Ancient Cybersecurity III: From Greek Fire-signalling to WWI Code-crafting

Martine Diepenbroek

Very early on in Greek history, extant sources reveal that mountain tops were clearly used both as watchtowers and as signalling stations from which secret and non-secret messages could be sent over long distances by lighting strategic fires. There are numerous extant references to this strategy for long-distance communication from the ancient world, in works not only of history but of literature too. To take just one famous example, Aeschylus – in his Agamemnon – described how Clytemnestra received word from Agamemnon that he was returning from Troy by means of beacons – signals that were speedily and efficiently sent from one hilltop station to the next.^[1]Aeschylus, Agamemnon 281–3. As it was only possible to send one prearranged message using this archaic system of fire-signalling, communicating parties needing to communicate on urgent matters faced significant limitations.

The 4^th-century BC Greek general and military author Aeneas Tacticus (“the Tactician”) accordingly invented a more useful system for sending more than one message by using water clocks and torches.^[2]Aeneas Tacticus’ system is discussed in a now lost work on military preparations, but its basic design has been handed down to us via Polybius at Histories 10.44. This passage, and others from Polybius cited in this article, can be read in English and Greek here. Communicating parties would fill large vessels of the same size with water. Then they would take a rod graduated into sections. In each section they would write the most evident and ordinary events that occurred in war (e.g., “Enemy approaching with ships from the west”).

The rod was placed into the vessel. When the two distant parties wanted to send a message to each other, one would first wave torches to attract the other’s attention. Then both groups would pull out corks from the vessels to let the water run out, which would cause the rods gradually to sink. When the intended message on the rod had descended to reach the top of the vessel, the sending party would wave a torch again to let the receiving party know that they had to put the cork back into the vessel and check the intended message. If all worked well, the receiving party could then read off the intended message.

What Aeneas Tacticus described was a highly inventive mode of long-distance communication that was a large improvement over the sending of one pre-arranged message. However, there are two significant downsides to Aeneas Tacticus’ method. First, it would have been extremely difficult to let the two water clocks run exactly parallel. Because of this it is hard to believe that Aeneas Tacticus’ mechanism ever functioned with perfect precision. What is more, with this method it remains the case that only pre-arranged messages could be sent – even though there was now a series of possible messages to choose from.

Polybius noticed these defects in Aeneas’ system. He stated that it would have been impossible to communicate by using Aeneas Tacticus’ method if anything unexpected occurred.^[3]Polybius, Histories, 10.45.1–5. Therefore, Polybius decided to improve upon this method by developing it into a more sophisticated system of fire-signalling – a system capable of dispatching with accuracy any kind of message.

Polybius described how both communicating parties had to take five tablets. On the tablets one would write all letters of the Greek alphabet from alpha to omega. They would also need two sets of five torches.

The dispatcher of the message would raise the first set of torches on the left side indicating which tablet was to be consulted (from 1 to 5). Then he would have raised the second set of torches on the right to indicate what letter on the tablet the receiver should write down.^[4]Polybius, Histories, 10.45.6–12. See for a parallel from Roman times Julius Africanus, Kestoi, 77.

Like Aeneas Tacticus, Polybius still used torches, but replaced the water clocks with tablets on which the letters of the Greek alphabet were written. One could then send messages letter by letter, whereby each fire signal represented one letter. Although Polybius’ method was still extremely laborious, it was clearly a significant improvement over Aeneas Tacticus’ system, since Polybius did not require a pair of water clocks to run in parallel. What is more, when using Polybius’ system any possible alphabetic message could be sent between communicating parties, instead of only a series of pre-arranged messages. Since only sender and receiver knew how Aeneas’ and Polybius’ method worked, it was perfect for transmitting secret messages, especially when the enemy was nearby.

The Polybius Square

Polybius’ system with the tablets has been developed throughout the ages into the so-called ‘Polybius square’, a method still used in modern communication security. A basic Polybius square consists of five rows and five columns, which gives 25 cells. In these cells the 26 letters of the modern English alphabet are written in their normal order from left to right, and top to bottom. To make the requisite number, the letters ‘I’ and ‘J’ are usually placed in the same block.

All rows and columns in the square have a number. In a basic square these are the numbers one to five for both rows and columns; every letter in the square thus gets a coordinate. The letter ‘A’, for example, can be found in the first row on the first column, which gives the coordinate 1-1, written as ‘11’. In this way, all the letters in the square have a coordinate between ‘11’ (A) and ‘55’ (Z). The coordinates can be compared to the place of the letters on Polybius’ tablets.

A modern message that is sent by a Polybius square looks like a series of numbers. The message “Send more troops before midnight,” for example, would appear as the following numerical sequence:

43 15 33 14 32 34 42 15 44 42 34 34 35 43 12 15 21 34 42 15 32 24 14 33 24 22 23 44

Since every coordinate contains two numbers (one for the row and one for the column) an encrypted text is created that is twice as long as the non-encrypted text.

To decipher the message, the recipient would take a Polybius square, look for the coordinates in the square, and check which letters correspond to these coordinates. The Polybius square has reportedly been used for simple cryptographic communications of this kind by the British army in the Boer War, and by both the British and German armies in the First and Second World Wars.

The ADFGX- and ADFGVX-ciphers

In 1918, the last year of the First World War, the system of the Polybius square was developed even further by the German military intelligence services. They invented the ADFGX- and ADFGVX-ciphers named after the only alphabetic letters that appeared in the ciphertext. Messages encrypted with the ciphers were transmitted by Morse code. The six letters were chosen to minimise transmission errors, since the letters sound very different from one another in Morse code.

In March 1918 the first system was introduced: the ADFGX cipher, which used a Polybius square of 5×5. This square was filled with 26 letters of the modern alphabet in random order agreed upon by sender and recipient. Again, the letters ‘I’ and ‘J’ were placed into one block.

The rows and columns of the Polybius square used for the cipher were then labelled with the letters ‘A’, ‘D’, ‘F’, ‘G’, and ‘X’ to produce a coordinate for each of the plaintext letters:  in this case, for example, the plaintext letter ‘Y’, was encrypted as ‘XF’:

In this way, a ciphertext was created that was twice as long as the plaintext, and that only contained the letters ‘A’, ‘D’, ‘F’, ‘G’, and ‘X’. The plaintext message “Send weapons quickly”, for example, would have been substituted into the following ciphertext:

GA DD AD FX FG DD AA DX FF AD GA DA DF AX GD FA FD XF

In June 1918, three months after the introduction of the first ADFGX cipher, the Germans added an extra row and column to the Polybius square to create a 6×6 grid which increased the level of sophistication of the method. Extending the grid meant that an extra letter was required to create the ciphertext. The letter V was chosen for this, since this letter, again, sounds different from the five other letters in Morse code. The newly created cipher was called the ADFGVX cipher and worked in the exact same way as its predecessor. Both the ADFGX and ADFGVX ciphers were the most advanced cipher systems that the German military intelligence used during the First World War. In fact, they proved to be amongst the toughest ciphers ever used in military secret communication until after WWI.

---

It is a big step from the oldest simple fire-signalling systems in Ancient Greece to the far more advanced ADFGX- and ADFGVX-ciphers used in WWI. Yet, the work of Aeneas and Polybius on fire-signalling did indeed form the basis for the ciphers that were used 2,000 years later. Moreover, Polybius squares are still used in communication security up to the present day. Yet another surprising place in which the influence of the Greeks and Romans survives in the 21^st century.

---

Martine Diepenbroek is a Dutch Ancient Historian who has recently completed a PhD in Classics and Ancient History at the University of Bristol. Her research focuses on ancient cryptography – especially on the Spartan scytale. She has previously published in KLEOS (The Amsterdam Bulletin of Ancient Studies and Archaeology) and in Ancient Warfare Magazine. She is currently in the process of publishing her PhD thesis with Bloomsbury (London). She has previously written for Antigone on the Spartan Scytale and the Caesar Cipher.

---

Further Reading

V. Aschoff, Geschichte der Nachrichtentechnik: Beiträge zur Geschichte der Nachrichtentechnik von ihren Anfängen bis zum Ende des 18. Jahrhunderts ( Springer-Verlag, Berlin, 1984).

M.L.M. Diepenbroek, QFrom Fire Signals to ADFGX: A case study in the adaptation of ancient methods of secret communication,” KLEOS: The Amsterdam Bulletin of Ancient Studies and Archaeology 2 (2019) 63–76.

W.W. Hyde, “The Mountains of Greece,” The Bulletin of the Geographical Society of Philadelphia, 13 (1915) 1–16, 47–64, 110–26.

R.M. Sheldon, Intelligence Activities in Ancient Rome: Trust in the Gods, but Verify (Routledge, London, 2005).