Some further notes on our poor plucky pigeon with his red Bakelite cipher payload.
People have been leaving some very interesting comments here, presumably spurred into action by today’s BBC piece on the mysterious bird. Moshe Rubin and Keoghly independently pointed out that “27” is a check figure to confirm that the sheet should contain 27 blocks of encrypted letters (as indeed it does), and both say that this was “standard fare for crypto“; while Keoghly adds that “1525/6 is likely [a reference] to [the] one time pad/offset used“. Bob Yeldham also helpfully notes that “the 37 & 40 on the rings denotes the year the bird was ringed that’s assuming the correct year ring was used, ringing was usually done 7 days after hatching.” Thanks to all of you!
What is also nice is that plenty of people seem to have rushed off to buy copies of Leo Marks’ very enjoyable book “Between Silk & Cyanide”, one of the best code-related books ever written, I’d say (even if some people have since asserted that Marks may have been an unreliable witness). Read it with care, for sure, but definitely read it! 🙂
I also ought to add my own thoughts from the last few weeks… and perhaps there’s an answer of sorts hidden in there.
My first reaction (that it might be a poem code sent from Douai early in the war) didn’t really stand up to analysis – having played with this in an Excel spreadsheet, I don’t think any kind of pure transposition is going to jump out from it. Though it resembles a poem code, Leo Marks specifically describes SoE’s sending cryptograms later in the war (particularly towards D-Day) that resembled the earlier poem codes, mainly to mislead Axis codebreakers: but by then Marks had (so he said) managed to persuade The Powers That Be to use one-time pads which were, to all intents and purposes, uncrackable then – or now, for that matter.
Having said that, what interests me (as an historian) is what we can tell from the rest of the message. For example, we know that the same message was sent back to the UK on two different pigeons, both of whose owners were members of the National Union of Racing Pigeons – “NURP 40 TW 194” (our dead pigeon) and “NURP 37 DK 76”. Yet we still have no word from the Racing Pigeon magazine (who presumably have all manner of racing-pigeon-related archives) on who those owners were, or where they lived.
Yet I think we can make a sensible guess at what happened here!
The message was addressed to X02, which (I believe) was Bomber Command in High Wycombe, Bucks (of which my aunt was once Mayor, incidentally), so it’s likely that we’re looking for locations in South-East England not too far from there. Hence I’ve already noted that our pigeon’s “TW” owner id could well have meant “Twickenham”, while similarly “DK” could very easily have been “Denmark Hill”. And we also know that the pigeon ended up in the chimney of David Martin’s house in Bletchingley (near Redhill) on its way home.
So let’s put it all into Google Maps and trace a possibly Twickenham-based pigeon backwards in time as the crow pigeon flies:-
If I’m right that “TW” is Twickenham (and it could just as easily be “Thomas Wilson”, I know, I know), then I think what we have here could be a pigeon flying back from Dieppe to Twickenham via Bletchingley. And one of the most [in]famous missions of WW2 was the Dieppe Raid of 19 August 1942, which went just about as badly as could be feared. So… could this have contained a desperate message sent back to Bomber Command to get help from the air on that dreadful occasion? Of all the possible scenarios discussed for our mystery pigeon’s journey, perhaps this is the one that has a (if you’ll forgive my pigeon-related pun) ring of truth!
Update: having said all that, if you follow the same line directly into the heart of France you hit Paris almost directly. Dieppe just happens to be the nearest point to Britain: all the same, until anyone knows any better, my pet name for this dead pigeon is going to be “Johnny Dieppe”, I hope nobody minds too much. 😉
As they sent 2 pigeons, are we thinking that the other pigeon successfully arrived? if so then the mission on the message was probably carried out and surely there will be a record somewhere on this?
Also it cant have been that important, surely, if they were relying on just two pigeons??
Narissa: you’re entirely right that I’d expect the other pigeon probably did get home safely. But where’s the fun in that? 🙂
I love your map, i was thinking along the same lines (no pun intended) but pop hooten park in and have a look at the line then.
I can’t see how that signature says “Stott” – I can only make out 2 t’s. Is it possible it’s Stf Sjt, i.e. Staff Sergeant?
Steve: it’s entirely possible, a number of sites read it as “Stot”. All of which is why I’d like to see a much higher-resolution image than the one we have, basically to eliminate all such ambiguities.
There was a story this summer that the driving reason behind the Dieppe raid was a failed attempt to capture an Enigma machine.
http://www.globalnews.ca/canada/feature/6442694158/story.html
Harold: fascinating, thanks! I missed that story at the time (though my wife now tells me she saw it but forgot to tell me), but better late than never! What a multi-layered cipher story this may yet prove to be! 🙂
My reading of the message, albeit probably completely wrong and almost certainly completely amaturish throws up a couple of interesting things.
The 1525/6 could just be a date and time. 1525 on the 6th of whatever month. Since pigeons fly 100 miles a day or thereabouts they’re likely to be where they’re going long before the 6th of the next month.
Lib 1625 makes sense in that respect. The pigeon was released an hour after the message was written.
It also ties in with the time of origin which is 1522, so the sender (blue) starts writing a message fills that field in at 1522 moves on to cyphering the message and is done in about 3 minutes.
Next the message is handed off and the pigeon tag numbers are noted (in grey) along with ‘lib 1625’ the time it was actually sent.
The block above is also the content of the message, so it makes sense that the date and the time are part of the message that would be decrypted then handed up to whoever was meant to get it.
Now as to the AOAKN – my guess would be this is the code used. And rather than start a message with START or END cyphered which leaks some information if there is a problem with your pseudo random code generation it makes perfect sense to just send the code again to indicate the message is complete.
It’s plaintext you’ve already sent, highly unlikely you’re going to randomly repeat the same 5 character block and it’s immediately followed by 27, which I think is just a checksum of the number of blocks sent.
It’s also likely that the same cypher systems would be used no matter what the actual transmission method. All of the above would work equally well if the same message were sent via radio, then a couple of pigeons get released as a backup. Finishing the message with the code and the number of blocks at least allows for partial recovery in the event that a radio message for example is received but is incomplete and you can’t ask for it to be retransmitted.
Here’s what I think; pigeon service message pads were carbon backed, the message received in the chimney is definitely the copy (it has a splodge on it and is harder to read / inconsistent ink in the coded area). See
http://cdn.thenextweb.com/wp-content/blogs.dir/1/files/2012/11/large-pigeon-message.png
The writer indicated that 2 copies were sent which would be both the original and the carbon duplicate which is why two pigeon numbers are listed… But he also writes ‘Late 16.22’ (zoom it to 200% and you can see it’s ‘late’ – remember it’s the carbon copy slightly rougher and looks like upside down now because now he’s in a hurry to actually send so the carbon changes colour).
From a ‘usability’ perspective it’s logical that the order of the pigeons when written down in a hurry, matched the order of availability for dispatch left to right and if stored in two separate pigeon boxes.
Therefore NURP.40.TW.194 got the original message and NURP.37.DK.76 got this one (and that’s our bird). DK could be for Dorking perhaps which is not far from Bletchingley.
I did think it could be a downed Lancaster (which would have carried two pigeons) see http://en.wikipedia.org/wiki/File:Avro_Lancaster_pigeons_WWII_IWM_TR_193.jpg.
It’s not an emergency message though, because they don’t get written at 15.22 but sent ‘late’ at 16.22, but it is a message that was ready to go at the right time and was then quickly re-time stamped and sent.
Finally it’s not ‘W Stot’, I think there is a letter missing before the final t (again that carbon transfer issue – more obvious at 400% view) and the military records database has no ‘stots’ I think it’s ‘W Stent’ but I can only find one and he was in the Indian army and pigeons don’t fly that far 😉
Sorry to cloud the waters…
What purports to be an archived WWII-era SOE silk onetime pad has the number 1578 printed on the top left ((http://www.duke.edu/web/isis/gessler/collections/silk-large-numbers.jpg), which may indicate that they were serially printed and the “1525/6” mark may indicate that pad 1525 was used, and perhaps line 6 of that pad? There would be no real danger of transmitting this info in clear, as without the specific pad the info would be useless and the sender was supposed to destroy their copy immediately.
Obviously the best placed to find the pad would be Bletchley/GCHQ, and if a onetime pad were used there is effectively no chance of recovering the message without it (unless you are prepared to wait a few trillion years…). Even with it, the appropriate WOK (worked out key) would probably be needed, as well as the four-letter prearranged short-codes sent with agents such as the Jedburgh teams (see http://www.campx.ca/georges-begue/silkcode.jpg for a frustratingly unreadable example), as without the latter XJHF might encode “enemy in sight” or “having tea and biscuits”!
Just a thought rather than Tunbridge Wells could the TW mean Twickenham??
Is it not the case that the repetition of ‘aoakn’ in the message is tremendously unlikely if it were encrypted using a one-time pad.
James: as a number of people have (rightly) noted, several war-time codes often placed some kind of key indicator as both the first and the last group of a message. AOAKN is therefore most likely a key indicator rather than a repeated message block. But all the same, that’s not 100% certain! 🙂
My thoughts are:
I cannot see a signature ‘Stot’
The handwriting is English round hand which I was taught in school.
This is a hastily written.
W.Flt Sgt
NURP 40 and 37 are the hatching years of the two birds.
The RAF pigeons numbered (by rings) in ‘Pigeons in the Great War’ were nearly all 2 years old.
27 is the number of groups of five letters
Half of the 27 groups have duplicated letters
Geoff: I don’t think it’s quite as easy as that – even if the signature is ambiguous, however you read it you then have to explain what it is you read… and as far as I am aware, “Sjt” was only used in the Army, not in the RAF. Moreover, different parts of the armed forces ‘recruited’ and used pigeons in quite different ways to serve their different needs, so it would also probably be wrong to generalise from one ‘wing’ to another. It’s a tricky old business! 🙂
Oh groan! Johnny Dieppe
Just goes to show when I last visited this particular discussion!
🙂