contingents, each under their own leaders in much the same way that they would fight for their own people. However, at Cannae Hannibal's centre consisted of alternate units of Spaniards and Gauls, clearly breaking up any tribal structure they possessed. Polybius uses one of the terms he also employs for the Roman maniple of 120-160 men, and the same term was used by later authors for the cohort of 480 in the Late Republican and Imperial army. This makes it probable that these 'companies' consisted of a few hundred men, certainly less than a thousand. 14
The mixture of contingents from different nationalities usually provided Carthaginian armies with a good balance of different troop types, with both close and loose order infantry and cavalry. Many of these contingents were of high quality, although their standard of discipline varied considerably. It was rare for troops whether serving as allies or for pay to fight without enthusiasm, and mutinies were uncommon. An additional element was provided by the fairly frequent use of war elephants who might well panic an enemy unused to them. The elephants employed were probably African Forest elephants, somewhat smaller than Indian elephants, but more amenable to training than today's African elephants. The elephant was the main weapon, using its bulk and strength to terrify or crush opposition, but Hellenistic armies also mounted towers on the animals' backs, from which crewmen hurled or fired missiles. There is no direct evidence indicating that Punic war elephants also carried towers, but Polybius' account of the Battle of Raphia in 217 BC implies that the African breed was capable of carrying the extra weight. The main danger with elephants was that they were inclined to panic and might then trample friend and foe indiscriminately. Hasdrubal is said to have equipped the drivers, or mahouts, with a hammer and a chisel-shaped blade, which they were supposed to drive into the animal's spine to kill it if threatened to stampede towards friendly troops. 15
Carthaginian commanders usually had well-balanced forces at their disposal, but the difficulty lay in co-ordinating the movements of these disparate elements. Orders issued in Punic had to be translated into various languages in order to be conveyed to the soldiers. Carthaginian magistrates, such as the suffetes, did not hold military commands. Instead generals were appointed, although it is not clear precisely by whom, and usually held command on a semi-permanent basis until they were replaced or for the duration of a conflict. Although not serving magistrates, it is clear that the commanders were drawn from the same social class who filled these offices and there is no reason to believe that ability, more than family connections and wealth, was the main reason for their selection. In the First Punic War the Carthaginians continued their traditionally harsh treatment of commanders who failed, several men being crucified for incompetence. In several cases this penalty was inflicted on them when they lost the confidence of the senior Punic officers under their command.
However, the long duration of the commands which they were given did mean that many Carthaginian commanders became highly experienced. The longer a general held command over an army the more efficient it tended to become. Gradually, the disparate elements composing it became accustomed to operating together, their leaders and the higher commander became familiar with each other and, at least to some extent, their languages. The army which Hannibal led into Italy in 218 was probably the finest Carthaginian army ever to take the field. Its efficiency was in part the result of its commander's ability as a leader, but was more the product of long years of hard campaigning in Spain under the leadership of Hamilcar, Hasdrubal and Hannibal himself. During this time its command structure had developed to a high level, and this, as well as its march discipline and ability to