July-September 1941: Defending Canadian Military Communications with Codes and Ciphers

"Convoy assembling in Bedford Basin, Halifax, N.S.," 1941. Canada. Dept. of National Defence / Library and Archives Canada / PA-128093

On July 28th, 1941, the D.M.O. & I. replied to Brigadier Pope, admitting that what Pope had surmised about the dangers of enemy interception of Canadian military communications was probably founded. In addition to accepting Pope’s suggestion restricting the use of R.T. sets, the Colonel proposed that a “simple code” be employed on the East Coast, as was already used on the West Coast, for these types of communications. The Canadian military needed a coordinated coast-to-coast policy on communications security. D.M.O. & I. wrote to Pope that “If you concur, I shall advise Atlantic and Pacific Commands accordingly, and request them to arrange co-operation of the Naval authorities in their localities.” On July 31st, Brigadier Pope noted on the letter that he concurred, and returned it to D.M.O. & I. for further action (CDWI00332).

On August 2nd, 1941, the Director of Military Operations and Intelligence wrote again to Brigadier M.A. Pope. He felt the matter should be referred to Captain Houghton, the Secretary of the Chiefs of Staff Committee, who would bring the communications security issue to both the Pacific and Atlantic Joint Service Committees for coordination across the different services. At the discretion of the Chief of the General Staff, D.M.O. & I. would arrange for this next step. C.G.S. approved the proposal that same day, and the responsibility went to Captain Houghton (CDWI00333).

Later that day, on August 2nd, the Chief of the General Staff sent a memorandum to the Secretary of the Chiefs of Staff Committee, Captain Houghton, that would become the basis of the letter sent out to the East and West Coasts, asking for input from the Joint Service Committees on possible solutions to this problem (CDWI00334). Captain Houghton wasted no time in drafting an edited letter and sending it on to the Secretaries of the Joint Service Committees on the Atlantic and Pacific Coasts by August 5th, 1941. The letter laid out the problem logically: 1) The Canadians had to use radio telephony and wireless telegraphy on occasion for coastal defence; 2) Although the sets in use were low-powered and wouldn’t normally be expected to propagate local communications across an ocean, evidence from Canada’s own wireless interception station in Ottawa suggested that even low-powered sets could transmit across enormous distances in certain atmospheric conditions; 3) In light of these discoveries, the Chiefs of Staff Committee requests that a “simple code” be decided on and employed in all wireless communications for coastal defence, to avoid giving the enemy information about Canadian ship movements (CDW00336).

"Front Cover of Field Service Pocket Book Pamphlet No. 2: Orders and Intercommunication," 1939.

A copy of the instructions given to Atlantic Commanders on September 2nd regarding the use of W/T and R/T sets was attached to the September 12th letter from Halifax (CDWI00340), and sent to National Defence Headquarters at the same time. It was important for Atlantic Command to reassure N.D.H.Q. that they were taking communications security seriously, as the initial concern about plain language transmissions had come from Canadian Army wireless sets in use on the East Coast. This document lays out the steps taken on the East Coast to improve communications security in the period before the construction of new communications was complete, and before orders arrived from N.D.H.Q. directing the use of codes for military wireless communications from coast to coast (CDWI00339). The instructions begin by laying out the problem of communications security, and when W/T and R/T sets should be used. Then, the Field Service Pocket Book (Pamphlet No.2, 1939)—provided to Canadian forces by the British War Office—is quoted at length.

Because the quality of CDWI00339 is poor, it may be helpful for readers to return to the original source of these quotes: it seems that these were taken from pages 18-19 of the aforementioned pamphlet, which is available online at this link: https://dai.mun.ca/PDFs/cns/FieldServicePocketBook1938.pdf (Pamphlet No.2 begins on page 39 of the PDF). The General Officer Commanding-in-Chief, Atlantic Command instructed the East Coast forces to use code names and ciphers for their transmissions, writing:

“The use of Playfair or a simple transposition cipher, with frequently changing keywords, provides an effective method of obtaining security in the use of W/T and R/T, and such cipher must be used whenever the slightest doubt exists as to whether or not a transmission will break security” (CDWI00339).

On September 13th, 1941, the day after N.D.H.Q. received the letter from Atlantic Command, the response from the Pacific Coast made its way from Victoria, British Columbia to National Defence Headquarters in Ottawa. Signed by representatives of the Canadian Army, Navy, and Air Force, the West Coast letter to National Defence Headquarters essentially threw the ball back into Ottawa’s court. The Pacific Joint Service Committee had referred the initial request from Ottawa to a Communications Sub-committee of the J.S.C., Pacific Coast, which had considered the matter in the intervening weeks and decided that Ottawa’s initial suggestion of a “simple code” used across both coasts was the best course of action. Pacific Command would await instruction from Ottawa on what that code may be, and provided no suggestions of their own (CDWI00338).

By the end of September, the Chiefs of Staff Committee had reviewed the “recommendations” received from the Atlantic and Pacific committees and come up with a unifying national policy on securing wireless communications. The decision was to adopt “one of two Codes that have been devised in the United Kingdom to meet the dangers to security alluded to.” Both of these codes would be manually implemented, with operators referring to codebooks to find the correct number group to substitute for each word or phrase they needed to transmit (e.g., ‘11825’ could be code for ‘convoy’). For now, the code S.P. 02383 would be used, but soon new codebooks would arrive from Britain so Canada could implement the newer “ECCO” code instead. Captain Houghton, Secretary of the Chiefs of Staff Committee, informed the Atlantic and Pacific J.S.C.s that:

“Pending the arrival of this second code, and in view of the extreme undesirability of plain language communications traffic continuing, orders have been given to R.C.N. Commands to issue S.P. 02383 for use forthwith, this being quicker and more secure than local codes and Playfair… Orders have been given to R.C.N. Commands that the use of P/L and local codes on R/T or W/T is to cease altogether on the bringing into force of S.P. 02383.” (CDWI00341)

From these documents, we can see how the Canadians adapted to changing technological possibilities as the war went on, and how they learned to defend Canadian military communications from potential enemy interception. We can also see the organizational and bureaucratic challenges posed when a policy needed to be implemented or changed: consultations needed to take place across Canada, the Chief of the General Staff had to sign off on new policies, and the swift implementation of these policies was sometimes hindered by circumstances, such as necessary construction projects and waiting for the delivery of new codebooks from England. Throughout the Second World War, Canadian military and intelligence personnel worked hard to bring their communications security up to scratch so that Canada would not become the weak link among its Allies.

BONUS CONTENT: What is a Playfair cipher?

Earlier on this page, document CDWI00340 mentioned that R.C.N. used a Playfair cipher to encrypt wireless communications across Canada’s East Coast in 1941. Oftentimes, when we think about encryption in the Second World War, we think of cipher machines, like the German Enigma machine, but cipher machines were not widely used by the Canadian military in this stage of the war (1939-1941). Code books full of number-group substitutions for various common words and phrases were preferred at this time (e.g., ‘17435’ could mean ‘midnight’). The downside of this method of concealing important messages was that code books needed to be widely distributed (often coming all the way from London), and could be stolen by an enemy, in which case, the code became pointless to employ—and then, an entirely new code would have to be devised, new code books printed, and distributed.

Playfair, by contrast, is a substitution cipher dating back to the mid 19th century whereby pairs of letters in a plain language message are substituted for pairs of enciphered letters—and the rules for letter substitution are set by an easily changeable keyword (which has to be pre-arranged between sender and recipient). Playfair is a very simple cipher to learn, and by the Second World War, it was already known to have a number of weaknesses that could lead to an enemy breaking the cipher and retrieving the secret message. Even so, if the Canadians changed their keywords regularly without reusing them, it would have made the German codebreakers’ jobs harder, and using Playfair was certainly more secure than using no cipher at all!

Playfair can be done by hand, and you, my dear reader, can try it out for yourself, and gain a deeper understanding of how an R.C.N. wireless telegraph operator might have needed to prepare a message before transmitting it, or quickly decipher an urgent incoming message from an Examination Vessel in the Atlantic. The following BONUS section walks you through the Playfair cipher step-by-step, and at the end, there are a couple of decryption challenges—how fast can you transform Playfair enciphered messages to plain language?

How does Playfair work?

First, you need to select a keyword. For this example, “Canada Declassified” will be my key. The keyword is used to construct a five-by-five grid of letters from the Latin alphabet with no repeated letters, and usually equating I and J to get a nice even 25 (although you could choose a different pair of letters to share a grid space). I construct my grid by starting with the letters in the keyword, writing them left to right in rows of five letters—without repeating any letters!

Keyword: Canada Declassified

Start the table of letters like this:

The rest of the grid is then filled in with the remaining letters in the alphabet, in alphabetical order. Remember I and J have to share a space in the grid!

Here is the completed grid:

Now that we have completed our table of letters, we can start to encipher a message. The message I want to encipher is “an orange consists of soft yellow juicy atoms” (this phrase is from a real message the Canadians intercepted in October 1941, and you can read about it on this page).

There are four basic rules to remember for enciphering letter pairs with Playfair:

1. If the two letters are the same—or it is the end of the message and there is an odd number of letters, leaving one letter on its own—replace the second one with X (or add an X, in the case where there is no second letter), and then follow one of the rules below. Most of the time, X is used in the case of repeated letters, but some people choose to use another uncommon letter for this purpose.
2. If the two letters are in the same row of the table, replace each with the letter directly to its right, maintaining the order of the pair. If you reach the end of the row on the right, wrap around to the left.
3. If the two letters are in the same column of the table, replace each with the letter directly below, maintaining the order of the pair. If you reach the end of the column on the bottom, wrap back around to the top of the column.
4. In all other cases, the two letters will form a rectangle between them, and you replace the first letter in the pair with the letter in the same row, in the opposite corner of the rectangle. Then replace the second letter with the letter in its same row, in the opposite corner of the rectangle. Be careful to maintain the proper order of the pair!

Let’s encipher my example message! First, we break it up into letter pairs.

AN OR AN GE CO NS IS TS OF SO FT YE LL OW JU IC YA TO MS

To encipher the first pair of letters, AN, I look at the grid I created above. Because A and N are in the same row, I follow rule number 2) to substitute them.

C A N D E

Therefore, AN becomes ND.

The next pair is OR. This pair forms a rectangle, with the opposite corner letters being K and U, so we follow rule 4). For the first letter in the pair, choose the opposite corner of the rectangle in the same row. Then, for the second letter, choose the opposite corner of the rectangle in its same row.

K M O   (be sure to do O first, because it comes first in the message)
R T U

Therefore, OR becomes KU.

I won’t do every letter pair separately, but I will highlight a couple more examples. To encipher JU, you need to remember that I = J, and then we follow the rectangle rule. (When you are deciphering a message, you won’t know whether a letter is I or J, but you will have to determine that from context.)

I/J F B
K M O
R T U

Therefore, JU becomes BR.

To encipher LL, you first replace the second letter (which is repeated) with an X. Then, you follow the rules as usual. You can use I or J in this case to substitute for L, but I always prefer I because it is a more common letter. 

L S  I/J
G H K
P Q  R
V W X

Therefore, LL (LX) becomes IV.

Here is the whole message in plain language and then enciphered:

Plain language:   AN OR AN GE CO NS IS TS OF SO FT YE LL OW JU IC YA TO MS

Cipher message: ND KU ND OC EG AI FI QF MB BH MY ZD IV HZ BR LN WD UM HF

If we were really sending this cipher message out over wireless telegraph, we would need to put it into morse code in order to transmit it, but we can ignore that step here, since our modern computers so kindly parse binary dots and dashes into readable letters for us.

Decryption Challenges!

Now that you’re an expert in the Playfair cipher, it’s time to try your hand at decryption. In order to decipher Playfair messages, you only need to follow the above instructions backwards. (So, instead of shifting to the right in a row, go left. Instead of shifting down in a column, go up. The rectangles stay the same. Remove and replace any X, if you suspect a double letter is the culprit for its appearance. Etcetera!)

Here are a few secret messages you can try to work out on your own. How fast can you make sense of these urgent incoming messages?

The three messages below were all received today on September 17th, 1941. Today’s keyword is “Rockcliffe”.

What happened?
GF CZ FO BD HO LG XI CP MY FR DB DH IC TM DB MU IC EH QU IH LG    1400 K/C    KI PD KQ BC IS GC HB MY   1630Z/17TH    IA DF HU:

What are the Germans saying about our convoy??
“KC DZ KW GL SA XE CO SA PV ET PE DH KE SI ES MY MY  1245”  BH HV

What are we going to do about it?
XH DG BK KB YI BR BI CB HX KC KP IO KG MF TQ AS CO BH HI VW LQ QX AP SC    Q.S. 10    QK DF XI IM PD PA LG BK HX KC KP

Good work, reader! You can check your answers against CDWI00342.

If you had fun with this exercise, and you’re interested in a nice, readable introduction to some of the most well-known ciphers throughout history (and the process of breaking them!), I recommend you check out Simon Singh’s The Code Book.