It may seem odd to begin a discussion of the capabilities and drill of Alexander the Great’s phalanx with a battle that occurred 126 years after his death but there is a good reason for doing so. Kynoskephalai has the advantage of having been written by a historian who also wrote on the formations and evolutions of the Macedonian phalanx, Polybios. Further we also have a parallel account based, in large part, upon this account by Titus Livius, which in turn allows us to evaluate his account of the other battles whose Polybian originals are lost, such as Magnesia and Pydna.
Also, Polybios has recorded some interesting details of the manoeuvres of Philip V’s army which are obscured in some translations and misinterpreted in modern accounts.
In his 1988 article ‘The Campaign and Battle of Cynoscephalae 197 BC’ (Journal of Hellenic Studies 108, pp 60-82) N G L Hammond details the strategic manoeuvres and identifies a battlefield which seems to fit the ancient descriptions and satisfy the supply requirements of both armies. This is available on JSTOR for ordinary members to read online and I recommend those interested sign up for the free access. I will assume his site is broadly correct not least because he surveyed the land in person and found what may well be related archaeology, possibly the remains of some Roman camps.
The first thing to point out is that, whilst being quite a full battle description, Polybios does not detail every manoeuvre on both sides nor, even for either side. However, there enough clues to reconstruct these with a fair degree of likelihood.
The situation is fairly unusual as both sides had encamped close to each other but in ignorance of the other’s presence. Both sides then sent out light forces to hold the pass in Philip’s case and to reconnoitre the ground in Flaminius’. Philip even sent some of his phalanx out to forage. These are characterised as belonging to the left wing of the phalanx (since he took the peltasts and the right wing with him), this is an important point as the ancient way was to put the best troops on the right. It would follow that these were the raw troops whom Philip had levied in the emergency. Ultimately, this would be Philip’s undoing.
An affair of outpost ensued when both advance forces ran into each other and the advantage swayed as re-enforcements were sent in from the respective camps.
Polybios, ignores the time that it would take to get troops from the camps into formation and then into action. We need not follow his example.
We have good descriptions of Roman camps and can work out the time it would take to deploy from them. Hammond’s survey found that the camps at Kynoskephalai were smaller than those described in Polybios and Hyginus and suggested that by analogy with the well preserved camps about Masada that the one full one and its partially preserved neighbour would have held ¼ of a legion and that there would therefore have been sixteen of these scattered around the Thetideum. The problem here is that ¼ is not a natural division of a legion, it would represent 2 ½ of each maniple. It is more likely, in my opinion that each held two cohorts, ie two maniples each of hastati, principes and triarii along with the requisite cavalry and velites. This would allow each camp to have the exact mirror symmetry of a larger camp and the troops would be able to issue in an orderly fashion and re-assemble in an equally orderly manner.
Hammond gives the dimensions as 66 by 88 by 80 by 97 paces (N, E, S, W sides); the Romans used anthropomorphic measurements and their passus of two gradi was equal to 2 1/2 Roman feet. Camps were laid out according to square acti which is 120 pedes by 120 pedes or 48 by 48 paces. 480 men required two acti, which is roughly what we have here, which would mean that one cohort is all that could be encamped here; 120 hastati, 120 principes 60 triarii , 120 velites and 30 cavalry. This would mean that Flamininus had his men construct forty camps, rather than the sixteen suggested by Hammond, and the Aetolians and other allies on top of that. For this reason I think we have to reject the idea that what Hammond found is part of Flamininus camp system. (This website discusses aspects of the Roman army in much greater detail than would be appropriate here, http://garyb.0catch.com/site_map.html).
It has been calculated that it would take four legions and their cavalry 1 hour 27 minutes to leave camp if only one gate were used and 43 minutes if two were used.
The army would then have to form up; in ‘triplex acies’ four legions (two Allied, two Roman), stretch for 1200 divided by six (usual depth) times six feet (frontage according to Polybios) or three feet (modern assumed fighting frontage) per legion, either 2400 feet or 4800 feet. The depth of the formation would be 15 times six feet plus the intervals between the lines 90ft plus say 60 feet (two ten yard intervals). So the man, who has to march the furthest, has to march 2418 feet (on a three foot frontage). 3 mph is 264 feet per minute, so, from a central position it would take him about nine minutes to reach his station, this does not delay the extracastramentation at all, even on a six foot frontage, as he will be in position before the last man leaves the camp. So we can say that it would only take an hour and a half to deploy a full Consular army for battle from the camp.
Flamininus’ army did not deploy fully in one go, but we can use the calculations above to establish a timeline for the battle.
Philip V’s manoeuvres can also be roughly timed, but first let us just look at the proposed site of the battle. Map 2
Hammond places it on the western spur of the hills (H), but this ignores the fact that Philip’s advanced force was guarding the pass, presumably the route now taken by the main road (P); as they fell in on the right flank of the phalanx and that phalanx deployed ‘as soon as they reached the heights and filled the line from the left’ ie the right wing deployed first it is more likely that the battle site is to the east of the road (B).
We cannot, yet make a guess at the length of Philip’s approach march, but as he had the peltasts and the right half of the phalanx we can work out the length of his line-of-march and how long his deployment would take.
10,000 men four abreast, (column by tetrarchiai) at a six foot interval would stretch for 15,000 feet, and at sixteen deep on the same interval they have a frontage of 3,750 feet ; the rearmost section, then, have to move 18,750 feet once the first files begin deploying. We are told that Philip set off at the double so allowing them to move at 4mph (352 feet per minute) we can say that this phase would take about 55 minutes. To convert this into an eight deep fighting line, the even numbered soldiers in each file step diagonally to the left and the ranks close up. This is the reverse of how Aelian describes the normal method of doubling the depth, which is what Philip does to reverse this manoeuvre. It seems to me more likely that Philip would initially deploy at half depth in order to cover the ground and then return to standard depth, rather than reforming his line thirty-two deep, a depth adopted by Antiochos at Magnesia and condemned as unwieldy and useless, one might expect one of the sources to have commented on its equal failure in the earlier battle, or even its initial success.
We are also told that when Philip first reached the ridge the fighting was close to the Roman camp and that while he was still deploying the Romans broke his advanced force, yet they did not press their advantage, instead they withdrew their own advanced forces through the gaps between the maniples, which allowed Philip’s force to rally on the right flank of the phalanx, despite Poly bios’ statement that Flamininus had all of his troops in line at this time one would have to doubt that. For one thing the whole force would take an hour and a half to get out of camp and yet Flamininus is stated to have only begun deploying his legions once his advanced party had been driven off the ridge by Philip’s forces, almost the same time that Philip reached the ridge with his forces.
Flamininus was no military incompetent were all his forces in order when Philip started deploying he would surely have attacked at once and overwhelmed the Macedonians much earlier in the battle rather than delaying to allow a victorious force to withdraw through his lines and thus losing the initiative. I would suggest that Flamininus halted and took the advanced force into his lines in order to allow more of his men to get out of camp.
Philip’s phalanx would have been deployed an hour after reaching the ridge, this movement was incomplete when Flamininus attacked with his left wing; if we allow Flamininus forty minutes he could have almost deployed the two legions of his left flank, with the assembled force he would overwhelm Philip’s advance guard, but from his position the whole ridge would seem occupied and thus himself outnumbered three to one (if Philip’s 10,000 were eight deep and in standard intervals they would occupy 3750 feet, Flamininus’ two legions would present a frontage of 1200 (at 3ft)).
While Flamininus was allowing his men through his lines, Philip readjusted his line. Hammond and others suggest he made his right wing 32 deep but this seems unlikely. More likely he saw the left wing mounting the ridge and needed to make room for it to deploy. Closing to the right means a move of 1875 feet for the leftmost file, which at 3mph would take 7 minutes and would certainly be achievable before Flamininus could reach the ridge.
The concomitant of Philip doubling his depth and closing to the right was that Flamininus could now see a force roughly equal in frontage (which in turn implies that the Romans were formed on a three foot frontage per man. It would be now that he decided to attack with the fully formed left wing, since the right was not committed, the logical reason would be that it had not yet extricated itself from camp.
Philip seems to have made one further movement. Polybios clearly states that the phalanx was unable to face about (metabole) when Flamininus’ unnamed tribune launched the decisive blow against its rear. This is a mentioned side-effect of being in ‘synaspismos’ and Plutarch (Vit.Flam. says
This could have been achieved by the same method as the previous halving of depth or by the insertion of the rear half files. Aelian would seem to imply that ‘insertion’ was used when moving from ‘pyknosis’ to ‘synaspismos’ . What seems certain is that the files did not simply close fully to the right as that would make the phalanx have a frontage of only 938ft which would leave it outflanked by the Roman line which had the flexibility to attack its flanks.τὸ βάρος τοῦ συνασπισμοῦ
The weight of their interlocked shields
The weight of the phalanx attacking downhill was more than the Roman left could take and Flamininus left his losing wing and led his now formed right into the attack; Philip stuck with his winning right wing a la Alexander.
Unfortunately for the Macedonians Nikanor, known as ‘Jumbo’, had not led the left wing to the ridge but merely sent them on as they assembled from foraging. Polybios is clear that it was lack of leadership that left this wing in a confused and undeployed state, combined with the broken terrain. Two things jump out from this, that the lower hegemonate of the Macedonian army lacked the nouse and authority of the comparable officers of the roman army; here it was a tribune who won the battle decisively (slightly more senior than a pentekosiarch), at Pydna the initiative went as low as the centurions. Also, the Romans could leave camp, form up and attack with two legions before the left wing of the Macedonian phalanx could march 1200 yards.
That the Macedonians made errors is not in doubt but it was the adaptability of the Romans that won the battle.
Polybios in his discussion of the phalanx states that the Romans fought on a six foot frontage and so that each legionary face ten pike points, yet in the above things work better if the Romans are on a three foot frontage. I think this can be reconciled if one bears in mind that in Book XII Polybios fails to realise that Alexander’s army at Issos did not enter combat at marching intervals; he was a practical soldier but not always alive to what was being said by his sources. Given that Philip may have adopted ‘synaspismos’ (a formation in which Antigonos Doson attacked uphill at Sellasia) a Roman on a three foot frontage would face ten pikes, since he does not report the move to ‘syaspismos’ it is possible that Polybios extrapolated from being told that there were two pike heads slamming into each scutum and, since he assumed the Macedonians were at a standard density he got the Roman frontage wrong. Almost heresy , since Polybios did observe the Roman army in the field. Mmmh…