Hall V · American Field American Revolution · 1778–1783 Codebook (never captured, never broken)

Culper Ring / Tallmadge Code Washington’s spy ring codebook · 1779

A 760-entry pocket codebook that let Washington read New York Harbor for four years — and that the British never compromised.

DesignerMajor Benjamin Tallmadge, US Army (Washington’s intelligence chief)
YearIn active use 1779 – 1783
MechanismCodebook of ~760 numbered entries (people, places, words)
OperatorsAbraham Woodhull (Samuel Culper Sr), Robert Townsend (Samuel Culper Jr), Anna Strong
StatusNo British break of the cipher itself recorded

Why This Matters

By 1778 George Washington’s army had been forced out of New York City and the British held the harbour, Manhattan, and most of Long Island. Washington needed reliable, sustained intelligence from inside that perimeter — ship arrivals, troop movements, the names of ranking British officers in the city. He turned to a 24-year-old Continental Army officer, Major Benjamin Tallmadge, who in late 1778 organised what is now called the Culper Spy Ring.

Tallmadge’s solution to the secrecy problem was a small leather-bound codebook with roughly 760 numbered entries. Each entry was a person, place, military term, or common word. “Gen. Washington” became 711 and “New York” became 727. The codebook was distributed only to Tallmadge, Washington, Abraham Woodhull (Samuel Culper Sr), and Robert Townsend (Samuel Culper Jr). It was used continuously from 1779 until the end of the war — and no British break of the cipher itself is recorded. Surviving copies of Tallmadge’s codebook are in the Library of Congress.

How It Works

The codebook is a parallel-list nomenclator. The plaintext list runs alphabetically through the words and proper names of interest; each entry has a number from 1 to roughly 760. The cipher list runs in numeric order, each number against its plaintext, so the recipient can decode.

Operators wrote messages using ordinary English for low-value content and substituted code numbers for sensitive words — names of officers, ship counts, place names. Numbers were embedded in otherwise innocuous letters (“I saw 711 yesterday and he asked after the 486 in 727” — “I saw Washington yesterday and he asked after the people in New York”).

Because the word list is alphabetical, an entry's number tells you roughly where its word falls in the alphabet. Entries 1–710 are common words; 711–763 are the proper nouns. Read straight from the codebook:

CodeMeaning
178enemysits in the alphabetical run 175 embark · 176 employ · 179 example
355ladythe entry behind the "Agent 355" legend — a word, not a person
596secret
711Gen. Washingtonfirst of the proper nouns
727New York

One absence is worth noting, because it is often reported otherwise: the codebook has no entry for “spy” at all. Its s-words run 596 secret, 606 speaker, then straight on to the proper nouns at 756 Spain. A number widely quoted as meaning “spy” — 178 — is in fact enemy, and could not be “spy”: at 178 the list is still in the e's.

For deeper concealment, Culper messages were written in James Jay’s invisible ink (“sympathetic stain”) between the lines of cover letters. The ink itself was Washington’s second-most carefully guarded secret of the war — only a handful of people knew the recipe.

The demo above uses a stable 215-entry codebook indexed from 100. Words not in the codebook fall through to per-letter codes in the 800s, bracketed by sentinels 998 and 999 — the same trick Tallmadge would have used (he had a small alphabet table for proper nouns the codebook didn’t cover).

How It Was (Not) Broken
No British cryptanalytic break
Complexity: Hard without a captured codebook

A 760-entry random-numbered nomenclator in casually English-mixed letters is genuinely difficult to break by frequency alone, especially when sensitive words are used sparingly. The British signals organisation in New York intercepted Culper letters but could not read them. The ring’s secrecy was compromised exactly once, when courier Caleb Brewster was nearly captured — by accident, not cryptanalysis.

Operational risks
Complexity: N/A — the system’s real risk was capture, not codebreaking

A captured codebook would have been catastrophic: the British would have read every Culper message in its archive. Tallmadge enforced strict need-to-know discipline, kept his own copy with the army, and never named his agents in writing. The ring’s identities were not all public until Robert Townsend was identified by historian Morton Pennypacker in 1929 — 150 years later.

What It Teaches Modern Cryptography
Culper lessonModern echo
Codebook + invisible ink = layered defenceDefence in depth — multiple independent security layers
Strict need-to-know distribution of code materialModern key management: split knowledge, dual control, HSMs
Embedding ciphertext in plausible cover trafficSteganography and traffic-analysis-resistant protocols
Operational discipline beats algorithmic complexityModern incident-response: most breaches are operational, not cryptographic
Quick Facts
DesignerMajor Benjamin Tallmadge, US Army (Washington’s intelligence chief)
YearIn active use 1779 – 1783
MechanismCodebook of ~760 numbered entries (people, places, words)
OperatorsAbraham Woodhull (Samuel Culper Sr), Robert Townsend (Samuel Culper Jr), Anna Strong
StatusNo British break of the cipher itself recorded
Sources

This page previously stated that “spy” was encoded as 178. Checked against the codebook itself, 178 is enemy, and no entry for “spy” exists. The claim has been removed and the real values printed above.

← Previous Book Cipher