crackciphers.com
[◑]

Compete to crack secrets lost to history—for fun, clout, and (no) profit


Links

Problems › 153 › solution › 1

A student’s two holiday messages under one homemade key

A catalog spelling concealed a pair

The first card was cataloged to “Gdna Mist,” but its physical address reads Edna Mist, the recipient of the second card. Both go to Piggott, Arkansas. Their similar signs and address hands make a comparison worthwhile; the repeated recipient and holiday art cannot by themselves identify a cipher key or a romantic relationship.[1]

Counting signs instead of apparent numbers

The signs resemble numerals, letters, and mathematical marks. Their appearance does not make them arithmetic. Native source review corrected the start of HCP1410 to ordinary Dec. 29, separated punctuation from a cipher sign, and recognized a hook-plus-dash as one composite character. Two supposed digraphs were actually separate signs. HCP1412 also extends to eleven rows: a first eight-row crop had left out three rows in the embossed lower field.[2]

Repeated runs fix a substitution

HCP1410 has the same twelve-letter run in its second and fourth rows. The recovery reads that run as COULD NOT COME; the surrounding rows produce MOTHER WAS SO SICK, I WILL NOT GET TO, and the school reference under the same one-sign-per-letter table. The first six-sign run is FRIEND. An early merger of its F-like first sign with the numeral-like 7 was corrected from the source, not retained as a second value for M.[4]

Recorded signReads
1A
CURL-C / CURL-SC
LONG-dD
2E
F-like signF
FILLED-OVAL (HCP1410); CLASS-02 (HCP1412, archived label G)G
H-like signH
3I
K-HOOKK
6L
7M
8N
4O
9R
EQS
DIVT
U-COMPU
OBW
X-like signX
y-like signY

The source reader described HCP1410’s relevant sign as a filled/slashed oval and used the label G in the HCP1412 source stream. The observed-key record assigns G to the former, while the released HCP1412 occurrence readings give the latter that same value. The aliases FILLED-OVAL and CLASS-02 keep those two source notations visible. Their common reading is part of the recorded solution; this editorial crosswalk does not independently establish that their pen strokes are identical.[4]

The observed table covers 20 letters. B, J, P, Q, V, and Z do not occur and remain unassigned. These are the observed key entries, not a reconstructed full alphabet for hypothetical further cards.

HCP1410: every address-side row

Physical rowRecorded signsReads
R1[open date: Dec. 29] | F 9 3 2 8 LONG-d : - | 3FRIEND:- I
R2CURL-C 4 U-COMP 6 LONG-d | 8 4 DIV | CURL-S 4 7 2 | X 7 1 EQCOULD NOT COME XMAS
R37 4 DIV H 2 9 | OB 1 EQ | EQ 4 | EQ 3 CURL-C K-HOOKMOTHER WAS SO SICK
R43 | CURL-S 4 U-COMP 6 LONG-d | 8 4 DIV | CURL-S 4 7 2I COULD NOT COME
R53 OB 3 6 6 | 8 4 DIV | [FILLED-OVAL] 2 DIV | DIV 4I WILL NOT GET TO
R6CURL-S 4 7 2 | DIV 3 6 6 | 7 y | EQ CURL-S [H: source-damaged] 4 4 6COME TILL MY SC[H]OOL
R73 EQ | 4 U-COMP [DIV with an attached descending slash/flourish]IS OUT[terminal flourish/punctuation]
Dec. 29
FRIEND:- I
COULD NOT COME XMAS
MOTHER WAS SO SICK
I COULD NOT COME
I WILL NOT GET TO
COME TILL MY SC[H]OOL
IS OUT[terminal flourish/punctuation]

The picture-side border adds the ordinary sentence I would like a line from you all. Scattered faint strokes over the picture’s blue field cannot be organized into a dependable stream and are not included in this reading.

HCP1412: the unchanged key predicts a whole card

The exact FRIEND ciphertext recurs at the end of HCP1412’s first row. The HCP1410 table then reads all 101 visible letter cells of HCP1412 without a writer-error repair or a new entry. The SCHOOL, COME, XMAS, and FRIEND strings check transfer across the two cards. GOING repeats twice within HCP1412, providing an additional internal check on that card.

RowRecorded signsTransferred reading
R17 y LONG-d 2 1 9 | F 9 3 2 8 LONG-dMY DEAR FRIEND
R23 1 7 | CLASS-02 2 DIV DIV 3 8 CLASS-02I AM GETTING
R31 6 4 8 CLASS-02 | F 3 8 2 | 3 8ALONG FINE IN
R4EQ CURL-S H 4 4 6 | 3 | 1 7SCHOOL I AM
R5CLASS-02 4 3 8 CLASS-02 | DIV 4 | CURL-S 4 7 2GOING TO COME
R6LONG-d 4 OB 8 | DIV H 2 9 2DOWN THERE
R7X 7 1 EQ | 3 OB 4 U-COMP 6 LONG-dXMAS I WOULD
R86 3 K-HOOK 2 | DIV 4LIKE TO
R9K-HOOK 8 4 OB | 3 F | y 4 U-COMPKNOW IF YOU
R101 9 2 | CLASS-02 4 3 8 CLASS-02ARE GOING
R11DIV 4 | CURL-S 4 7 2TO COME

Historical reading and independent check

The earlier card, postmarked 24 November 1908, reports that the writer is getting along well in school and plans to come down for Christmas. The later card says the writer’s mother was too ill for a Christmas visit and that school must finish before another visit. An independent source reader checked the native occurrences; the complete transfer has no plaintext mismatch. HCP1410’s H in SC[H]OOL remains physically damaged, supported by surviving strokes and the clean second-card H rather than claimed as unobscured ink.[4]

What is not claimed

The two address-side messages are read; HCP1410’s picture-field marks remain unresolved. Its cancel shows 1909 but a weak month resembling NOV conflicts with open Dec. 29 if both dates belong to that year. No exact mailing day is promoted. The source suggests a shared writer and system but does not identify the writer, age, gender, school, or a courtship. “Friend,” school, illness, and “you all” establish a bounded family and household context.

Both portal records were “Not solved” with null keys on 29 August 2026. Bounded ID, recipient/address, phrase, public-cipher, and local-index checks found no prior exact-item solution, operative key, or further Edna Mist sibling. This is new to those checked sources, not proof that the messages had never been read.[3]

References