About History: Testudo

Etymology The Roman testudo means “tortoise” not “turtle”, even if Rex Harrison, playing Julius Caesar in the 1963 film “Cleopatra”, uses the latter term. To be fair, he was delivering a line from script written by an American, which may sound a little condescending until one realises that in North America “turtle” is used to denote the whole group of creatures. In fact, the order’s name, Testudines, is based on the Latin word testudo meaning “tortoise” and was coined by German naturalist August Batsch in 1788. The word “turtle”, however, is borrowed from the French word tortue or tortre which is seemingly used interchangeably to mean either “turtle” or “tortoise”. In Britain, where tortoises and turtles are not native, “turtle” is reserved for sea turtles as opposed to freshwater terrapins and land-dwelling tortoises.

The ancient Roman army used the term testudo because their soldiers’ shields formed an all-round shell-like defence reminiscent of a tortoise. The front rank of men, possibly excluding those on the two flanks, held their shields from about shin height to their eyes, creating a contiguous “wall” protecting the formation's front. Those men in the succeeding ranks would raise their shields above their heads using the shields to not only protect themselves but also the head of the man preceding them. There has been some debate as to whether shields were underlapped as shown in the colour image above or overlapped as shown in the black and white image from a relief panel on Trajan’s Column in Rome. In fact, contemporary imagery from the Roman period depicts shields being used in both styles so either manner is attested and thus correct.

From experience overlapping shields is easier to execute, even while advancing, but critics have claimed that arrows, for example, could breach the formation passing under the shield rim. As can be observed in the accompanying image of a small testudo, gaps in the overlapped shields are minimal. The biggest danger is the much larger opening where the front rank men’s heads can be seen. While one man has to keep watch to “steer” the formation, the threat to the others can be minimised by tilting the head downward so the crown of the helmet is presented and the face protected. From bitter experience, a well placed arrow can still penetrate the testudo, but such shots are the exception.

The men along the formation’s flanks, however, would present a shield wall to either side of the testudo. If necessary, the soldiers at the rear of the formation could stand backwards with their shields held as the front ranks to protect the formation's rear. In this manner the testudo offered all-round defence against opposing infantry and excellent protection against arrows and other missile attacks, albeit at the cost of reducing the speed and mobility of the formation.

Battlefield use The two principal uses of the testudo most often quoted are as a defence against enemy missiles or projectiles, or to close on defended positions or fortifications during sieges. The former was most likely employed by stationary troops as Cassius Dio implies in his description of Mark Antony’s campaign in 36 BC:

“This testudo and the way in which it is formed are as follows. The baggage animals, the light-armed troops, and the cavalry are placed in the centre of the army. The heavy-armed troops who use the oblong, curved, and cylindrical shields are drawn up around the outside, making a rectangular figure; and, facing outward and holding their arms at the ready, they enclose the rest. The others, who have flat shields, form a compact body in the centre and raise their shields over the heads of all the others so that nothing but shields can be seen in every part of the phalanx alike and all the men by the density of the formation are under shelter from missiles. Indeed, it is so marvellously strong that men can walk upon it and whenever they come to a narrow ravine, even horses and vehicles can be driven over it.” (Cassius Dio, “Roman History”, XLIX, 30)

Dio does not suggest that Anthony’s troops were moving but the testudo can be used on the march. Because of its density or compact nature of the formation, the primary drawback was that it was more difficult to fight in hand-to-hand combat and because the soldiers are required to move in unison, speed was sacrificed. Yet it is documented that the steadily advancing testudo was used in siege warfare to close on and storm a defended position. Tacitus, for example, recorded its use during the siege of the city of Cremona by the troops commanded by Marcus Antonius Primus, part of Titus Flavius Vespasianus’ army. During the attack the troops advanced under the rampart “holding their shields above their heads in close ‘tortoise’ formation” (Tacitus, “Histories”, III, 27).

The testudo was not invincible. Once again Cassius Dio provides an account of a Roman testudo being defeated by Parthian cataphracts and horse archers at the Battle of Carrhae in 53 BC:

“For if [the legionaries] decided to lock shields for the purpose of avoiding the arrows by the closeness of their array, the [cataphracts] were upon them with a rush, striking down some, and at least scattering the others; and if they extended their rank to avoid this, they would be struck with the arrows.” (Cassius Dio, “Roman History”, XL, 22)

Forming a testudo severely restricted the Romans ability to engage in melee combat. The Parthian cataphracts exploited that weakness and repeatedly charged the Roman line, which caused panic and inflicted heavy casualties. When the Romans tried to open their formation to repel the cataphracts, the latter rapidly retreated, and the horse archers resumed shooting at the legionaries, who were now more exposed. In effect the Parthians were able to pin the soldiers of Marcus Licinius Crassus’ army in place and systematically engage them with cavalry charges and withering volleys of arrows from a distance.

Mythbusting As an aside, after the defeat at Carrhae an estimated 10,000 Roman prisoners of war appear to have been deported to Alexandria Margiana (Merv) near the Parthian Empire's northeastern border in 53 BC. Employed as border guards they reportedly married local women. Nearly twenty years later the nomadic Xiongnu tribal confederation [1], led by their chieftain Zhizhi Chanyu, had migrated east and established a state in the Talas valley, near modern-day Taraz, Kazakhstan. Fearful that Zhizhi was planning to build a great empire, in 36 BC the Chinese Han Dynasty launched a pre-emptive attack and besieged Zhizhi’s wooden-palisade fortress. According to the Chinese account of one Ban Gu, during the subsequent Battle of Zhizhi about “a hundred men” under the Zhizhi’s command fought in a so-called “fish-scale formation”. In the 1940s, Homer H. Dubs, an American professor of Chinese history at the University of Oxford, hypothesized that this might have been the Roman testudo formation and that these men, who were captured by the Chinese, founded the village of Liqian (Li-chien, possibly from “legio”) in Yongchang County. Dubs proposed that the people of Liqian [2] were thus descended from these Roman prisoners. To date, no artifacts that might confirm a Roman presence, such as coins or weaponry, have been discovered in Liqian. Dubs' theories therefore have been rejected by modern historians and geneticists on the grounds of a critical appraisal of the ancient sources and recent DNA testing of Liqian residents. This “rural myth”, which quite regularly resurfaces on social media, has spawned at least three works of popular fiction. Alfred Duggan used the possible fate of the Roman prisoners as the kernel of his novel “Winter Quarters” suggesting they were employed as frontier guards on the eastern border of the Parthian Empire. Ben Kane used the recurring myth as the basis for his novel “The Forgotten Legion”. While in “Empire of the Dragons”, Valerio Massimo Manfredi deviates slightly from the premise placing the starting point of his novel in AD 260, some 200 years after Crassus’ ill-fated campaign. In Manfredi’s version a small group of Roman soldiers escape the Parthians but end up as the personal bodyguard of an exiled prince accompanying him on his epic journey through the forests of India, the Himalayan mountains, the deserts of central Asia and all the way to the heart of China.

Phoulkon Despite this apparent weakness the testudo could still be an effective defence against less well-trained cavalry. Thus we encounter in Plutarch’s version of Mark Antony’s campaign in Parthia in 36 BC a description of the testudo being used as a static defensive formation:

“Then the shield-bearers wheeled about, enclosing the lighter armed troops within their ranks, while they themselves dropped on one knee and held their shields out before them. The second rank held their shields out over the heads of the first, and the next rank likewise. The resulting appearance is very like that of a roof, affords a striking spectacle, and is the most effective of protections against arrows, which glide off from it.” (Plutarch, Parallel Lives: Anthony, 45)

The description is reminiscent of the much later “phoulkon” (Greek: φοῦλκον), in Latin: *fulcum, a close-order infantry formation used by the military of the late Roman and Byzantine Empire. The term “phoulkon” is first attested in the Strategikon of Maurice, a military manual written in the AD 590s. Although written in Greek, the author of the Strategikon “also frequently employed Latin and other terms which have been in common military use” (Rance, 2004, 267). Like other military terminology found in the manual, “phoulkon” is likely a Greek transliteration of a hypothetical Latin word *fulcum, a term not attested in any surviving texts (Rance, 2004, 286). The only other early Byzantine author to use the term was Theophanes the Confessor, who describes Rhazates arranging his troops in three phoulka when facing Heraclius's army at the Battle of Nineveh in AD 627 (Rance, 2004, 310–311). Later Byzantine writings, such as De velitatione bellica and Praecepta Militaria, describe keeping a portion of troops, either cavalry or infantry, in phoulka to serve as guard while the rest of the army dispersed for pillaging or foraging. These later usages appear to have evolved to simply mean a “battle formation”, rather than Maurice's specific description of a shield wall tactic (Rance, 2004, 321–324).

Before close contact with the enemy, and while outside archery range, on the command “ad fulco” (αδ φουλκω) the infantry closed ranks and formed a shield wall from the first two lines (Rance, 2004, 271–272). As they advanced, light infantry from the rear would shoot arrows at the enemy while the heavy infantry could hurl martiobarbuli darts or throw their spears before closing and engaging in hand-to-hand combat with the spatha (Rance, 2004, 274–275). In the presence of enemy cavalry, the first three ranks of the phoulkon would form a shield wall and thrust their spears outwards while fixing the ends in the ground. The third and subsequent rear ranks would hurl projectiles, while the light infantry shot arrows (Rance, 2004, 276–280).

Though only the Strategikon explicitly describes this formation as a phoulkon, such tactics appear to have been established Roman practice. In the Strategikon we read:

“If the enemy [cavalry], coming within a bow shot, attempts to break or dislodge the phalanx...then the infantry close up in the regular manner. And the first, second and third man in each file are to form themselves into a phoulkon, that is, one shield upon another, and having thrust their spears straight forward beyond their shields, fix them firmly in the ground...They also lean their shoulders and put their weight against their shields so that they might easily endure the pressure from those outside. The third man, standing more upright, and the fourth, holding their spears like javelins either stab those coming close or hurl them and draw their swords.”

In his work Ektaxis kata Alanon (“Deployment against the Alans”) Arrian describes an almost identical tactic used centuries earlier against the Alans:

“If [the enemy cavalry] do approach, the first three ranks, closing their shields together and exerting pressure with their shoulders, should receive the attack as steadfastly as possible and locking together very closely, pressing themselves together as firmly as they are able. The fourth rank should throw javelins overhead, while the third rank should strike with their spears or throw them like javelins unstintingly at both horses and riders.”

Other tactical uses Testudines were also employed to cross obstacles if we believe Cassius Dio’s comment: “…whenever they come to a narrow ravine, even horses and vehicles can be driven over it.” One suspects that this might be something performed in extremis when no other options were immediately available. Scrambling up a testudo to scale a palisade or wall appears feasible although curved shields would not offer a particularly good climbing surface. Moreover, in the face of a determined defender it seems a riskier option than using scaling ladders which are attested and would be far more effective.

A fifth use is revealed by Plutarch comment “enclosing the lighter armed troops within their ranks”. The same sentiment is expressed in Cassius Dio’s description of the testudo being used to protect “[t]he baggage animals, the light-armed troops, and the cavalry…placed in the centre of the army.” Two further uses of testudines have been suggested, namely as a defence when reforming troops during lulls in battle and in critical situations during a melee. The former, presumably resembling the aforementioned “phoulkon”, would rely on the troops breaking contact with the enemy to give them the time and space to re-order their ranks. The latter example seems counter-intuitive in the classic all-round defensive testudo which provides little option for hand-to-hand combat but less so if using “phoulka”.

Later usage Testudines were a common feature in the Middle Ages. The following examples serve to prove the tactical sense of using shield formations in battle, especially during sieges where attackers are vulnerable to missiles from the defensive works above (Rance, 2004, 286; 310-311; 321-324):

• A testudo was used by Muhammad's forces during the Siege of Ta'if in AD 630 (Muir, 1861, 145).

• The Carolingian Frankish soldiers of Louis the Pious used one to advance on the walls of Barcelona during the siege of the city in AD 800–801.

• During the siege of Paris in AD 885–886 the Norse besiegers would undoubtedly have used their shields to protect against the defenders’ missiles as the approached the walls and during any breaching assault.

• East Frankish soldiers under King Arnulf of Carinthia used a similar tactic during the siege of Bergamo in AD 894, as did the Lotharingians under Conrad the Red at the siege of Senlis in AD 949, and by Lotharingian defenders at the siege of Verdun in AD 984.

• Just over a century later and the Crusaders of Count Raymond IV of Toulouse were employing the same tactic during the siege of Nicaea in AD 1097 (Bradbury, 1992, 280).

The testudo formation was again employed by Mediæval Arabs, who called it the dabbāba or “crawler”. It was employed by Abu Abdallah al-Shi'i in the AD 906 siege of Tobna (in modern-day Algeria), where he used it to protect sappers as they advanced to the city walls to undermine and collapse a tower, creating a breach for their allies to enter the city (Halm, 1996, 110).

Although in popular culture testudines are associated with the Romans, the formation has a much longer history. This makes perfect sense when one considers that this obvious and simple tactic allows shield armed troops to not only safely manoeuvre, whether on the battlefield or during a siege, but also protect themselves against missiles and projectiles. Indeed, testudines can be observed being employed by modern riot police in various countries across the globe. And why would they not – the testudo just works. Bon appétit!

---

References:

Bradbury, J., (1992), “The Medieval Siege”, Woodbridge: Boydell Press.

Cassius Dio, “Roman History”, Book XL, chapter 22, Volume III of the Loeb Classical Library 1914 edition, available online at LacusCurtius (accessed 25 March 2025).

Cassius Dio, “Roman History”, Book XLIX, chapter 30, Volume V of the Loeb Classical Library 1917 edition, available online at LacusCurtius (accessed 25 March 2025).

Halm, H., (1996), “The Empire of the Mahdi: The Rise of the Fatimids”, (Translated from the German by Michael Bonner), Leiden: E.J. Brill.

Muir, W., (1861), “The Life of Mahomet and History of Islam to the Era of the Hegira: With Introductory Chapters on the Original Sources for the Biography of Mahomet and on the Pre-Islamite History of Arabia”, Smith, Elder & Company.

Plutarch, “Parallel Lives: Antony, chapter 45”, Volume IX of the Loeb Classical Library 1920 edition, available online at LacusCurtius (accessed 23 March 2025).

Rance, P., (2004), “The Fulcum, the Late Roman and Byzantine Testudo: The Germanization of Roman Infantry Tactics?”, Greek, Roman, and Byzantine Studies 44, pp. 265–326.

Tacitus, “Histories”, Book III, chapter 27, Volume II of the Loeb Classical Library 1925 edition, available online at LacusCurtius (accessed 25 March 2025).

Endnotes:

1. The Xiongnu (Chinese: 匈奴) were a tribal confederation of nomadic peoples who, according to ancient Chinese sources, inhabited the eastern Eurasian Steppe from the 3rd century BC to the late 1st century AD. ▲

2. Zhelaizhai (traditional Chinese: 者來寨) is a village on the edge of the Gobi desert in Gansu province, China. The area was renamed after Liqian, an ancient county, and is located in Jiaojiazhuang township, Yongchang County. Some of the modern-day residents of Zhelaizhai, now known as Liqian village, claim to be the descendants of the accounted for Roman prisoners of war after the Battle of Carrhae. ▲