[In the original document the following text is written in a hand other than Turing's and is attached to page 143.]
This process worked well initially. The popular trigraphs were at the top of columns on the centre pages of the K-book, but the German instructions were to mark any trigraph as it was used, and not to re-use it. Thus the repeat rate of the new trigraphs with those known to have been used gradually dropped. The part of the K-book used for this trigraph selection contained all trigraphs in hatted order. I. J. Good in 1941 devised a quicker method, using the non-randomness of the free-choice dummy letter. JELM 8/2/78
Back to Turing's Treatise on the Enigma. Chapter VII.