to unravel the enemy's plans.

History’s first ‘information war’ — Army intelligence officers try to unravel the enemy’s plans.

Secret Inks! Carver and the Maud Letter

The "Maud" letter. (Image source WikiCommons)

The “Maud” letter. (Image source WikiCommons)

In early 1918, a German agent accidentally switched two secret messages and mailed the wrong one to an address in New York. It was intercepted by postal department censors who suspected it contained a secret message in invisible ink. Dr. Emmett K. Carver ran the military intelligence secret ink lab in New York City. Using a combination of chemical reagents known to develop certain German invisible inks and a new test using iodine vapor, Carver exposed the secret message and managed to photograph it just before it disappeared forever. This message, signed by an operative with the alias “Maud”, was intended for one of the most notorious female German spies in America, Madame Marie de Victorica. Her subsequent arrest and confession led to the collapse of the largest German spy ring in the United States.


SIDEBAR: France’s Top Cryptanalyst Cracks the Germans’ Toughest Code

A new German code was introduced just prior to the 1918 Spring Offensive. One French cryptanalyst worked tirelessly to crack it.

A new German code was introduced just prior to the 1918 Spring Offensive. One French cryptanalyst worked tirelessly to crack it.

ONE OF THE First World War’s greatest code-breaking coups belongs to France.

As part of the preparations for the Kaiser’s enormous 1918 Spring Offensive, the German general staff established a whole new encryption system – one of apparent simplicity, but real subtlety. This new cipher would be in constant use until the Armistice. Dubbed ADFGX by the Allies, the code carried a larger volume of messages than any other cipher in the history of the war.

The letters ADFGX on two axis were the basis of an ingenious 1918 German code. By sending two letters, in the case of this diagram, F and X, the recipient could decode the solution: D.

The last German code of WW1 was based on the letters A,D,F,G and X, which were placed along two axis of a 25-character grid. Pairs of these primary letters worked like coordinates on the grid, which recipients could then use to plot the intended letter. In the case of this diagram, ‘F’ and ‘X’ represent the letter ‘D’. Entire messages would be made up of long sequences of ADFGX pairings. (Image source: WikiCommons)

ADFGX is what is known as a fractionating cipher system in which each letter is enciphered using two or more letters, making the coded message or “ciphertext” longer than the uncoded dispatch or ‘plaintext’. The ciphertext is then transposed, breaking up the enciphered pairs of letters. The ADFGX cipher system was used for German command messages at the division level and above.

George's Painvin. (Image source: author)

Georges Painvin. (Image source: author)

Georges Painvin, the French military’s code wizard, worked virtually non-stop from March 5 through the first week of April, trying to crack ADFGX. He lost nearly 30 pounds in the process. On April 6 intelligence officers at American headquarters were handed one of Painvin’s deciphered messages from April. It was accompanied by a note explaining completely the complex method of enciphering and giving the keys by which all messages for that date could be read.

Painvin’s method of solving the ADFGX cipher relied on finding messages with stereotyped beginnings that the Germans used often. These messages would have similar patterns in the ciphertext, making them easier to put into columns.

DooleybookcoverJohn F Dooley is the author of the recently released Codes, Ciphers & Spies: Tales of Military Intelligence in World War I (Springer/Copernicus). He is also the author of A Brief History of Cryptology and Cryptographic Algorithms and Software Development and Professional Practice. The William and Marilyn Ingersoll Professor of Computer Science at Knox College, his research interests include the history of cryptology, computer security, and software development in small teams. He lives with his wife and cats in Galesburg, Illinois.

Leave a Reply