The unfortunate life of the Army’s Cysquare cipher…

When I was trawling through WW2 pigeon / cipher-related documents at the National Archives, I found a brief mention of a medium grade Army cipher called “Cysquare”. I half-remembered the name and that it was (unsurprisingly) a cy[pher] based around a square, but couldn’t remember the context at all.

However, when I later searched for it on the Internet, it all started to come back to me… because Cysquare had been devised by my cryptological hero, master codebreaker Brigadier John Tiltman (‘The Brig’). Very much as the NSA PDF paean to him would have it, few would disagree that Tiltman was indeed “A Giant Among Cryptanalysts“.

Tiltman himself wrote a short introduction to his Cysquare (recently declassified, though brutally redacted) called “A Cryptologic Fairy Tale“, on the basis (a) that he’d lost the original material and had had to make up some examples, and (b) that it had a happy ending (which I guess was that the Allied Forces won the war). Here’s how The Brig described it (though with a few linefeeds added to make it acceptable to our modern short attention spans):-

Sometime later in 1941 I produced the “Cysquare” which was accepted by the War Office as a low-echelon cipher to replace the “Stencil” cipher and issued to the Eighth Army in North Africa, Figures 8 and 9 [both redacted, *sigh*] give photographs of two pages of the printed instructions.

The grille has 676 (26 x 26) squares. Each column and each line contains 10 white (permitted) squares, with the exception of 3 “plus” lines containing 20 white squares each and 3 “minus” lines which contain no white squares at all. The key for the day consists of 26 letters of the alphabet in random order with the numbers from 1 to 26 written under them also in random order. For each message the operator selects a 4-letter indicator from a random list of such groups provided him for use in turn. The indicator in the case of the example given [but redacted] is GMBX. The numbers corresponding to this indicator are 11 19 20 7, i.e., position 11, line 19, column 20, taking out number 7. The grille could be used with any of its sides at the top. Position II indicates that the grille is used as shown with numbers 8 to 13 at the top.

The numerical key for the day is written from left to right at the top of the grille and from the bottom upwards on the left hand side. The plain text is written into the grille starting at the next white square after the square described by the line coordinate 19 and the column coordinate 20, using the elements of the key to define the corresponding lines and columns. If and when the operator reaches the last white square in the grille he proceeds from the top left-hand corner. He then takes out the columns of letters starting at the top of the grille and in the column designated by the taking out number, i.e., in this case 7.

The message is written out in 4-letter groups preceded by the 4 letter indicator and followed by the number of letters, the indicator repeated, and the time and date. No message of more than 220 letters was permitted. If a message handed in for transmission exceeded this length it had to be divided into parts, none of them exceeding 200 letters in length.

However, somewhat contrary to Tiltman’s story’s name, Cysquare itself didn’t really have a happy ending. For a start, a number of people thought that cryptanalysts such as Tiltman shouldn’t be messing about with making their own cipher systems, and so there was a certain amount of resistance to it from within, right from the beginning.

The second problem was to do with implementation: even though it relied on disposable pads containing pre-printed grilles, somewhere along the line someone had the bright stupid idea that they could economize by getting the Army cipher clerks in North Africa to reuse the pads, by writing messages in pencil and then erasing them with a rubber. However, before very long it became impossible to tell a blank square from a dark square – everything in the grille ended up fifty shades or grey (so to speak). Hence the cipher clerks refused to use the system, and it was quietly abandoned.

However, when Germans captured Cysquare pads and implementation notes, their cryptographers rather liked it. And so in 1944, a new system started to appear in German messages: the Rasterschlüssel (also known as RS44), a system derived directly from Tiltman’s Cysquare cipher. Of course, Tiltman quickly recognized it: and had the Germans not made some mistakes when designing their pads’ grille designs, they might have been extremely hard to decipher.

So… was Cysquare a success, or a disaster? It was certainly quite secure (if somewhat awkward to use): but in the end, it nearly gave Germany a cryptographic edge late in the war.

For fans of the pigeon cipher story, it seems unlikely that its message used Cysquare… and so the search for that goes on.

PS: there’s an 2004 article in Cryptologia by Michael J. Cowan called “Rasterschluessel 44 — The Epitome of Hand Field Ciphers“.