In Fig. 78, we have a “checkerboard” which, primarily, is intended as a transformation device; that is, a means for replacing single letters with syllables, and, consequently, for replacing five-letter incoherent groups with ten-letter pronounceable groups; under the European agreement, the price of transmission is the same
| Figure 78 L N R S T b a t t a l i o n A C U L P E ER NE UL LU NE AR RI OR NO E R A B D F I G H I J K O M N O Q S Regrouped: U T V W Y Z ERNEULLUNE ARRIORNO. |
for both, and the pronounceable groups are less likely to result in transmission errors. The alphabet is first reduced to 25 letters (in this case by the omission of X), and is written into a 5 x 5 square. The five vowels, written at one side, will then serve to designate the five rows, while five other letters, written across the top, will designate columns. Any letter found inside the square may thus be pointed out by naming the two letters which will indicate its column and row. In the given example, A can be replaced with EN or NE; T with UL or LU, and so on.
The fact that two interchangeable substitutes have been provided for each letter of the alphabet has led many persons to use this device, absolutely without modifications, as a simple substitution key. Yet it must be plain that any decryptor, taking his preliminary frequency count, will discover, before going very far, that this count is being made on only ten different letters, and thus can represent only one possible kind of encipherment. A frequency count taken on the pairs, with no distinction made between a given digram and its reversal, will afford the necessary proof; after that, the average decryptor will usually replace the pairs with single letters (or numbers), just as he would in dealing with printers’ symbols, or other inconvenient characters. The checkerboards which are actually intended for encipherment purposes ordinarily use digits for pointing out columns and rows. Where the digits at the side are the same as those across the top, it becomes necessary to observe an order, as column-row, or row-column, and this, using only five digits, is ordinary simple substitution, in which every letter has one substitute. But if the five digits at the side are different from the five written across the top, then the order is immaterial, and any number may be interchangeable with its reversal; that is, 17 or 71 can represent the same letter.
This encipherment might not be spotted so promptly as the case in which only ten letters are present out of a possible 26. But if the count is made on a chart, as recommended at the beginning of the chapter, it is very readily detectible that there are two separate groups of digits, neither one of which has ever formed any combination within itself, every number in the cryptogram being composed of one digit from each group. Thus we see plainly the trail which is left by co-ordinates.
| Figure 79 The KEY-PHRASE Cipher (a) Plaintext: a b c d e f g h i j k l m n o p q r s t u v w x y z CIPHER: O N E W H O H A S P A S S E D O N I S A M O N G U S (b) CIPHER....... O M S S May represent: A U I I Full F L L Fuss P M M Fuzz V S S Pull Z Z Puss |
Checkerboards, of course, can be used to better advantage. But, before leaving the simple for the complex, we must not overlook the celebrated key-phrase cipher, which discards the idea of multiple substitutes in favor of multiple originals! This cipher, shown in Fig. 79, is said to have been used for serious purposes. Its only difference from the ordinary simple substitution lies in the nature of the cipher alphabet, which must be a plaintext sentence, or phrase, containing the necessary 26 letters. The mysterious pronouncement, “One who has passed on is among us,” is the earliest example of which the writer has any recollection; those of later years have been largely proverbs, or other familiar sayings: “Journeys end in lovers’ meeting”; “Prosperity is just around the C.” As any cryptogram-letter may have five or six different originals, it is readily understood why the cryptograms of the key-phrase cipher are seldom seen without their word-divisions; yet, curiously enough, their translations are almost never ambiguous.
As to their decryptment, the student who cares to try the appended example will find that it is hardly more difficult than one of the simpler “aristocrats.” The method is about the same for both, keeping in mind that the frequency shown by any cryptogram-letter is either the frequency belonging to one letter or the exact sum of the frequencies belonging to several. Here, however, the reconstruction of the key simultaneously with the identification of substitutes is a very important adjunct to solving; the cipher-alphabet, being pure plaintext, can often be built up long in advance of solution. It might be added that this cipher, with or without word-divisions, is readily distinguished from all others by the make-up of its frequency count, which, as a rule, consists chiefly of the high-frequency letters in unusual numbers.
Passing now to the more difficult cases, we will glance at a few of those ciphers which are truly multisubstitutional; that is, which provide multiple substitutes for all or most of the plaintext letters. This is usually accomplished by the use of two-digit numbers, of which one hundred are possible: 01-02-03. . . . . .98-99-00. These one hundred numbers may be assigned as substitutes to the twenty-six letters, in proportions roughly approximate to their normal frequencies, as suggested in Fig. 80; or most of them may be so assigned, and the rest reserved as substitutes for digits, punctuation, and so on. For security, however, they must never be assigned in regular order, as in the figure, or even by any methodical process, but absolutely in incoherent order. Thus, while the form indicated in Fig. 80 will be
| Figure 80 A 11, 12, 13, 14 B 15 C 16, 17 D 18, 19 E 20, 21, 22, 23, 24 (Etc.) |