Polybius tells us that in 216 BC the Senate decided for the first time in Rome's history to muster an army of eight legions, four for each consul. However, in his account of the Gallic invasion of 225 BC he also appears to say that the consuls were given four-legion armies, but this may be a misreading of the text from a statement that there were four legions in total, two for each consul. Another important difference was that in 225 the two consuls were not supposed to join forces and it was luck, rather than design, which brought both armies into contact with the Gauls at the battle of Telamon. In 216 BC both consuls with all eight legions were expected to fight together. Not only were there more legions than usual, but each unit was increased in size to 5,000 infantry and 300 cavalry. As usual, the number of Roman infantry was matched by the allies who also provided a higher proportion of cavalry.
This scene from Trajan's Column shows Roman legionaries harvesting grain, wearing the segmented armour of the Imperial Period. Maintaining an adequate supply of food was a continual problem for army commanders and a major factor in their decisions during the campaign before Cannae.
Polybius tells us that the army finally mustered 80,000 infantry and 6,000 cavalry, making it one of the largest forces ever put into the field by the Romans.5 Polybius' testimony is clear, but Livy says that the different sources for the battle included a huge variation in the size of the Roman army. He mentions the Polybian figure of eight legions of 5,000 foot and 300 horse, but according to another tradition, the army of four legions commanded by the consuls of 217 outside Gerunium was only augmented by a levy of 10,000 new recruits. If the latter account is true, then the army at Cannae may have been no larger than around 50,000 men. Several modern scholars have favoured this version, arguing that, since numbers were frequently exaggerated in ancient accounts or distorted as the manuscript was copied over the centuries, it is always better to accept the lowest figure. In this case at least, such a view is surely mistaken. Polybius is our earliest and most reliable source and his account is utterly consistent in assuming that there were eight legions. It would be very rash to reject such positive testimony from this source. Livy's narrative also assumed the higher estimate, when for instance he mentions that twenty-nine tribunes were killed in the battle and names six of the survivors. Since there were six tribunes per legion, this would indicate that at least six legions were present. Both sources also emphasize how great an effort the Romans made in this year in their determination to crush Hannibal. An army of 50,000 men would have given them no more than rough parity with the enemy and would not have been much bigger than the army soundly defeated at Trebia. The course of the campaign and battle certainly make far more sense if the full eight legions were present. It is possible that a reinforcement of about 10,000 men was needed to bring the army at Gerunium up to the strength, compensating for losses in the last campaign as well as the need to increase the size of each legion. In addition four complete legions plus allies were raised and sent from Rome to join the army. This reconstruction would accommodate both traditions, but must remain conjectural.6
The Roman legions in this period were not the long-lived institutions of the later professional army and appear to have been re-numbered each year. It is extremely rare for our sources to explain in detail when units were raised, disbanded, destroyed or incorporated into other units. The army at Gerunium consisted of the legions allocated to Servilius Geminus at the beginning of 217 and the units raised by the dictator, a total of four legions. Geminus' army was based around some of the survivors of Trebia, most probably Scipio’s two legions, for Flaminius seems to have taken Sempronius Longus' men. These units had been badly mauled, first in the fighting with the Boii and Insubres, and then at the hands of Hannibal, so it is virtually certain that they had received strong drafts of replacements. In the aftermath of Trasimene most if not all of Geminus' cavalry had been killed or captured. Since then, all four legions had campaigned with Fabius Maximus and Minucius, winning some minor skirmishes, but also suffering some defeats. These were the best troops in the army. Many of their men were experienced, and, most importantly, that experience had been gained recently, alongside the same comrades and under the command of familiar officers. However, their service had not been dominated by successful fighting and it was only this, combined with good training and leadership, that was believed to raise an army to the peak of its efficiency.7
The four new legions were all recruited in late 217 or early 216. It took a good deal of time to organize each legion, dividing the recruits into their subunits of century and maniple, decury and turma, and appointing officers to command them. It took even longer for them to train and drill, gradually developing trust in each other and their leaders to create a confident and effective unit. In spite of the earlier defeats it is clear that the Romans had not become daunted by their foe, but rather grimly determined to achieve victory. Livy says that this mood was reflected when the tribunes of each legion led the men in taking a formal oath 'Never to leave the ranks because of fear or to run away, but only to retrieve or grab a weapon, to kill an enemy or to rescue a comrade.’ In the past he claims that this oath, the sacramentum which in a similar form would survive in the Roman army for centuries, had been taken voluntarily by the soldiers in their own centuries.8
Another indication of the mood of the times was the high number of senators and equestrians serving with the army. The Roman aristocracy earned the right to rule the Republic in peace through its willingness to provide leadership and risk its lives on the State's behalf in wartime. Between a quarter and a third of the Senate was present at Cannae, and most of the remaining senators had sons or other close relatives with the legions. Marcus Minucius Rufus, Fabius' Magister Equitum and a former consul, was one of many distinguished men serving as a military tribune or on the staff of one of the consuls. The tribunes of this year were, in general, far more experienced and capable than was usual. Of the six named tribunes who survived the battle, four went on to hold the consulship. It is also instructive to look at the four praetors elected for the year. Three, Marcus Claudius Marcellus (consul 222, praetor 224), Lucius Postumius Albinus (consul 234, 229, probably praetor in 233), and Publius Furius Philus (consul 223), were former consuls and the other, Marcus Pomponius Matho, had already held the praetorship in the previous year.
This scene from the altar of Domitius Ahenobarbus seems to show the enrolment of soldiers or perhaps the census of all citizens held every five years. The census listed each citizen's property and thus made clear the capacity in which he would serve when required for military service.
Marcellus and Albinus, both of whom were given field commands, in Sicily and Cisalpine Gaul respectively, were of the same generation as Fabius Maximus, men who had reached maturity during the First Punic War. The Romans were relying upon experience in the present crisis. It is another indication of Varro's widespread popularity that he achieved the consulship in a year when the magistrates were such a distinguished group.9
The field army in 216 was four times larger than a normal consular army and no one had experience of handling such a huge force. Tradition, created on those very rare occasions when both consuls had fought together, dictated that command should be held by each consul on alternate days, emphasizing just how improvised the system for controlling this army would be. It is doubtful whether the four legions massing at Rome had much time to spare for training together and certain that there was no opportunity to integrate this force with the existing four-legion field army when the two were united. The Roman army in the Cannae campaign was large, and its officers and soldiers brave and enthusiastic, but there had simply not been the time to weld them into a coherent force. The army was capable of only the very simplest of manoeuvres or tactics, far less flexible and supple than Hannibal’s veteran force. This difference was to have a fundamental influence on the forthcoming campaign.
The army sent against Hannibal was not the only one fielded by Rome in 216, although it was by far the largest. In all between fifteen and seventeen Roman legions were in service by the summer of 216, a total of perhaps 75,000-85,000 men, around a quarter of the number of citizens Polybius says were eligible for military service in 225. Supporting these were a similar number of allied soldiers.10