Polybius tells us that the Roman army at Cannae numbered about 80,000 infantry and just over 6,000 cavalry. The figure for the infantry was clearly based on the assumption that there were eight legions of 5,000 foot supported by the same number of allied soldiers. Whether all of the legions and alae were in fact at this theoretical strength on 2 August is questionable, but this figure probably provides a reasonable approximation of the army's size. If all of the legions had their full complement of 300 cavalry then there should have been 2,400 Roman horse at the battle. The remaining 3,600 or so men were allied cavalry.
Normally the allies provided three times as many cavalrymen as the Romans, but this did not occur in the extraordinary circumstances of 216 BC. One of Livy's sources stated that in this year the allies provided twice as many cavalry as the Romans.8 This may simply have been a rough approximation, but it is possible that not all of the legions were able to recruit the full 300 horsemen and that as a result more of the 6,000 were allied cavalry. Despite the great effort mounted for this campaign, the overall proportion of cavalry in the army was lower than in most other Roman field armies. This may well reflect the casualties suffered in earlier engagements, notably the defeat of Centenius in 217. Another problem which doubtless restricted the number of cavalry with the army was the difficulty of finding so many mounts at such short notice.9
Not all of the Roman army was deployed for battle: 10,000 men were left outside the main camp. As yet the Romans did not know whether Hannibal would accept their offer of battle and transfer his army to the right bank of the Aufidius. Polybius tells us that this was Paullus’ decision and that, whilst they would also guard the baggage, this force was intended to pose a threat to Hannibal’s camp. This would either persuade Hannibal to weaken his army by leaving a strong garrison or, if he did not, allow the Romans to storm the Punic encampment. The loss of baggage, equipment, supplies and camp followers would have been a fatal blow, impossible for Hannibal's army to recover from if his army failed to win an outright and overwhelming victory. The attribution of this plan to Paullus once again challenges the tradition that he was less aggressive than his colleague. It is unknown who provided this force of 10,000 men, and whether it consisted of detachments from some or all of the units in the camp or of complete units. Some cavalrymen are mentioned in this camp in the aftermath of the battle which, if they were not fugitives, suggests that the covering force included both horse and foot. It has sometimes been suggested that these men were the triarii from the entire army, since 600 men from each of the eight legions and a similar number from the alae would total 9,600 men. There are a few recorded occasions when the triarii were given the task of protecting the army's baggage, but there is no indication that this was standard practice. Given the aggressive role planned for this force by Paullus, the triarii would not seem the most suitable men to carry this out, for they were not normally used as a strike force. It would also have meant sending some men from the smaller camp which seems rather unlikely. A much more plausible solution would be to see the 10,000 as one legion supported by an ala, but certainty is impossible. Appian, whose account of the battle is generally unreliable and confused, claims that 3,000 men were left as a covering force for the smaller camp. This may have been the case, but this position faced no immediate threat given that the main army was deployed in front of it. It may be that the semi-armed servants and camp followers there were considered sufficient protection, but once again we have no clear information.10
Battle scene showing Greek hoplites from the Nereid Monument in the British Museum. This relief depicting a battle between phalanxes of Greek hoplites demonstrates the problems of showing a massed battle in a two dimensional medium. The formations are shown with men standing one behind the other and there is no attempt to indicate the files of men on either side. This problem would recur in many famous works of art, for instance in the Bayeux Tapestry.
When Varro had united the two sections of the Roman army he formed them into battle formation with the Roman cavalry on the right, the Roman and allied heavy infantry in the centre and the allied horse on the left. Probably, as we have seen, his flanks rested on the river and the high ground around Cannae. There is little or no information on the formations normally employed by Roman cavalry in this period. Polybius, in a criticism of another historian's account of Alexander the Great s victory at the battle of Issus, claims that if they were to be effective then cavalry should never be deployed in more than eight ranks. He also states that wide intervals between squadrons were essential to provide them with the freedom to manoeuvre, so that when formed eight deep 800 cavalrymen occupied a frontage of 1 stadium (roughly a furlong), which works out at about 2m (c. 6-7 feet) per horseman. This is probably rather too generous for the Roman cavalry wings at Cannae. At no stage during the battle are either the Roman or allied cavalry recorded as having mounted a serious attack. They were heavily outnumbered and their role seems to have been merely to protect the flanks of the infantry and prevent them from being outflanked by the enemy horse. In such a defensive role there was no need to maintain such large intervals between squadrons or to form only eight ranks deep. In fact some of the Roman cavalry may even have dismounted and fought on foot, although the tradition is rather confused over this point. Allowing 1.5m (5 feet) per horseman and assuming that the 2,400 Roman cavalrymen on the right flank were formed ten deep, then they will have occupied a frontage of 360m and a depth of perhaps 40m. Employing the same calculation, the 3,600 allied cavalry on the left would have needed 540m by 40m.11
As far as we can tell, the Roman legions and allied alae deployed in the normal triplex acies. However, Polybius specifically tells us that Varro ordered two major changes from the normal drill, reducing the gaps between the maniples in each line and making each maniple very deep, so that each was ‘many times deeper than it was wide'. The standard size of a maniple of hastati or principes in a normal legion was 120 men, but we do not know whether there was a standard formation for this unit. Polybius’ statement here, as well as practical utility, suggests that a maniple was not normally formed with greater depth than frontage. If the Romans preferred to have equal numbers in each rank then a formation of twenty men wide by six deep or fifteen men by eight deep would be prime candidates. If each legionary was allocated a frontage of 1m (c. 3 feet) and a depth of 2m (6-7 feet) then the maniple would cover 20m by 12m or 15m by 16m respectively. The scant evidence for the formations of the later professional army suggest systems of drill based upon multiples of three or four, the latter being standard for Hellenistic armies. However, it is possible that the third century BC Roman army had no standard system of drill and that the depths of maniples was determined for each battle by a legion's or an army's commanders, as had been the case with most Greek hoplite armies.12
The legions at Cannae were unusually large with 5,000 foot apiece. Polybius tells us that when the size of a legion was increased, the number of triarii always remained the same at 600. The remaining 4,400 men were supposed to be divided equally between the hastati, principes and velites, giving each approximately 1,466 men. This would give an average strength for a maniple in the first two lines as about 146 men. If there were in fact seven legions and seven alae making up the line, allowing for one of each left in the larger camp, this would give 20,524 men in each of the first two lines and 8,400 in the third, a total of 49,448, supported by 20,524 velites. Such a strong force of light infantry ought to have given the Romans a distinct advantage in the skirmishing at the beginning of the battle, and, although there may be other reasons why this was not the case, it is possible that there were fewer velites. We know so little about the internal organization of the alae that it is impossible to say whether in fact these included roughly the same proportion of light infantry as the legions. It is also possible that, as in so many other respects in the 216 campaign, the normal procedures had been modified and the legions were themselves composed differently. In either case perhaps we should reduce the number of skirmishers by several thousand and add these to the heavy infantry. Most commentators on the battle estimate the number of close order infantrymen at around the 50,000 to 55,000 mark, but precision is impossible.13
We do not know precisely how deep the Roman centre was at Cannae, and various suggestions have been made, usually ranging from about fifty to seventy ranks. In some cases the lower figure has been based upon the assumption that the triarii were left in the main camp. Polybius' statement that the maniples were ‘many times' deeper than wide is fairly vague, but it is difficult to see it being applied to a formation much wider than five or six files across and less deep than twenty-nine or twenty-four ranks. If we assume that each maniple had a frontage of five men and a depth of twenty-nine then it would have occupied an area of 5m by 58m. The gaps between maniples in the same line were normally equivalent to the frontage of a single maniple, but at Cannae this was significantly reduced. Assuming an interval equal to half the width of a maniple, then the ten maniples of hastati in one legion at Cannae would have occupied about 75m by 58m, and the entire first line of the army about 1,050m by 58m. To cover the same frontage the principes would have been in an identical formation, but the less numerous triarii formed around ten to twelve deep, giving a total depth to the Roman centre of perhaps seventy-four ranks. If those scholars who suggest a somewhat shallower formation of around fifty ranks are closer to the mark, then the frontage of the infantry centre would have to be expanded to around 1.5km.
This scene from the altar of Domitius Ahenobarbus shows animals waiting to be sacrificed. In the centre are two legionary infantrymen, who could easily represent troops from any of the three lines of heavy infantry in the legions at Cannae. To the right a cavalryman stands in front of his horse.
The Roman centre at Cannae concentrated an exceptionally large number of men on a very narrow frontage and it is important to understand why such an unorthodox deployment was adopted. In any formation only the legionaries in the front rank could effectively employ their weapons. Soldiers armed with long spears might be able to reach the enemy from the second rank, but only the triarii at Cannae were equipped in this way. Men in the ranks behind the first could throw missiles over the men in front, although restricted visibility must have made it difficult to aim and it was a question of hoping to hit somewhere in the enemy mass. The pilum, the heavy throwing spear carried by legionaries and perhaps some allied soldiers, had a maximum range of just under 30m and was most effective at about half that distance. Therefore any soldiers in the ranks behind the eighth in a Roman maniple would have had difficulty in throwing their pila without running the risk of hitting their own front ranks. A deep formation did not offer the most effective use of a unit's weaponry which could best be served by a much shallower formation of perhaps two or three ranks, with the front rank to do the actual fighting and the others to replace casualties. Yet such shallow formations were exceedingly rare and most military theorists felt four ranks to be the minimum and recommended six or eight. Shallow formations tended inevitably to be wide and the wider a unit's frontage, the harder it was to move across the battlefield at any speed and remain in formation. A broader formation encountered more obstacles, since no battlefield was ever perfectly flat, and required a good standard of drill and the close supervision of its officers to prevent a unit from falling into disorder. As a result, over any distance, a narrower, deeper column would move more quickly whilst retaining its order than a wider, shallow line.
There were other reasons why troops tended to fight in deeper formations, which went beyond the purely practical. A column many ranks deep was an intimidating sight as it approached the enemy, even if many of the men within it would not actually be able to fight. As importantly, the close proximity of their comrades all around them encouraged the men forming the column.
There appears to be a strong herd instinct within human beings, so that even today there is a marked tendency for men under fire to bunch together for mutual comfort, despite the fact that this tends to make them more of a target. Only rigorous training has proved able to control this instinctive reaction. The physical presence of their comrades encouraged men, but also made it difficult for them to flee. The front rank could not run until all the other ranks behind them had given way. The men in the centre and rear of the formation were removed from actual physical danger in direct relation to the depth of the formation. Deeper formations did not fight any better than shallow formations, but they did possess longer endurance in combat, simply because it was that much more difficult for the actual fighting men in front to escape. Greek military theorists recommended placing the best and bravest soldiers in the front and rear ranks, the former to do the actual fighting and the latter to prevent the rest of the unit from escaping. We cannot be sure whether or not they had this function in the third century BC, but in the latter Roman army the centurions' second in command, the optiones, were stationed at the rear of a century to prevent the men from running away, if necessary physically forcing them back into place with their symbol of office, the hostile staff. An especially deep formation was one way of keeping questionable troops in the battle for a longer period, increasing the chance that the enemy’s morale would crack first. Throughout the ancient world, and indeed for much later military history when troops continued to fight in close formation, there was a direct link between the quality of troops and the depth of their formation. Highly trained and well motivated soldiers were able to fight in much shallower formations than was ever possible for less experienced and poorly drilled units.14
The inexperience and lack of training of much of the Roman army at Cannae in part explains the decision to form them in such depth. It would have been exceptionally difficult to keep together a more conventional and shallower triplex acies formation, with each of the three lines stretching for several kilometres. The reduction in the frontage of each maniple and even more importantly of the intervals between them removed much of the manipular formation's flexibility, but it did make it possible to move so many men in a more co-ordinated manner. The great depth also gave the Roman infantry phenomenal staying power, making it extremely difficult, if not impossible, for an enemy to defeat them in a straightforward frontal attack. Finally we must remember that the Romans' formation may have been limited by the ground chosen for the battle. According to the above calculations, the entire Roman army occupied a frontage of about 2km or 2,100 yards (360m for the Roman cavalry, 1,050m for the infantry centre, and 540m for the allied cavalry = 1,950m). Precision cannot be claimed for a figure based upon so many conjectures and assumptions, but even if this is something of an underestimate it would still suggest that the Roman army was more than capable of being fitted into the area between Cannae and the presumed more northern course of the Aufidius. Delbriick and Lehmann, the chief advocates of a battlefield on the left bank, similarly calculated the frontage of the Roman army as under 2km in which case it could have fitted into a plain between two physical obstacles which protected its flanks.15
The Roman plan for the battle was simple and unsubtle, but not unreasonable or by any means inevitably doomed to failure. At Trebia a large section of the Roman legions had cut their way straight through the Punic centre, defeating not just Gallic warriors, but also the Africans, Hannibal’s best infantry. At Trasimene, in spite of the massively unfavourable position and lack of organization, the Roman heavy infantry had held off the enemy attacks for hours and inflicted significant losses upon them. Throughout the same period, the Roman and allied cavalry had performed consistently badly, winning only a few minor engagements. In this battle the Romans were once again outnumbered in cavalry, but had a massive advantage in infantry. It was therefore logical to rely most upon their foot in the coming battle. The problem they faced was how to bring the weight of their infantry to bear without exposing its flanks to Hannibal's superior and mobile cavalry which had so easily swept around the Roman flanks at Trebia and robbed them of any real advantage gained by breaking the enemy centre. The answer, and it seems the reason why Varro chose to fight on the opposite bank of the River Aufidius to where Hannibal had offered battle, was to deploy not in a wide open plain, but in a narrower, more confined space. In this way the left wing was protected by the hills around Cannae and the right by the river. Although heavily outnumbered, the Roman and allied horse could not be outflanked and the enemy cavalry would be forced to attack and defeat them in a frontal charge. The Roman wings were not required to beat the enemy, but simply to stay in position for as long as possible. They were there to give sufficient time for the massed infantry in the centre to deliver an overwhelming hammer blow against Hannibal's foot. If the Punic centre could be overwhelmed, then it would matter little if the Roman wings at last gave way, for on their own the Carthaginian cavalry would not be able to do much more than harass the legions. The selection of the ground at Cannae was intended to allow the heavy infantry to smash their Punic counterparts. We do not know whether Varro alone or perhaps with the assistance of Paullus, Geminus or some of the other experienced men with the army conceived this plan. It was not complicated or especially imaginative, and in fact the very close formation of the Roman foot sacrificed the usual tactical flexibility of the legions. Elsewhere Polybius commented that the Romans as a race tended to rely instinctively on ‘brute force' (bia) when making war and that sometimes this had led to terrible disasters. Their plan at Cannae would seem to be a prime example of this trait.16 The cavalry wings were the vulnerable spots, for they needed to remain in place long enough for the infantry to win. This was always going to be difficult, for infantry combats seem usually to have lasted for hours, whereas cavalry encounters were faster and more fluid. It was no coincidence that the two consuls took direct command of the wings, Varro leading the allies on the left and Paullus the Romans on the right. That Paullus was stationed with the citizen cavalry rather than the allies has been one of the chief arguments put forward for claiming that it was he who held overall command on 2 August, since it is assumed that this was a more prestigious post than controlling the non-citizen allies. This is in fact extremely tenuous, for there does not appear to have been any set place from which a Roman consul was supposed to lead the army. Most certainly there was no convention for where each consul should be when both were present with the army, for this was such a rare event. Roman generals tended to station themselves wherever they felt that they could do most to influence the battle and thus usually where they anticipated its crisis to occur, hence the consuls' presence with the cavalry at Cannae. The centre was placed under the command of Servilius Geminus. There were also many tribunes and prefects, so that the Roman and allied foot were led by a very large number of senior officers concentrated along a small frontage. We do not know how the legions and alae were arranged. Livy says that the Roman legions were on the right and the allied foot on the left, but it is hard to know what to make of this. Conventionally the legions held the centre and the alae were split on either side, but there was no precedent for an eight legion army to know whether this would be followed in these circumstances. One possibility is that the different armies formed up side by side, so that the legions and alae used to working with each other remained together. One attractive idea is that the centre of the line was formed by the proconsuls' forces, so that the best and most experienced troops formed the heart of the Roman attack, but once again this is purely conjectural. Whatever the precise details, the unprecedentedly large Roman host can only have presented an intimidating sight to the watching Carthaginians.17
A closer view of the soldiers on the altar of Domitius Ahenobarbus. All three wear mail shirts like that depicted on the Gallic warrior from Vacheres. Those of the two infantrymen have clearly depicted doubling on their shoulders. Both men carry heavy semi-cylindrical shields with barleycorn shaped bosses either of wood or bronze. Their helmets are probably of Attic or Etrusco-Corinthian pattern.
A relief depicting an Iberian cavalryman from Osuna in Spain. Much detail has been lost but he clearly carries a short sword. A good number of Hasdrubal’s heavy cavalry on the left wing were Spanish.
The Romans could not be sure that Hannibal would accept their offer of battle in this confined position, hence the strong force left outside the larger camp. It is even possible that Varro did not expect the Carthaginian to fight and saw this largely as a morale boosting operation for his own soldiers, rebuilding their confidence after the humiliation of the day before. This is certainly possible for such gestures were common before the battles of this period, but it is far more probable that the Romans did want to fight the battle on this ground of their own choosing. Whatever their intentions, the sources imply that Hannibal ordered his army to move out almost as soon as he saw Varro's columns leaving the main camp. The Carthaginian sent out his light troops to form a protective screen for the main body. This was standard practice for most armies and it is more than probable that the complicated process of forming up the Roman army was carried out behind a line of velites and perhaps some of the cavalry. Polybius noted that Hannibal's army crossed the Aufidius at two points, which makes it very likely that it was divided into two columns for deployment. The Carthaginians then formed into battle order, probably within a kilometre of the Roman line.18
Hannibal is said to have had 10,000 cavalry and 40,000 infantry at Cannae, but we have no precise figures for the various contingents making up this total. The cavalry were a mixture of Spanish and Gallic horse, both of whom fought in close order, and the Numidian light cavalry. The Gauls had all been recruited from the Cisalpine tribes after Hannibal arrived in Italy and he is said to have had 6,000 Spanish and Numidian horse after crossing the Alps. At most the Numidians may have accounted for two thirds of this total, but they must have suffered some casualties in the 218 and 217 campaigns and their strength was probably somewhere between 3,000 and 4,000. At Cannae the Numidians were placed on the Punic right flank opposite Varro and the Allied cavalry and it seems probable that they roughly equalled these in numbers. All of the Gallic and Spanish cavalry, some 6,000-7,000 men, were massed on the left flank, giving them a numerical superiority of two or three to one over their Roman counterparts. Probably the Punic horse occupied roughly the same frontage as the Roman and allied cavalry, but it is unlikely that they were quite so densely packed. The Numidians normally fought in small, widely spread bodies which advanced and retreated in each other's support, always avoiding close contact but harassing the enemy with missiles. The Gauls and Spaniards were almost certainly divided into several lines of squadrons, for if cavalry became too crowded then they tended to merge into one mass which was difficult for its leaders to control and inclined to panic and stampede. The behaviour of these horsemen during the battle makes it clear that they were kept closely in hand by their officers.
The initial deployment of the two armies at Cannae, shown to scale under the assumption that the River Aufidius ran further north in 216. Each army occupied a frontage of some two kilometres. cramming some 130,000 men into a narrow valley, but giving the Romans secure flanks. Hannibal's formation made best use of the differing qualities of his soldiers.
Hannibal's infantry were formed with the African foot on the flanks and the Gauls and Spanish in the centre. It is difficult to know how many men should be deducted from the total of 40,000 to account for the infantry skirmishers. He had had 8,000 of these men at Trebia and although his army had been significantly augmented by Gallic tribesmen since then it is questionable how many skirmishers these provided. The warrior culture of the Gallic tribes placed most emphasis on close fighting and light infantry seem to have played little part in inter-tribal warfare. If there were still only 8,000 light troops at Cannae, then the close order foot mustered something like 32,000 men, once again divided into Gauls, Africans and Spanish. Hannibal had 20,000 foot when he arrived in Italy, consisting of the Libyans, Spanish and light infantry, and none of these contingents had as yet received any reinforcements. Perhaps there were around 6,000 light infantry, some of them Spanish, 4,000 Spanish close order troops and 10,000 Libyans. All of these had suffered some casualties by August 216. If there were 8,000-9,000 Libyans and around 3.000-4,000 Spanish at Cannae, then that would suggest something like 19.000-21,000 Gallic warriors in the main line.
The Carthaginian centre consisted of the Gallic and Spanish foot, perhaps 24,000 men in all. These were intermingled, companies or units of each being deployed alternately. Polybius uses the Greek word speirai (spe...rai) for these units, a term which he also sometimes employs for the Roman maniple and which would later become the term used for the 480 man cohort adopted by the professional Roman army. It is doubtful that these bands were of uniform size, and anyway either there must have been more Gallic than Spanish units or each unit was significantly larger, but it is most likely that this refers to a group of a few hundred men and probably less than a thousand. The African foot were divided into two roughly equal bodies and stationed on the flanks near the cavalry. They were formed either in one deep column or in several lines one behind the other. Although this is not explicitly stated by any of our sources, it is highly probable that the Libyans were in fact behind the main line of Gauls and Spanish and concealed from the Romans' view. The likelihood is that they formed one of the two columns used by the Punic army in its deployment. The Africans, issued by Hannibal with captured Roman equipment, were approximately equivalent in size to two legions, one behind each of the Carthaginian army's flanks.19
Once all the units of the army were in place, Hannibal made one major alteration to its formation, advancing the units in the centre of the main line so that this bulged towards the enemy. The most probable interpretation of the descriptions in our sources is that the companies in the very centre of the line advanced to form a line further forward and that the companies on either side were echeloned back. It must have been obvious that the Romans were relying on the densely packed mass of infantry in their centre to win the battle. Hannibal’s foot were greatly outnumbered and in the past had had difficulty standing up to the legions even on equal terms. At Trebia Hannibal's foot had formed a single line so that there were no reserves to plug the gap when the Romans broke through his centre. At Cannae he could have chosen to make his centre as strong as possible by concentrating all the foot into one line. Yet he could not rival the depth and therefore the endurance of the Roman infantry and in the end such a line was likely to give way. Another option would have been to copy Roman practice and divide the foot into two or more lines stationed one behind the other, so that as the troops in the fighting line became weary they could be reinforced by fresh reserves, something Hannibal would choose to do at Zama in 202 BC. However, once again, this could at best delay the inevitable Roman breakthrough. Instead of delaying the clash with this overwhelming force, Hannibal kept his centre relatively shallow and then moved it forward so that the Roman charge would reach it more quickly. His objective was to concentrate the Roman effort at the very centre of his line, knowing that it must inevitably break and that the Romans would pour through the gap they had created. Then, it was hoped, they might make themselves vulnerable to flank attacks from the Libyan foot, his best, most disciplined men. Whilst this was going on the Numidians on the left were supposed to keep the Latin cavalry busy, and at the same time the concentrated force of the Gallic and Spanish horse mounted a huge, direct attack at the Roman right. These were to smash the Roman cavalry and then, remaining in good order, threaten the rear of the Roman army.20
In essence Hannibal hoped to use the Romans' own strength against them, drawing them in to be surrounded and destroyed. It was a complex plan, contrasting sharply with the brutally simple tactics of the enemy. It also made very heavy demands on both his soldiers and their officers. Hannibal himself, supported by his youngest brother Mago, took up position in the centre with the Spaniards and Gauls for it was vital that these warriors held out as long as possible. The Numidians were led by Hanno according to Polybius and Maharbal in Livy's account, whilst the vital task of leading the squadrons on the left was entrusted to Hasdrubal. The events of the battle were to demonstrate the great superiority which the Carthaginian army derived from the command structure and mutual trust between leaders and forged by years of campaigning together. In spite of this, the plan was fraught with risks and by no means as certain of success as is sometimes implied in modem accounts.
Hannibal's tactics were tailored to the specific conditions of the battle, with the Romans, and especially their infantry, deployed in great depth on an exceptionally narrow front. Over the years some highly fanciful attempts have been made to see this as the fruition of long held plans, perhaps even based on the naval battle of Ecnomus or first conceived by Hamilcar Barca, passed on to his sons and experimented with at Trebia and Ibera in 215. There is no reason to believe that either Hamilcar or his capable sons attempted to conform rigidly to previously conceived plans. It is worth considering when Hannibal decided to form his army in this way. The decision to advance the centre of the Spanish and Celts may have been made on the spot, late in the stage of the army's deployment, but the concentration of all of his heavy cavalry on the left and the positioning of the Libyan foot on the flanks must already have been decided upon before the army marched in its deployment columns out of camp. Organizing the army and issuing orders so that each contingent knew where it was supposed to be took time. It seems unlikely that Hannibal could have devised these complex tactics and the means of implementing them after he had seen the Romans begin to march out of camp and cross the river on the morning of 2 August. There was no reason for him to have formed his army in this way and on such a narrow frontage when he offered battle in the plain north of the river. Had he known that the Romans were planning to form on the narrow plain between Cannae and the river then the lack of space would have given a fairly clear idea of their likely deployment. This raises the intriguing possibility that the Carthaginians had seen Roman officers looking at the ground on the previous day - a highly likely activity if they were considering fighting on it. If this is right then it would further support the idea that Paullus, who was in command on 1 August, was far less reluctant to fight a battle than our sources suggest. In some way Hannibal does seem to have known or guessed how and where the Romans would fight and devised his plan accordingly, or perhaps his soldiers and officers were so superior to the enemy that he was able to react to the Roman plan and still form up within the time taken for the great enemy host to deploy.21