22G Page 64
have occurred. A good example is a message giving a sheet of QKP numbers whose letter count is given in Fig. 22(IX). (R4 p. 16.)
It is unreliable to reject any letter count with significant bulges however oddly arranged these bulges appear to be. A few of the last German messages ever sent on Tunny gave some new wheel patterns and consisted almost entirely of the words NOCKE and KEINE separated by commas e.g.
(d) P counts on i-2 impulses.
The best Pi bulge is on P3 = x for punctuation (single or double) and on P5 = dot for language. Normally P1, P2, P4, P5 → dot and P3 → x but if 5's and 8's in P are very strong, they may be sufficient to negative the bulges on P1, P2, P4 and P5.
Fig. 22(X) shows the one and two impulse bulges for the 3 messages (type A, type B, type C) whose full counts are given in Figs. 22(VI), 22(VII), 22(VIII), and average bulges for a set of messages described in R5 p. 86.
|Fig. 22 (X)|
(e) ΔP counts on 1 and 2 impulses.
Bulges on ΔP1 are of interest only in the case of ΔP2 on messages with Χ2 limitation. ΔP2 → cross in messages strong in single punctuation. Double punctuation will normally cancel out the tendency for ΔP2 → x, but only rarely produces a comparable bulge on ΔP2 = dot.