enough PSi' is recovered to give about 20 elements of each of the Psi
patterns these can be projected forward.or backward (or.both. if text is
available in both directions) to their next respective cycles and will overlap
sufficiently to give a small stretch of Psi key. This has to be slid back
and forth in an effort to find plain text because the exact number of
intervening dots in MT is unknown. Furthermore, since Psi' key, rather
than Psi is needed, additional assumptions must be made and tested as to
various possibilities of extension. .These.projected stretches of Psi key
are bordered by areas in which 4 of the Psi patterns are known so that,
after the stretch has been properly located and extended, additional text
can be found fore and aft on 2 generatrices (again with proper extensions)
This method can be used before enough of the Psi pat!terns are recoverefl to
overlap on all 5 wheels at the next cycle but with.the necessity of
reading on 2 generatrices superimposed upon the problem of locating the
stretch and determining the extensions it is even more difficult. In
practice what is done is to.look for a new and independent break at
approximately the right interval and to confirm it or expand it, or both,
with the aid of the pattern fragments. It should be observed that the
P5 limitation is a help rather than a hindrance in this type of work
because in working forward it limits the places where a motor dot has to
be tried and in working backward it sometimes determines the fifth plain
text irnpulse. This is true as against no limitation at all; the X2
lindtation alone would be evenmore of a help. In theory this method
could be used on a number of different messages and the Psi pattern
fragments thereafter combined but inpractice, if the fragments are large
enough to piece together successfully, it would almost necessarily be easier
to proceed by extending and projecting.as.described above. Actually the
solution was achieved wholly on one message.
5. The enclosure "Breaking the Psis.in the P5 Era" includes the
section's own description of the foregoing procedure but is devoted prirnarily
to the recovery of Psi patterns with the aid of a depth. When it was written
they had never actually been recovered without a depth.
6. It should be easily possible to use the machine for dragging
probable words to assist in the.initial steps involved in recovering Psi
patterns without depths. If a 10 letter word were dragged through a
de-chi the machine would be set to show all positions where 3 or more
repeats were yielded in the Psi' key, that is, it would stop if the Psi
patterns were contracted down to 7 or below. I have discussed.this
suggestion with Maj.Tester and Mr. Newman and they both feel it would
be a great help in getting started. Mr. Welhman advises that there
would be no difficulty in adapting the machine to accomplish this provided
it is borne in mind during construction. I calculate the number of