General Report on Tunny

These consists of the wheel-sheets and most of the Colossus sheets of Mullet 25/4; and some miscellaneous exhibits. Mullet 25/4 is rather easier and more straight forward than the average wheel-breaking job. The margins of Colossus sheets have been drastically reduced.

(a) The rectangle (ch 24).

This was evidently a Garbo rectangle (24B(c)), but the Garbage is not preserved. At bottom right is the 9 x 9 flag and its convergence, used as a start for converging the rectangle (24D(c)). When converged the rectangle is easily significant; the 1+2 double bulge is 758 (or 759), 615 being sufficient according to the crude computery text; the leading term of significance test IV (24E(d)) is 258, so that it is unnecessary to calculate the ζ terms. Raw means made from a raw tape. (33B).

(b) Checks on Colossus.

The rectangle wheels, ΔΧ₁, ΔΧ₂ with low-scoring characters doubted (25D(a)), are set up on Colossus (wheels AA --- Figs II, III) and the score is checked: doubting reduces the double bulge to 705. The message is spanned (run ) in 200's for possible message slides: none is found. The two readings in each pair are 1 + 2 = ., 1 + 2 = x.

(c) Initial runs for ΔΧ₅, ΔΧ₄.

The first short run is C2, 5=/1=2. It is just significant (25B(a)): x = 5.1. Note the check ∑x_i= ν - 2x norm with a discrepancy of 2. The pencilled figures are the pippages (25a) for the various characters, i.e. score minus norm. The wheel ΔΧ₅A, heavily doubted, is set up (Fig VI). A bold run, 4= /5=1=2, form Χ₄ is comfortably significant producing wheel ΔΧ₄A (Fig V).