This latter can be seen by looking at the cryptogram (e), where all segments, as soon as selected, have been circled out. In finding a column which would complete the very evident word SYSTEM and, at the same time, furnish a letter suitable to precede HA, we find that this is the end-segment of the cryptogram, and would leave only two letters — far fewer than the number needed for furnishing another column.
At (f), we have extended the rest of the columns by two (and one) letters, except that there is a gap in sequence on the next-to-last line. At (g), we have transferred the letter which will fill this gap, leaving a misplaced X at the top; and, at (h), we have placed this X where it belongs and are now ready to transfer the two misplaced columns and recover the key. This key, as before, is found by numbering the segments of the cryptogram, and assigning these key-numbers to the correct columns in the adjusted block. It is usually possible to go further, and learn the long words on which such keys might have been based.
Concerning digram-tests, Ohaver suggests another which is more quickly made than the frequency test, and which the writer, so far, has found fully as reliable. Using “C” for “consonant” and “V” for “vowel,” he speaks of this as his VC-CV or “mixed” test. A digram like HA is a cv digram, one like AT is a vc digram, and others are vv and cc digrams. His theory is this: Since almost two-thirds of the digrams used in the language will be of mixed formation, that is, either vc or cv digrams, it stands to reason that the set-up containing the largest number of “mixed” digrams would probably be the correct choice. The student may look it over in Fig. 45.
As to possible variations, a cipher with a new name is not necessarily a different cipher. Fig. 46 shows a cipher originated many years ago by the cryptologist
| Figure 46 The Myszkowsky Cipher (a) Keyword: CURTAINS C U R T A I N S C U R T A I N S C U R T A 4 19 11 16 1 7 9 14 5 20 12 17 2 8 10 15 6 21 13 18 3 R E G R E T C H A N G E I N S Y S T E M S Cryptogram: E I S R A S T N C S G G E H Y R E M E N T. (b) Keyword: PARADISE P A R A D I S E P A R A D I S E P A R A D 14 1 17 2 7 12 20 10 15 3 18 4 8 13 21 11 16 5 19 6 9 R E G R E T C H A N G E I N S Y S T E M S Cryptogram: E R N E T M E I S H Y T N R A S G G E C S. |
E. Myszkowsky, and advertised by its inventor as non-decryptable. The key-word here is repeated often enough to furnish one key-letter for each text-letter, nulls being added, when necessary, to prevent the complete unit which would result if key-word and text were allowed to end at the same point. This long series of key-letters is then treated as a single word, and is converted to a numerical key in the usual way, all A’s receiving the first numbers, all B’s the next numbers, and so on. The message of the figure is very short: REGRET CHANGE IN SYSTEMS. Try enciphering this in the ordinary columnar transposition, using first the key-word CURTAIN, which contains no repeated letters, and afterward the key-word PARADISE, which has a repeated letter A. In the second case, what happens to the two columns belonging to the A-numbers? Suspecting a Myszkowsky encipherment, how could you go about unscrambling the two? Suppose there were three?
Fig. 47 plays another variation on the columnar theme. This cipher, originated by a member of the American Cryptogram Association, follows the rules of columnar transposition in all respects except that pairs alternate throughout with single letters (The text is: CHIEF WANTS YOU TO INTERVIEW SMITH). Can you pick out at a glance the really vulnerable feature of this cipher, and formulate a special method for its solution?
| Figure 47 The "AMSCO" Cipher (A.M.Scott) R A C K E T 5 1 2 4 3 6 CH I EF W AN T S YO U TO I NT ER V IE W SM I T HX Order for taking off: I YO V HX / EF U IE /... Cryptogram: I Y O V H X E F U I..... |