battle
Battle of Cannae
216 BCE ยท Cannae
The most studied battle of antiquity and Hannibal's tactical masterpiece. Drawn south into Apulia by Hannibal's seizure of the Roman supply depot at Cannae, the Roman consuls Aemilius Paullus and Terentius Varro engaged with an army roughly twice the size of the Carthaginian force. Hannibal placed his weakest infantry, Iberians and Gauls, in a forward crescent in the center, with his veteran African heavy infantry held back on the wings. As the Roman center pressed forward, the crescent flexed back without breaking; the African veterans wheeled inward against the Roman flanks while the Carthaginian cavalry, having broken the Roman horse on both wings, returned to attack the Roman rear. The Roman army was encircled and systematically destroyed. Casualty figures are contested, Polybius reports 70,000 Roman dead, Livy somewhat fewer, but on any plausible reading the proportion of Roman forces lost in a single day at Cannae is among the highest in the history of warfare. Three of the four senior Roman commanders fell. Capua and most of the Italian south defected to Hannibal in the months that followed. Yet Rome did not negotiate. The battle ended Roman field engagements on Hannibal's terms but did not end Roman strategic capacity, a survival the ancient historians treat as the pivot of the entire war.
Context
Related events
- Battle of Zama
matched pair
Read the full narrative โ
The Battle of Cannae
Cannae was Hannibal's tactical masterpiece and a strategic limit. The battle's military design (the deliberate double envelopment, the trading of a thinner Carthaginian line for an annihilating victory) is rightly remembered. But the battle's strategic legacy lies less in what Hannibal achieved than in what Rome refused: the negotiation that any other ancient state would have offered after losing 50,000 men in a single afternoon. That refusal, and the structural manpower depth that made it possible, set the limit on what Cannae could accomplish.
Participants
- Hannibal Barca
commander ยท Carthaginian
- Hasdrubal (Hannibal's cavalry commander)
cavalry commander ยท Carthaginian
- Maharbal
cavalry commander ยท Carthaginian
- Mago Barca
subordinate commander ยท Carthaginian
- Lucius Aemilius Paullus
commander ยท Roman
- Gaius Terentius Varro
commander ยท Roman
What we know
Each claim below is tagged with its epistemic confidence and linked to the ancient and modern sources that support, qualify, or contradict it.
Major Italian allies of Rome, including Capua, the second-largest city of Italy, and most of the Greek south, defected to Hannibal in the weeks and months after Cannae.
Scholarly consensus: Universally accepted. The pattern of Italian defections in the year after Cannae is well-documented and is generally read as evidence that Hannibal's broader strategy, to break Rome's confederation rather than to take Rome itself, was at least partly working.
Sources
- โ Polybius of Megalopolis, Histories 3.118 โ Polybius lists the principal defections, Capua, Atella, Calatia, most of Bruttium and Lucania, the Hirpini, and parts of Apulia, and treats this as the immediate consequence of Cannae.
- โ Titus Livius (Livy), Ab Urbe Condita 22.61, 23.1-7 โ Livy gives a fuller and more dramatic account of Capua's defection in particular, including the negotiations and the city's conditions.
- โ Eve MacDonald, Hannibal: A Hellenistic Life โ MacDonald reads the defection pattern as evidence of how thinly Rome's hegemony in central and southern Italy had been built, and as evidence that Hannibal's strategy of detaching allies was a coherent alternative to direct attack on Rome.
Roman losses at Cannae were among the highest of any single day in the history of warfare, perhaps 50,000โ70,000 dead and 10,000 captured. Carthaginian losses were under 10,000.
Scholarly consensus: The catastrophic asymmetry is firmly attested. The specific totals vary across the principal sources and are treated by modern scholars as approximate. The order of magnitude, tens of thousands of Roman dead, is not seriously contested.
The debate โ
- **Polybius (3.117)** reports approximately 70,000 Roman dead, 10,000
captured, and roughly 5,700 Carthaginian losses.
- **Livy (22.49)** gives approximately 47,500 Roman infantry and
2,700 cavalry killed, with similar figures for captives, and
approximately 8,000 Carthaginian dead.
Modern scholarship typically treats Polybius's figure as the upper
bound and Livy's as the lower; the difference partly reflects whether
troops who escaped to nearby Roman strongholds and were captured later
are counted as battle casualties or separately. The proportion of
Roman forces lost is the more telling figure: on Polybius's totals
~75% of the Roman field army was killed or captured in a single day.
The asymmetry between Roman and Carthaginian losses, roughly 10:1,
is striking and may itself be partly inflated. Even at Livy's figure
of 8,000 Carthaginian dead, a tactical victory of this magnitude with
losses under 20% is extraordinary and unparalleled in the period.
The senior Roman casualties are well-attested and not in dispute:
the consul Aemilius Paullus, the proconsul Servilius Geminus (consul
in 217), the former magister equitum Minucius Rufus, and a large
proportion of the Roman senatorial class who served as military
tribunes all fell at Cannae.
Sources
- โ Polybius of Megalopolis, Histories 3.117 โ Polybius reports approximately 70,000 Roman dead, 10,000 captured, and ~5,700 Carthaginian losses; treats the Carthaginian losses as heaviest among the Iberian and Gallic infantry of the forward crescent.
- ~ Titus Livius (Livy), Ab Urbe Condita 22.49 โ Livy gives lower Roman totals (~47,500 infantry plus ~2,700 cavalry killed) and a higher Carthaginian total (~8,000), partially reflecting different counting conventions.
- ~ Plutarch of Chaeronea, Parallel Lives (selected, Cato the Elder, Fabius Maximus, Marcellus, Aemilius Paullus) Aem. Paullus 14, Fabius 16 โ Plutarch preserves figures broadly consistent with Polybius and adds detail on the senatorial casualties.
- ~ Adrian Goldsworthy, The Fall of Carthage, The Punic Wars 265โ146 BC โ Goldsworthy surveys the figures and treats the proportion of forces lost, rather than any specific total, as the measure that matters militarily.
Hannibal's deliberate tactical design at Cannae was a double envelopment: a forward crescent of weaker Iberian and Gallic infantry that gave ground under Roman pressure, while veteran African heavy infantry on the wings held position and then wheeled inward against the Roman flanks, with the cavalry, victorious on both wings, completing the encirclement from the rear.
Scholarly consensus: Universally accepted; the Cannae double envelopment is the canonical ancient and modern account of the battle's tactical structure. Disagreements concern detail (which Carthaginian wing won the cavalry fight first; how deliberate the crescent's flex was) rather than the basic disposition.
Sources
- โ Polybius of Megalopolis, Histories 3.113-117 โ Polybius gives the fullest tactical account: the crescent disposition with the weakest infantry forward, the African heavy infantry held back on the wings, the cavalry envelopment from both flanks, and the systematic destruction of the encircled Roman army.
- โ Titus Livius (Livy), Ab Urbe Condita 22.46-47 โ Livy reproduces substantively the same account with characteristic rhetorical heightening of the Roman experience inside the encirclement.
- โ Adrian Goldsworthy, The Fall of Carthage, The Punic Wars 265โ146 BC โ Goldsworthy provides detailed modern military analysis of the tactical execution, treating the maneuver as a deliberate design that depended on the cavalry's ability to defeat the Roman horse and return to the infantry battle in time.
Maharbal urged Hannibal to march directly on Rome immediately after Cannae and, when Hannibal refused, replied "you know how to win a victory, Hannibal, but you do not know how to use one."
Scholarly consensus: Modern scholarship overwhelmingly treats the episode as later literary invention, most likely Livian or developed in the post-Polybian Latin tradition, rather than historical record. The substantive question the anecdote raises (whether Hannibal could or should have marched on Rome) is real and serious; the specific exchange is almost certainly not.
The debate โ
*Fabius* and in later Roman literary tradition. The famous formulation,
*vincere scis, Hannibal; victoria uti nescis*, is a tightly
constructed Latin epigram, of the kind characteristic of Livian speech
composition.
Reasons modern scholarship is skeptical:
1. **Polybius does not preserve it.** Polybius wrote earlier (~150 BCE
vs. Livy's ~25 BCE), interviewed participants and their families,
and gives the post-Cannae council without this exchange. An anecdote
this dramatic would not have escaped Polybius's notice if it had
been current in the historical tradition he worked from.
2. **It serves a Roman literary purpose.** The exchange dramatizes the
thesis, central to Livy's narrative, that Rome was saved by
Hannibal's failure to seize the moment after Cannae. The anecdote
externalizes onto a Carthaginian voice the criticism Roman
historiography needed to validate Roman survival.
3. **The military assessment behind the anecdote is questionable.**
Hannibal lacked siege equipment for an assault on Rome, his army
had just fought a major battle, and the march from Cannae to Rome
would have taken weeks. The "if only Hannibal had marched on Rome"
scenario is a counterfactual the ancient sources favor more than
modern military analysis does.
The site treats the anecdote as Roman literary tradition rather than
historical reportage. The underlying strategic question, what
Hannibal could realistically have achieved by marching on Rome, is
treated separately as an open question.
Sources
- โ Titus Livius (Livy), Ab Urbe Condita 22.51 โ Livy preserves the exchange in its classic Latin form, with Maharbal's epigram as the dramatic close of the Cannae chapter.
- โ Plutarch of Chaeronea, Parallel Lives (selected, Cato the Elder, Fabius Maximus, Marcellus, Aemilius Paullus) Fabius 17 โ Plutarch preserves a version of the same exchange, working from the Latin tradition.
- โ Polybius of Megalopolis, Histories 3.118 โ Polybius's account of Hannibal's post-Cannae deliberations does not include the exchange with Maharbal. The silence is generally treated as significant.
- ~ Adrian Goldsworthy, The Fall of Carthage, The Punic Wars 265โ146 BC โ Goldsworthy treats the anecdote as later tradition but uses it as a frame for the substantive military question of whether marching on Rome was a real option.
Cannae did not produce Roman political collapse. The senate refused to negotiate with Hannibal, refused to ransom captured prisoners, and within months had raised replacement legions and resumed the war effort.
Scholarly consensus: Firmly attested in both Polybius and Livy. Modern scholarship treats this as the war's true pivot, the moment that demonstrated Rome's strategic depth and structural resilience, and the point after which Hannibal's path to victory effectively closed.
Sources
- โ Polybius of Megalopolis, Histories 3.118, 6.58 โ Polybius emphasizes Rome's refusal to despair as the defining Roman virtue. Book 6, the analytic excursus on the Roman constitution, uses Cannae as the test case for the constitutional structure that enabled the survival.
- โ Titus Livius (Livy), Ab Urbe Condita 22.55-61 โ Livy gives an extended account of the senate's response: refusal to negotiate, refusal of ransom, replacement levy of younger citizens and even slaves, the prosecution of survivors who had fled, and Varro's reception on his return.
- โ Plutarch of Chaeronea, Parallel Lives (selected, Cato the Elder, Fabius Maximus, Marcellus, Aemilius Paullus) Fabius 17-19 โ Plutarch's Fabius traces the return to Fabian strategy after Cannae and the gradual restoration of Roman strategic position through avoidance of decisive engagement.
- โ Dexter Hoyos, Hannibal's Dynasty: Power and Politics in the Western Mediterranean, 247โ183 BC โ Hoyos identifies the post-Cannae Roman response as the structural reason Hannibal's strategy could not succeed: a victory of this scale should have ended the war and, in any other state of the ancient Mediterranean, would have.
Cannae transformed Roman strategic and tactical thinking in the long run, driving the return to Fabian avoidance of decisive engagement, the systematic destruction of the Roman senatorial-tribunate command pool, and the eventual Scipionic reforms that produced the more flexible manipular practice on display at Zama.
Scholarly consensus: Strong; standard reading in modern military histories of the Roman Republic, though the specific causal chains from Cannae to later Roman tactical developments are inferred from outcomes rather than directly documented in the sources.
Sources
- โ Polybius of Megalopolis, Histories 6.58 โ Polybius treats Cannae as the test case for the Roman constitution and notes the transformation of senior Roman command following the loss of so much of the senatorial class in the battle.
- โ Adrian Goldsworthy, The Fall of Carthage, The Punic Wars 265โ146 BC โ Goldsworthy traces the tactical lineage from post-Cannae Roman reforms through the Iberian campaigns to Scipio's reformed maniples at Zama, treating Cannae as the catalyst for changes that took two decades to mature.
- โ Dexter Hoyos, The Carthaginians โ Hoyos identifies the post-Cannae senatorial losses as a structural transformation of Roman politics in their own right, opening command opportunities to a younger generation that included Scipio Africanus.
At Cannae the Roman force outnumbered Hannibal's army roughly two-to-one in infantry but was inferior in cavalry, approximately 80,000 Roman infantry and 6,000 cavalry against 40,000 Carthaginian infantry and 10,000 cavalry.
Scholarly consensus: The asymmetric advantages, Roman infantry preponderance, Carthaginian cavalry preponderance, are firmly attested. The specific totals are Polybius's figures and are debated by modern scholarship; some recent reconstructions place the Roman force somewhat lower, but the order of magnitude and the cavalry imbalance are not seriously contested.
The debate โ
allies, roughly 80,000 infantry and 6,000 cavalry, and the
Carthaginian force at approximately 40,000 infantry and 10,000 cavalry.
Livy (22.36) is broadly consistent.
Modern scholarship treats Polybius's totals as plausible but inflated:
Brunt's *Italian Manpower* and Goldsworthy's *The Fall of Carthage*
both note that mobilizing eight full legions for a single field army
was unprecedented and that Polybius may be retroactively reasoning
from the constitutional limit of consular forces. Lower estimates
(~50,000โ70,000 Roman effectives) have been proposed but no consensus
alternative has emerged.
What is not disputed is the *structural* situation: a deep, thickened
Roman line, Varro is reported to have ordered the maniples drawn up
closer than usual to maximize shock, facing a thinner Carthaginian
line with cavalry strength concentrated on the wings.
Sources
- โ Polybius of Megalopolis, Histories 3.107, 3.113 โ Polybius gives the Roman army as eight legions plus allies, with ~80,000 infantry and ~6,000 cavalry; Hannibal's force at ~40,000 infantry and ~10,000 cavalry.
- โ Titus Livius (Livy), Ab Urbe Condita 22.36 โ Livy is broadly consistent with Polybius's figures, with minor variations in the breakdown of Roman versus allied troops.
- ~ Adrian Goldsworthy, The Fall of Carthage, The Punic Wars 265โ146 BC โ Goldsworthy treats Polybius's totals as the upper bound and surveys lower modern estimates. The cavalry imbalance and the Roman infantry preponderance are not affected by the uncertainty in totals.
Editorial
Open questions
-
What was Hannibal's actual strategic objective in Italy, and was Cannae meant to achieve it?
Hannibal left no surviving statement of his war aims; the sources that attempt to reconstruct them (especially Polybius and Livy) write decades or centuries later, with access to events but not always to motivation. The result is that modern scholarship can describe what Hannibal did but must infer what he was trying to do. The inference is genuinely contested, because the same actions are consistent with several distinct strategic theories.