Enigma Settings Sheets

Responding to some of Craig's follow-up questions from earlier this week:

-----Original Message-----
From: Bletchley Park discussion [mailto:BLETCHLEYPARK-0lvw86wZMd/[email protected]] On
Behalf Of Craig Heath
Sent: Monday, February 20, 2012 5:20 PM
To: BLETCHLEYPARK-0lvw86wZMd/[email protected]
Subject: Re: [BPARK] Enigma Settings Sheets

<snipped questions>

> I presume the settings sheet provided to Phil by Tony Sale was from the BP
archives, perhaps from a wartime "pinch"?

I also assume that it came from the BP archives, although I did not
specifically ask him at the time.  I do not know its provenance before that.


> How does one go about getting access to these primary sources?  I know
there's a digitisation project going on at BP:
> http://www.bletchleypark.org.uk/edu/archives/ ,
> I suppose I should just go and ask them, as I have the convenience of
living just an hour's drive away
> :-)  Does anyone know about getting access to the TICOM stuff?  It's many
years since I last visited the US National Cryptologic Museum, 
> do they hold those materials?

I had to get most of my material the old-fashioned way, poking about in
paper records in the US National Archives.  This was supplemented a great
deal by the generosity of other researchers who are acknowledged in the
article.  Back then, an awful lot of expensive and time-consuming
photocopying was involved.  At least these days both US and UK National
Archives allow the use of digital cameras.

Living in NJ, it is inconvenient and expensive for me to get to Kew.  I've
managed about 2 or 3 days there in the last 10 years.  I've been able to
visit BP a bit more often, but I have never visited the archive there.

I hadn't looked at the BP page Craig references in a while, and I see they
would now like a GBP25 donation to follow up on inquiries.  Since it's in a
good cause and cheaper than buying plane tickets, I think I'll give it a try
with some current questions I have.  A recent e-mail went unanswered, but
looking at the page I now see why.

US NCM does have some archives, and I believe David Kahn recently donated
most of his papers to them.  NARA of course has much larger holdings of
source documents, but NCM has been very good about letting me access their
machines.  In particular, all of my photos of UKW D and more recently of the
Kryha machine were taken in NCM's conference room.


>... I reread that part of your article and I see you refer to a document
that is held by the US National Archives and Records Administration,
> RIP 403 ... "Special British Reports on German Cryptography" - I don't
suppose that document is available online anywhere?  I'm particularly
curious 
> about how they would create sets of 12 reflector pairs from sets of 10
plug-board pairs, if anyone can explain further?

I do not believe this document is available online.  It is a collection of
BP memos sent to the US during the war.  The relevant memo here, dated March
9th, 1945, is the very last one in the file, and is too large for me to
retype in its entirety.  Here are a few relevant quotes:

- Repeats of stecker pairings, either whole or in part, as stecker for other
keys have been a feature of G.A.F. cypher procedure since 1942.  Since the
institution of the D reflector such sets of D pairings as have been
recovered have been compared together in the hope that repeats between them
will also occur.  So far, repeats have not been found.

- A day's stecker employs only ten pairs of letters and because of this
difference it was impossible to make comparisons in the stecker catalogue in
its old form.  A new catalogue has been compiled ...

- When considering this process two points should be borne in mind:-
(1) When constructing a D plugging J and Y must be paired together.
(2) For some unaccountable reason the G.A.F. cypher office have a
deep-rooted objection to pairing consecutive letters together, either in a
stecker or a D.

- The process [for transforming stecker pairings into D pairings] would seem
to be this:-
(a) Lift all pairs not involving J and Y directly into the D.
(b) The letters to which J and Y are paired in the stecker are paired
together in the D.  If these letters are consecutive ... The process is
altered - see (d) below.
(c) Take the unsteckered letters in alphabetical order and pair them 1 and
4: 2 and 5: 3 and 6.  These obviously cannot give consecutive letters.
(d) Where operation (b) gives a consecutive pairing take the offending
letters and pair them with the 1st and 4th unsteckered letters.  The
remaining unsteckered letters are paired 2 and 5: 3 and 6: as in (c) above.



So my previous speculation was slightly off.  Consecutive letters seem to
have been a G.A.F. obsession throughout, rather than just inherited from
their steckering habits.  Note that the descriptions in the memo use German
notation throughout (thus J and Y are the fixed B/O pairing observed in GCCS
notation).

Phil