These transmissions were picked up by Tromsoe and retransmitted in Enigma to Wilhelmshaven on a special frequency created for the purpose. Before being recyphered in Dolphin, these messages in L.M.T. were decyphered and then recyphered in another hand cypher: the result was cyphered on Dolphin with information as to the originator, and time of origin and serial number of the L.M.T. message. Unsuccessful attempts were made to break both the L.M.T. and the second hand cypher with a view to obtaining reencodements, but even though this was unsuccessful, excellent cribs were available. Most of the reencyphered weather message consisted of 35 letters, separated into 5 letter groups by commas; the first 5 letters were predictable and the last 7 almost invariably XX, XXXXX. As these letters were spelt out in the form XANTXANT etc a 29 letter crib resulted. At a later stage this spelling out of the letters was abandoned and a more abbreviated form of preamble adopted; the message started with the time of origin and serial number of the L.M.T. original, followed by the 35 letters of text, the first 5 and last 7 of which were predictable. This brought a great restriction in length which made cribbing much easier; a 25 group message had to be BASSGEIGER as he alone had a 2 figure serial number and a name short enough to make this possible; a 25 group message from Bassgeiger was impossible. One of the worst difficulties was intercepting the L.M.T. transmissions so that we could use their times of origin as cribs but satisfactory results were eventually obtained by quadruple-banking the frequencies and enlisting the assistance of R.S.S. The crib died when it ceased to be reencyphered on Dolphin, appearing on some other cypher instead; it ranks with the Biscay Weather, Bereich 7, Zustand ostwaertiger Kanal and one or two others as one of the most remarkable cribs we ever possessed; during the period in which it was available, we ran on an average only 21 cribs a month to break the 15 wheel orders of each month's Dolphin keys: similar statistics for other periods are not available but it is most improbable that this performance was equalled at any other time.
March 1944 passed without further cryptographic event with the exception of the arrangement of a meeting between 2 U Boats, one of which was to receive from the other a copy of new bigram tables, known as Quelle. As the recipient was expected home in the middle of May, May 1st was thought a likely date for a change of tables though the possibility of an earlier change could not be ruled out; in any case a change was certainly due, the current tables having been in force for a year. Plans were drawn up for tackling the problem, now made somewhat more complicated by a multiplicity of keys but simplified by the increasing traffic. A change of May 1st would have suited us well enough as it would have left plenty of time to get clear before the opening of the Second Front early in June; in fact no change took place until July 15th by which time Quelle had been captured and we suffered no inconvenience.
In April the Mediterranean U-Boats ran out of keys and Turtle proceeded to reuse March keys with a different Stichwort. Turtle had been allowed to lapse in February but work on March was started as soon as this news was received with a view to obtaining free re-