As in the initial advance, neither side gained a decisive advantage in the missile battle and eventually the two lines met. Perhaps this was gradual, the Romans edging forward whilst throwing their pila, or more sudden as their officers were able to lead them in a charge sword in hand. The hastati had lost some of their order during the advance and subsequent javelin combat, but it still may have been possible, in spite of the noise and confusion, for centurions and tribunes to urge the maniples closest to the enemy to charge together. The Celts and Spaniards may have came forward to meet them, for it seems to have been unusual for infantry to remain entirely on the defensive against other foot, unless they were formed in an especially dense formation which made movement difficult. Shouting was redoubled as each side tried to appear as confident and frightening as possible when they at last closed.
Hand-to-hand combat is especially difficult for us to visualize accurately and all too often conjures up images which have more to do with Hollywood than with reality. In cinematic epics the two armies rapidly intermingle, every soldier fighting aggressively in combats which always end in the death of one of the participants. The whole swirling scene rarely lasts for more than a few frenzied minutes in which huge casualties are inflicted on both sides. In recent centuries hand-to-hand combat has been very rare, even in battles involving armies armed primarily with edged weapons, and massed fighting between formed units of men has been virtually unknown, since almost invariably one side or the other fled before actual contact. There is a good deal of evidence to suggest that in ancient battles it was also not uncommon for one side to rout before a blow had actually been struck or after very little fighting, but it is also certain that sometimes the opposing sides fought longer combats. From the descriptions in our sources, skeletal remains of the dead from battles fought with edged weapons, and comparison with modern studies of the behaviour of soldiers during the stress of combat, it is possible to reconstruct a picture of how such combats were fought. The evidence from more recent periods suggests that only a minority of soldiers even in elite units actually fought with the intention of killing the enemy. Another, probably smaller minority invariably failed to cope with combat, whilst the majority fought in a limited way, their priority to defend themselves far stronger than the urge to wound the enemy. As a result hand-to-hand combat between massed units was probably a good deal more tentative than our imagination or Hollywood images might suggest.36
The charge across the last few metres separating the two sides was accompanied by increased shouting and culminated in the noise of shield striking shield. Neither side needed to be moving very quickly to create the audible clash of arms described in some sources. It is extremely unlikely that men ran straight into each other hoping to barge into and knock their opponents over, for this risked losing their own balance and a man on the ground during a melee was immensely vulnerable. Nor did the men in the ranks behind push them on in such a physical sense as has been argued by many of the studies of hoplite warfare, since this in turn would only have unbalanced the front ranks. When the rear ranks of a formation pushed too closely behind the men in the lead it put these at a severe disadvantage, preventing them from fighting properly and causing heavy casualties as a result. Attackers began the charge at a run, but if the defenders stood or advanced as steadily to meet them, it seems that both lines checked their pace and then walked or shuffled into actual contact. They would only accelerate their running charge if the enemy gave way before them and it was a question of chasing and striking at their helpless backs.37
Attic helmet of a type common amongst the legions and alae at Cannae. The tubes on either side - one is missing as are the cheek pieces - held tall feather plumes intended to make the wearer look taller.
When the two lines met, the battle became in many respects a series of small duels fought between the individuals facing each other in the opposing front ranks. All of the soldiers engaged at this stage of the fighting at Cannae were primarily swordsmen and only men in the front rank were capable of reaching the enemy with their weapons. Both the Romans and the Spanish used swords with comparatively short blades which could be used to deliver either a cut or a thrust effectively. Many of the Gauls employed a longer type which was primarily intended for cutting. Some of these blades lacked a point altogether and many lacked any significant counterweight, such as a heavy pommel, so that they were unbalanced and end-heavy.
The Etrusco- Corinthian helmet was another type commonly used in this period. It was a derivative of the Corinthian helmet much used by Greek hoplites and had added cheek pieces to give some protection to the face.
This added to the force of a blow, especially a downward slash, but made the sword awkward to wield. The natural first blow for a right-handed man using such a long sword was a downward, diagonal cut to the right side of his opponent's head or shoulders. After the sword had been raised again, it would then have been easier to aim a series of straighter slashes down at the enemy's left side. A warrior fighting in this way inevitably exposed his right arm and some of his right side as he did so. The shorter swords used by the Romans and Spanish could be used in a similar way to deliver downward cuts, although they had less reach and weight than the Gallic long sword, but were also effective as thrusting weapons. Ancient swordsmen stood in the opposite way to a modern fencer, with their right arm furthest from their opponent, since it was vital to have their shielded left side protecting against any threat. As a result, a man could not put the weight of his own body behind a lunge without turning his less well defended right side towards his opponent. Normal thrusts, delivered with no more than the strength and force of his right arm, were unlikely to penetrate an opponent's shield with sufficient force to carry on and inflict a wound. It was therefore necessary to aim blows around the enemy's shield, striking perhaps at his head or face, right arm or lower legs.
A Roman legionary stood with his left foot forward and slightly crouched in the normal fighting position. From the front he was protected by his long, plywood shield which covered his body, left arm and upper legs. Additional protection for his torso came from a mail cuirass or metallic pectoral. His head was the most vital area not covered by the shield and this was why the helmet was the next most important piece of defensive equipment after the shield. The various patterns of bronze helmet in use by the Romans and their allies at this time offered good protection from a blow on the top of the head and, if provided with cheek pieces, a limited degree of coverage for the face. His left leg, the one nearest the enemy and as a result vulnerable, was often fitted with a metal greave. As far as we can tell most of the Spanish and Gallic warriors at Cannae lacked both helmets and body armour. In tribal warfare such expensive equipment tended to be the preserve of chieftains and the wealthy who were always a small minority, although it should be remembered that these well equipped men would tend to take their place in the front rank of a formation. We do not know whether many of these men wore captured Roman armour or whether Hannibal had reserved this exclusively for his Libyans. Most warriors relied exclusively on their long, flat shields for protection, making it all the more devastating a loss if a shield had been lost to a Roman pilum.
It was very difficult to disable an opponent with a single blow; either a heavy strike to the head, a massive thrust past shield and through any armour to the body, or a hit on the leg breaking the bone and causing the victim to fall. Attempting to deliver such a strong cut or thrust exposed the attacker to greater risk of wounding, especially as his right arm, and perhaps part of his right side, lost the protection of his shield. It was less risky to deliver weaker attacks to the unprotected extremities of an opponent, even though this was unlikely to kill him quickly. There is some skeletal evidence from battlefield graves in the ancient period, and rather more from the middle age. The pattern and type of wounds is remarkably consistent and suggests how hand-to-hand combats were fought. (However, it is important to remember that only wounds which involved damage to bone are preserved in this record. Injuries to fleshy parts of the body or the stomach would not leave any trace. Bearing in mind that our information is derived from battlefield grave finds, we have no record of the men who suffered wounds but survived the battle.) The sheer physical force of some blows surprised many specialists, but such single, almost certainly fatal, injuries were rare. Usually the dead suffered a number of lesser wounds, none of which were incapacitating, before being finished off by a heavier blow to the head. The most common were hits to the lower legs, especially the left leg nearest the enemy, the right arm, undefended by a shield, and the left side of the head. Even a number of such light wounds did not seriously impair the man's ability to continue fighting.
La Tene Celtic sword. Long bladed slashing swords much like this example were carried by many of the Gallic warriors fighting for Hannibal at Cannae. These weapons were end heavy, adding to the force of a downward blow, but also making them awkward to wield.
We should imagine the two front ranks separated by a metre or so, prodding and cutting at each other in a constant clatter of blade against shield, helmet and sometimes flesh. Once again individuals hoped that their appearance - physical size, expression, plumes, shiny armour, impressive hair or beard - and the noise they made would intimidate their opponents and aid their victory. Cato the Elder, who served during the Second Punic War as a cavalryman and junior officer, although he probably was not at Cannae, always maintained that a soldier's bearing, confidence and the ferociousness of his war cry were more important that his actual skill with a blade. The majority of men took care to shelter as much as possible behind their shields, warding off blows and occasionally delivering a careful attack themselves. Such men inflicted only minor wounds, weakening an opponent but not putting him out of the fight.
A minority of soldiers fought with far more aggression, aiming savage cuts or thrusts at the enemy, and it was these who inflicted nearly all of the serious injuries, although they in turn suffered a higher proportion of casualties. Some men, especially amongst the Romans, may have used their heavy shields to buffet and unbalance the enemy, punching with their whole weight behind the boss. Ideally, when a man in the opposing front rank was killed or knocked to the ground, the victor stepped into his place. This was highly dangerous for he risked attack from the men to the front and sides in the second rank of the enemy formation, but it was also the best way to begin the enemy's rout. As soldiers began to feel that they were no longer protected on their flanks by their comrades and that enemies were amongst them, their nervousness could quickly turn to panic. It was at this stage in the fighting, when a unit turned and fled, that most casualties occurred, and this was the single most important factor in explaining the far higher casualties always suffered by the losing side in an ancient battle. Men in flight lost the vital protection of their shields and the victors were able to strike freely at their backs. The sight of enemies, who until recently had posed a direct threat to them, turning their backs seems to have encouraged the majority of soldiers, the ones who fought with the intention of staying alive, to act aggressively and expunge their fears in a one-sided massacre of all they could catch. Minor wounds, most of all wounds to the legs, suffered during the fighting could now prove fatal, for the weakened men were often slower to run and more likely to be caught and finished off by the vengeful pursuers. Blows delivered when the victim was helpless were stronger and more closely spaced. It appears that it was not uncommon for the attackers to strike repeatedly at the fallen enemy, so that as many as seven or eight massive cuts were delivered to the skull, any one of which would probably have proved fatal (see illustration on page 158). The savagery of such attacks on already defeated enemies is a powerful reminder that battles, especially hand-to-hand battles, are not fought by calm soldiers fighting coldly, carefully and logically, but by frightened, vulnerable and emotional human beings.38
The Montefortino helmet was probably the most common type of headgear used by the legions in this period, and remained in service for several centuries. Based originally on a Gallic design, and so quite possibly also in use amongst Hannibal’s Celts, it had a high bowl which offered good protection against downward cuts.
This was the most dramatic end to a hand-to-hand encounter, when men from one unit cut their way into the enemy ranks, created a panic and inflicted a brief massacre as the defeated group turned to flee, but it was not the most common outcome of a fight. More often in the initial clash neither side was able to gain such a decisive advantage and if any men tried to break the enemy ranks they were swiftly killed themselves. The very fact that both sides had sufficient confidence to meet in hand-to-hand fighting in the first place, rather than being persuaded to retire or flee by the enemy's intimidating advance and the volleys of missiles, which they had hurled, made it unlikely that either would swiftly give way. Hand-to-hand fighting was physically strenuous and emotionally draining. If one side did not quickly collapse then the actual combat could not continue for more than a few minutes. Instead the two sides seem to have drawn apart, perhaps little more than a few metres, for even at such a short distance they were out of the reach of the enemy's hand-held weapons. There they drew breath, shouted at the enemy and, perhaps, threw any remaining missiles at them.
After each such lull, one side or the other would surge forward into contact again and another brief flurry of actual hand-to-hand fighting occur. If no outside force intervened, then victory would eventually go to the side which endured the stress of staying so close to the enemy for the longest and was still able to urge enough of its men forward to renew the fighting. The pauses in the fighting most probably grew longer and longer as it became more difficult to persuade the weary soldiers to advance and fight another time and another. The great emphasis the Romans placed on encouraging and rewarding individual boldness in their soldiers acknowledged the very real need for aggressive soldiers who would lead a fresh charge forward and try to fight their way into the enemy's formation. These situations also made great demands on an army's officers to lead the way, even though this might mean that they suffered casualties at a disproportionately high rate. The Romans had an optio behind each century to hold the men in place and a centurion in the front rank to urge them onwards. Centurions were supposed to be selected for their determination and skill as leaders rather than individual prowess in fighting, and as infantry combats drew on this sort of stubbornness was especially important. There were six tribunes per legion and the majority of these appear to have fought with the heavy infantry in battle. They were not tied to any one position, but moved around the battle line, encouraging the men and committing reserves as necessary. In addition to these men there were senior officers. Servilius Geminus began the battle with the infantry and he was subsequently joined there by Paullus, who is described as moving to crisis points in the line, leading local charges and fighting hand to hand, and always urging on his soldiers to greater effort. The Roman army at Cannae had an especially large number of senior officers concentrated on a limited frontage to inspire the men and control the battle. Hannibal's decision to advance the middle of his line and provoke a battle first in the very centre acted to reduce even more the width of the initial contact and so concentrate the attentions of so many officers in this limited area. On the Carthaginian side, both Hannibal himself and his brother Mago acted in a similar fashion to the Roman officers, keeping close to the fighting to inspire and direct their men, and there were presumably many junior officers and tribal chieftains performing the same task.39
Early in the fighting it might have been possible to persuade the entire line to advance together, but this sort of order was swiftly lost in the chaos of noise, dust and confusion. As the struggle went on it is likely that it was hard to persuade more than an individual maniple or company to surge forward and renew the fight at the same time, and eventually things may have degenerated further so that only small groups managed to act together. Close contact with the enemy caused unit formations to degenerate into loose masses, and we should never imagine combats as fought by neat blocks of men, the soldiers in perfect rank and file. Under the pressure of combat the less enthusiastic soldiers tried to escape, edging to the rear, whilst the boldest pressed forward. The majority massed in the middle, ready to follow the bold few if their attack proved successful or the more nervous if these started to flee. The fighting line was not a solid wall of men, but a row of increasingly rough groups clustered together, each man’s position a reflection of his keenness. Its openness allowed officers to move around with some freedom, only occasionally actually fighting hand-to-hand themselves, although they were always at risk of being hit by missiles or singled out by a lone attacker. Facing this was another similar line formed by the enemy, the two usually separated by a short distance, save where a unit or group had managed to build up sufficient aggression to charge into contact. It was difficult for most soldiers to know how well even their own unit’s fight was going unless they were in the front rank, and only if they could make sense of the overwhelming noise could they possibly gauge the progress of the fight elsewhere in the line. This created a permanent state of nervousness, since men knew that, if a serious breakthrough occurred in their line and it collapsed into flight, then the men most likely to be killed by the pursuing enemy were the ones who hesitated before they ran.
The third phase of the battle. Hasdrubal rallies his men from the pursuit of the Roman cavalry and reforms his squadrons. The Numidians continue to skirmish with Varro's Italian cavalry. In the centre, the pressure of the Roman advance has started to force back the Gauls and Spaniards. Their most advanced units have now been pushed back level with the main line.
Once again, our sources are vague as to how long the fighting continued after the Romans had reached the advanced centre of the Punic line, but both Polybius and Livy testified to the stiff resistance put up by the Gallic and Spanish infantry to the Roman juggernaut. Here, as at Telamon and some other battles, the tribal warriors belied the literary stereotype of the wild barbarian whose initial ferocity rapidly declined as he grew weary. In numbers the Punic centre was roughly equal to the Roman hastati and, since they occupied a similar frontage, was presumably deployed in much the same depth. Depth gave a formation great resilience in combat and this, along with the presence of so many senior officers, encouraging the men and sharing with them the risks of combat, prolonged the fight. Livy tells us that, ' ... at first equally matched in strength and confidence, the Gauls and Spaniards stood firm for as long as their formation held. At length the Romans, surging forward again and again on an even front and in dense array drove back the advanced wedge [curved line] formed by the enemy which was too thin and weak to hold.'40
The Gauls and Spaniards had no immediate supports, whilst the hastati were just the first of the three Roman lines. The manipular system was intended to allow the reinforcement of the fighting line with fresh troops, with the intention that their enthusiasm would persuade the whole line to surge forward into contact against the weary enemy. The reserve lines could reinforce the fighting line if it was coming under pressure, or advance to exploit any successes and breakthroughs it managed to achieve. The many senior officers, tribunes, prefects and above, who had pressed forward to oversee the fighting in the centre of the line were there not just to inspire the men and witness their behaviour, but also to control the commitment of the second and third lines. However well the Punic infantry fought, in the end, the Romans' weight of numbers would come to bear as more and more maniples were fed into the combat. Eventually the pressure grew too great and the Celts and Spaniards began to give way. They did so slowly at first, perhaps moving back after each flurry of fighting, but still facing the enemy. We read in accounts of other battles of the ancient world of lines which were forced back several hundred metres or even more than a kilometre, but still maintained a front and did not dissolve into rout. At Cannae the Punic centre at first gave way gradually in this fashion, but, as the Romans poured more and more men into the main line to exploit this success, the line broke and ran. As usual, the Carthaginian foot seem to have suffered very heavy casualties as they fled from a vengeful enemy.41