In the spring of 1942 the first shift system for a crib was discovered. This was a very important discovery which affected straight cribbing for the rest of the war. The reason for failing to make this discovery earlier was doubtless the same reason for our failing to notice in early days cribs of a type which were in regular use later; the obvious cribs were very good and we simply did not look very hard for the less obvious, a reprehensible but very understandable state of mind into which we seem frequently to have lapsed. The fact was that, when a key was breaking regularly and satisfactorily, there was very little incentive to do energetic research work.
The incentive which led to the discovery of the shift system was a deterioration of the Channel Weather cribs, soon to be finally killed. The position was that they were developing too many different forms and it was at this stage that it was noticed that certain forms occurred at regular intervals. The discovery was made on Cherbourg weather and was rapidly exploited elsewhere. This system involved 4 operators and I very much regret that all records of this original shift system have long since been destroyed. The interesting thing is that all later shift systems, some of which have been proved beyond a shadow of doubt, have worked on a 3-day cycle and have involved 3 operators and it would be interesting to re-examine the original shifts in the light of this evidence to see if our conclusions were only in part correct. The most common form of 3-day cycle, and the only one to be proved in detail, divides the day into 3 shifts (approx. 0-9, 9-16, 16-24) and works as follows: /BAC/ACB/CBA/BAC/ACB/ etc. As an example of the uses of these shifts we may take the twice daily weather report sent out to Arctic U Boats. This message stated the day and the month for which it was a forecast and the month appeared as a jumble of names and numbers, in fact 2 forms of the crib had to he run each time, one with April and one with VIER. On being divided into shifts on the above principle, however, it was discovered that one man said April consistently and the other 2 VIER so that on any day we knew which would occur. Also one man was security conscious and was responsible for nearly all the "horrors" and it was best to leave the crib severely alone when he was on duty.
The practical use of crib records, that is to say, the identification of cribs and the decisions as to what was and was not worth running always remained a job for high-grade labour, rather surprisingly, as in theory it is simple enough. An attempt to "mechanize" cribbing with the help of mathematical formulae was a lamentable failure and disappeared amidst howls of derision though in justice to the inventor of the system it must be said that this was not really due to the essential faultiness of the system as he proposed to use it, but rather the fact that it was misapplied by the inexpert and ignored by