Below is a sample of some of the unit sizes given by Livy for the year 171 BC, during the Third Macedonian War.
Livy (42 55) 171 BC. Eumenes set out with a force of 4,000 infantry and 1,000 cavalry...Athenaeus with 2,000 infantry remained in Chalcis. The Apollonian sent 300 cavalry and 100 infantry, the Aetolians made up one division of cavalry, the Thessalians sent "not more" than 300 cavalry, and the Achaeans furnished 1,500 men.
Livy (42 57) the Romans despatched a reconnaissance force of 100 cavalry and the same number of infantry (100). Later, Livy mentions that this Roman force comprised two troops of Gallic horse, and "about" 150 light infantry made up of Mysian and Cretans. Perseus, uncertain of the strength of the Roman force, sent from his main body two squadrons of Thracian cavalry, two squadrons of Macedonian cavalry, two cohorts of Cretans and two cohorts of Thracians. In this engagement, Livy gives the Roman losses at "about" 30 men belonging to Eumenes.
Livy (42 59) The Macedonian centre consisted of the king was surrounded by the "agema" and the troops of the "sacred" cavalry. In front of these were stationed 400 plus slingers and javelin men...In front of the Roman wings were stationed 200 Gaulish troops, 300 Cyrtians troops of Eumenes, and 400 Thessalian cavalry stationed a short distance beyond the Roman left wing.
Livy (42 60) According to Livy, the Romans lost 200 cavalry and "not less than" 2,000 infantry, and "about" 600 Romans were made prisoner. Livy (42 60)
These numbers, and many more are based on Roman and allied units within a consular army and they also prove that the Roman legion of 60 centuries is, for want of a better word, the legion's macro-organisation. Livy's figures above are adhering to the legions' micro-organisation, which has far more than 60-centuries per legion. This micro-organisation I stumbled across some years back but did not properly fully comprehend its significance until now. The micro system has now allowed me to decipher those numbers found in the Third Macedonian War that have been causing me a lot of grief for some years, in fact, too many years.
Livy's (42 57) figure of "about" 150 light infantry (Mysian and Cretans), converts to 144 hastati. The engagement only involves velites, hastati and cavalry. At Livy (42 60) the Romans loss of 200 cavalry equals 192 cavalry, and the "about" 600 Romans made prisoner converts to 576 hastati, which when divided by the 144 hastati already mentioned, equals 144 hastati x 4. Now Livy (42 57) gives the Roman losses at "about" 30 cavalry, which converts to 24 cavalry and the 192-cavalry divided by 24 cavalry produces 8 cavalry squadrons, so those 400 Thessalian cavalry converts to 384 cavalry (16 cavalry squadrons). And as the whole Macedonian army has been borrowed from a consular army, Livy (42 66) has the Macedonian army leave the battlefield with 300 Macedonian infantry lost and 24 Macedonian cavalrymen belonging to the sacred cohort. So, this converts to 288 hastati (144 x 2) and a Roman or allied cavalry squadron of 24 cavalrymen fabricated as Macedonians.
The golden rule seems to be that the numbers in the macro-organisation of the legion ends in a zero, and the micro-organisation ends in any number from one to nine (ex.288).
The Third Macedonian War has for the Roman introduce new military doctrines. Livy claims the 5,200-man legion for Macedonian was increased to 6,000 men, but this is hogwash as is the 5,200-man legion. The true increase in the size of the legion is that each legion has been allocated garrison troops (both infantry and cavalry), as there is a myriad of data that supports this, in fact Livy even mentions garrison troops. So, rather than taking the garrison troops from the legion, for the Third Macedonian war they brought them with them.
Steven, you seem to use casualty (inc prisoner) figures as a guide to unit strengths here. What is the thinking there, as I'd expect casualties to be distributed across the force? Or do the sources imply whole units wiped out or surrendering?
Quote from: Erpingham on Nov 21, 2025, 02:46 PMSteven, you seem to use casualty (inc prisoner) figures as a guide to unit strengths here.
That is what the ancient writers are doing. Those casualty figure are not real casualties. Ancient sources just deduct or select from the force a unit size and convert them into killed or captured. The common pattern I have found is the ancient sources like to convert the allied infantry and allied cavalry in a consular army, into killed or captured.
The bottom line is the ancient writers do not use historical casualty figures for the Romans or their enemies. They probably don't have any. The one exception is the 3,505 men killed in Pyrrhus army as claimed to be taken from Pyrrhus' own diary could be historical. However, that figure is very close to the Roman camp guards, but to arrive at 3,505 men, it involves rounding the individual troops types, then having some being omitted, like half or a quarter, a method I found quite convoluted, and it does not follow the simple methods used by the ancient writer's sources.
In some battles, like Cannae, Livy's causalities figure of 45,500 infantry killed simply omits the standard velites that were on the battlefield facing Hannibal's army, thereby only including the specialist velites. And yes, there are two types of velites. Livy's figure of 45,500 causalities is deadly accurate, no rounding whatsoever, just took the actual figure of the Roman and allied infantry, deducted the camp guards and the stand velites for 14 legions. Presto, 45,500 infantry, consisting of the specialist velites, all the hastati and the princeps. No triarii, they guard the camps with their allocated velites, both standard and specialist. Every figure including Hannibal's losses at Cannae have been lifted from the Roman army. In fact, in many cases, the missing men are converted into enemy causalities (killed or captured.
Found in the "Acts of the Disputation of Archelaus (1-2)," in 277 or 278 AD, Archelaus, the bishop of Caschara was offered a sum of money for a large number of Christian prisoners (men, women and children) captured by the soldiers of the camp. The captured Christians consisted of "some" 7,700 prisoners, of which 500 were wounded, and "about" 1,300 Christians were killed.
Again, these number conform to the organisation of the Late Roman army as do many of the data concerning the Christian martyrs. Some like 252 soldiers or 1,104 soldier are spot on. In fact, 1,104 soldiers convert to 1,008 infantry and 96 cavalry. The 1,080 infantry includes officers and supernumeraries, as does the 252 infantry. So, the practice is common place and my first example is the battle of Silva Arsia in 509 BC. For the causalities for Silvia Arsia, the method is to deducted the camp guards from the main Roman army, and then from the residue, deducted the camp guards again, and then round the total figure ending in 20 men to zero.
Another is the battle of Actium. Octavian has eight legions, five cohorts of praetorian unaccounted, and 230 ships. These are figures found in the ancient sources. If you know Roman military doctrine for the Roman navy during this period, there should be over 60 reserve ships in Octavian's fleet, which are rounded and converted to Cleopatra's fleet of 60 ships. Anthony's fleet is given at 170 ships, so Octavian's 230 ships minus Cleopatra's 60 ships leaves 170 ships. With the five cohorts of Octavian's praetorian guard, the actual size of Octavian's fleet can be reconstructed, which includes its battle array and organisation. This information can also be confirmed with other references to the Roman fleet for this period. So, all in all, the size of Mark Anthony's fleet is unknown, as is the size of the Macedonian army at Cynoscephalae and Pynda.
Quote from: Erpingham on Nov 21, 2025, 02:46 PMWhat is the thinking there, as I'd expect casualties to be distributed across the force?
The problem here is your methodology. There is an old saying "expectations lead to disappointment."
I've been continuing my examination into this micro-Roman army organisation. I made a table of the system and it makes identifying the data as presented in the ancient sources.
In 171 BC, Livy has both consuls levy four new Roman legions, two Roman legions for each consul, with each legion having of 5,200 infantry. The two Roman legions levied for Macedonia had a special provision and were raised from 5,200 infantry to 6,000 infantry. The allied contingent for Macedonia was also raised, with the consul Licinius Crassus, transporting to Macedonia 16,000 allied infantry. For Italy, Livy writes that the two Roman legions levied for Italy each amounted to 5,200 men, and were allocated 12,000 allied infantry. Livy's infantry figures amount to 60,800 infantry (32,800 Roman and 28,000 allied infantry).
The 60,800 infantry when divided into centuries of 80 infantry, produces 760 centuries.
The 12,000 allied infantry (150 centuries) allocated to Italy should be allocated to Macedonia, which would accompany the 12,000 Roman infantry (150 centuries), thereby giving a total of 24,000 infantry (300 centuries). Livy's gives the allied contingent to Macedonia at 16,000 allied infantry (200 centuries) consist of both Roman and Italian infantry levied for Macedonia, which includes their allocated garrison forces. It has been reduced to 16,000 because some of those infantry have been, given to Eumenes. Oh, look, Livy claims that Eumenes and Masinissa's troops, was "not more" than 7,000 infantry and 2,000 cavalry. So, 16,000 infantry plus 7,000 infantry makes 23,000 infantry. Now we are talking. This is how the ancient sources work, you have no numbers for Eumene's forces, so give them a proportion of the Roman army in Macedonia. And funny how those numbers conform to the Roman military organisation, so not just random number that have no organisation behind them.
Livy gives the size of the Roman army for Macedonia in 171 BC at 39,000 men (37,000 infantry and 2,000 cavalry). Take away the 16,000 infantry and this leaves a residue of 21,000 infantry. So, 2,000 infantry are missing from 23,000 infantry. Livy reports that in 171 BC, the Macedonian army amounted to 43,000 men (39,000 infantry and 4,000 cavalry). So, after deducting the 16,000 so called allied infantry from the 39,000 Macedonian infantry, bingo, this leaves a residue of 23,000 infantry. Hm, where have I seen that before. Whoever Livy's source is I like him, in many incidences, he keeps it tight. For example, Livy provides a list of the Macedonian allies that amount to 13,500 men, consisting of:
2000 caetrati (Agema)
about 3000 caetrati
about 3000 Thracians
2000 Gauls
3000 Free Thracians
500 Aetolians, Boeotians
13500 men
As they are Macedonian allies, Livy' source has plundered the Italian troops belonging to the consular army in Macedonia, and the 13,500 Macedonians has been rounded from 13,520 men. That is tight.
In 171 BC, Eumenes, accompanied by this three brothers, Attalus and Athenaeus, arrived with his fleet at Chalcis. From Chalcis, Eumenes set out with a force of 4,000 infantry and 1,000 cavalry to unite with the consul Publius Licinius Crassus, while Athenaeus with 2,000 infantry remained in Chalcis. Livy further writes that other contingents of troops form all the Greek States arrived at Chalcis, there "numbers so small that they have passed into oblivion." The Apollonian sent 300 cavalry and 100 infantry, the Aetolians made up one division of cavalry, the Thessalians sent "not more" than 300 cavalry, and the Achaeans furnished 1,500 men armed in the Cretan style. Livy's given figures amount to "about" 9,200 men (7,600 infantry, and "about" 1,600 cavalry).
Besides the troop numbers, an important clue is the mention of the three commanders, Eumenes, Attalus and Athenaeus. This means the 7,600 infantry must be proportioned into three bodies Having already know what this force is, you divided the 7,600 infantry by two, which leaves two bodies of 3,800 infantry. So, one body of 3,800 has been rounded to 4,000. With the next body of 3,800 infantry, you deduct the 1,500 Cretan style troops, which leaves 2,300, which has been rounded and converted to the 2,000 infantry that remained at Chalcis with Athenaeus. Once this is done, it is just a simple matter of allocating the troops types to those numbers. The 100 Apollonian infantry has been rounded and convert from 96 allied triarii, which belong to those 1,500 Cretan style troops. How they got stranded by themselves is anyone's guess. The 1,600 cavalry has been rounded from 1,560 cavalry. A lot of cavalry numbers provided by Livy are actually infantry. The Thessalian cavalry during this engagement played an important part, and because Livy or his source had no actual numbers for the Thessalian cavalry, Roman infantry numbers were substituted.
In 171 BC, the praetor Gaius Cassius Longinus was given command of the fleet, and sailed from Rome to Dyrrhachium with 40 quinqueremes. To this fleet, the allies provided 83 ships, consisting of:
1 trireme from Rhegium
2 ships from the Locri
4 ships from the Sallentines
10 ships from Dyrrhachium
12 ships from Issa
54 light ships from Illyria (Gentius)
83 ships
A few paragraphs latter, Livy mentions 17 ships being assembled at Chalcis:
2 Punic quinqueremes
2 Pontic triremes
4 ships from Chalcedon
4 ships from Samos
5 Rhodian quadriremes
17 ships
Well, this listing is so accurate that I was able to identify which Roman ships were what and the troop types they carried, which not much effort. Literally, paint by numbers. Again, this is why I like this source, a straight transfer of the information.
At the battle of Callinicus, Livy mentions Perseus posted 400 slingers and javelin men, in his front line. The Roman centre was commanded by Quintus Mucius with a picked body of volunteer cavalry. To their front were posted 200 Gaulish troopers and 300 Cyrtians belonging to Eumenes. On the Roman left were 400 Thessalian cavalry. Livy's numbers amounted to 1,300 men, which has been rounded and converted from 1,344 allied infantry, which is keeping to tight, and following my table, no triarii are present. As in the other example, the 400 Thessalian cavalry, are infantry and more importantly, allied princeps. Livy's 1,300 men deducted from the unrounded figure of 2,300 men left at Chalcis earlier, leaves a residue of 1,000 men. So, 1,300 and 1,000 = 2,300 men plus 1,500 returns as to the figure of 3,800 men explained earlier. Again, recycled numbers.
In 171 BC, the praetor Lucretius was besieging Haliartus. A picked force of 2,000 men, was to act as a diversion and attack a breach in the wall, while the rest of the army would attempt a simultaneous assault all around the walls to scale the remaining walls. The attack proved successful, with the Macedonian defenders amounting to "about" 2,500 combatants that had taken refuge in the citadel eventually surrendered. Livy gives Lucretius a force of 10,000 marines and Athenaeus 2,000 troops, which can be ignored. The 2,000 picked troops attack the breach in the wall, which leaves 8,000 troops and when divided by the three remaining walls, allocates each wall 2,666.66 troops. Now without rounding, the result is as Livy states "about" 2,500 troops per wall, which has been converted to those Macedonian troops that took refuge in the citadel. So, the Romans attacking one wall were the ones who run to the citadel. How amusing.
In another battle in 171 BC, Livy writes that some authorities believed a great battle was fought and that 8,000 of the enemy were killed, two Macedonian commanders, Antipater and Sopater were captured, 2,800 Macedonians made prisoner and 27 military standards captured. For the Romans, "above" 4,300 were killed, and five standards belonging to the left wing were lost. Now those five standards are extremely important, because when the unrounded figure is divided by five, it reveals the number of infantry and cavalry, what organisation is in play, and confirms they are all allied infantry and allied cavalry. Also, if we take the 4,300 Romans killed and deduct the 2,800 Macedonian made prisoner, we are left with those 1,500 men, which again are those men in the Cretan style. So, again, recycled numbers. The 2,800 Macedonians made prisoner was earlier given by Livy as consisting of 200 Roman cavalry lost, "about" 600 Roman prisoners and "not less than" 2,000 infantry killed.
In 170 BC, Livy has part of the defenders of Uscana that surrendered consisted of 4,000 Romans and an Illyrian cohort of 500 men. Livy (43 18) Now if you round down the 4,000 to 3,800 and add the 500 Illyrians, you get 4,300. Again, recycled numbers. In 170BC, we find some familiar number popping up again, like 1,500 Romans prisoners, which are the same 1,500 Cretan style troops of 171 BC.
What this micro-organisation system turns out to be is the Roman campaign system, that is the campaign divisions the men are undertaking being organised into their own little centuries, maniples and cohorts. This is another interesting aspect of my research, and one that I cannot explain. Each period of the military history of the republic, concentrates on a certain military doctrine and the numbers involved for that doctrine. For example, the Third Samnite War, details the number of camp guards that are used as an emergency force. At the battle of Beneventum, the consul, Manius Curius ordered the camp guards to attack the elephants. Not all the camp guards are committed, and the number of Samnite killed and captured are based on these numbers, in fact the numbers are quite accurate. The data in the Third Macedonian war concentrates a using the campaign division data.
There were three areas of contention I have before I am completely finished my research, the first is writing a conclusion to the Publius Scipio's African campaign, the second is the two reforms of the Late Roman legion and the third is the Third Macedonian war, which I have now eliminated, which has been a major thorn in my side. My method of research is the blitzkrieg method. If I ancient sources offer stiff resistance, I just isolate it and keep moving. Why I was finally able to understand the data in the Third Macedonian war was due to exposing how the various campaign system for the soldiers actually worked, and the number of men in each campaign division. This could not have been achieved without the use of Pythagorean mathematical principals.
Some scattered updates on the Third Macedonian War, which thankfully is now finished.
In 172 BC, following Livy, the praetor Gnaeus Sicinius was ordered to raise 8,000 allied infantry and 400 allied cavalry. The propraetor Aulus Atilius Serranus was chosen to receive the soldiers at Brundisium by the thirteenth of February and then send them to Macedonia. In order that the praetor Gnaeus Sicinius should have an army ready to cross over into Greece, the praetor Gaius Licinius on authorization of the senate wrote to the consul Gaius Popilius to send both the Roman second legion and 4,000 allied infantry and 200 allied cavalry to Brundisium by the thirteenth of February.
Taking the premise that the 2nd legion also had 4,000 infantry, Livy's infantry figures would amount to 16,000 men, and when corrected, consist of 8,000 Romans (2 legions) and 8,000 allies (2 legions). Following the ancient writers practice of allocating each legion 5,000 infantry, 4,000 infantry are missing, but this can be explained.
Again in 172 BC, for the coming war against Macedonia, Livy writes that the praetor: "Gaius Licinius was ordered to refit and repair 50 old quinqueremes stored in the ship-yards at Rome. If Gaius Licinius could not refit all of the 50 old quinqueremes, then the praetor Gaius Memmius in Sicily was to make up the difference and sent them to Brundisium. The praetor Gaius Licinius was also ordered to enrol, from the Roman freedman class, sailors for 25 (Roman) quinqueremes. The praetor Gnaeus Sicinius was also ordered to raise 25 allied quinqueremes. The propraetor Aulus Atilius Serranus was chosen to receive the soldiers at Brundisium by the thirteenth of February and then send them to Macedonia. With this fleet and army, the praetor Gnaeus Sicinius was instructed to hold the theatre of operations in Macedonia at Appollonia until his successor arrived to replace him. From the ship-yards (at Rome), Livy has 38 quinqueremes launched under the command of Lucius Porcius Licinius. Another 12 quinqueremes were sent from Sicily.
Now the naval section is a real pig's breakfast. What is actually going on here is that there is a praetorian army of two legions (10,000 infantry) at Appollonia in Greece, under the command of the praetor Gnaeus Sicinius, who is waiting to be replaced by the consul Publius Licinius Crassus. With the arrival of the consul with his 16,000 infantry; there would be now 26,000 infantry under the command of the consul. This means that 6,000 infantry have to return to Italy, thereby reducing the praetorian army to 4,000 infantry, and when combined with the consul's 16,000 infantry, will produce a consular army of 20,000 infantry organised into four legions each of 5,000 infantry. So far, so good.
For 171 BC, Livy writes that Quintus Marcius Philippus (Livy provides no ranking), returned to Greece with a fleet of 50 quinqueremes. Following Polybius that a quinquereme was able to convey 120 marines, 50 quinqueremes x 120 marines = 6,000 marines. Therefore, we can confirm that 6,000 infantry from Gnaeus Sicinius' praetorian army returned to Italy, leaving 4,000 infantry in Greece. A little problem now arises. Livy has Quintus Marcius Philippus as a consul in 186 BC, and consul for Macedonia in 169 BC. In 171 BC, the praetor Gaius Lucretius Gallus was given command of the fleet for Macedonia, and sailed from Rome to Dyrrhachium. This would indicate that Quintus Marcius Philippus name has been mistakenly substituted for the propraetor Gnaeus Sicinius, who after being replaced by the new praetor Gaius Lucretius Gallus, is returning to Italy with the 6,000 infantry.
For 171 BC. Livy writes that both consuls levied four new Roman legions, two Roman legions for each consul, with each legion having of 5,200 infantry and 300 cavalry. The two Roman legions levied for Macedonia had a special provision and were raised from 5,200 infantry to 6,000 infantry. The allied contingent for Macedonia was also raised, with the consul Publius Licinius Crassus, transporting to Macedonia 16,000 allied infantry, 800 allied cavalry and a further 600 allied cavalry previously commanded by the propraetor Gnaeus Sicinius, who remained in Macedonia until replaced. For Italy, Livy writes that the two Roman legions levied for Italy each amounted to 5,200 men, and were allocated 12,000 allied infantry and 600 allied cavalry. Livy's given figures amount to 53,600 men, consisting of 50,400 infantry (22,400 Roman and 28,000 allied infantry) and 3,200 cavalry (1,200 Roman and 2,000 allied infantry):
Italy 10400 Roman infantry 600 Roman cavalry
12000 allied infantry 600 allied cavalry
Total 22400 infantry 1200 cavalry
Macedonia 12000 Roman infantry 600 Roman cavalry
16000 allied infantry 800 allied cavalry
Total 50400 infantry 600 allied cavalry
3200 cavalry
Legions of 5,200 infantry means Livy's source is definitely Polybius, who is the originator the 5,200-infantry legion, which is another of his mistakes. Macedonian legions increased to 6,000 infantry is also incorrect. infantry. The four new Roman legions raised is also correct. The cavalry has 500 cavalry being double counted so needs removing. In 170 BC, the legate, Appius Claudius Centho, eager to erase his earlier defeat in Illyria, with a force of 6,000 men, which included a contingent of Chaonians and Thesprotians, attacked a Macedonian stronghold in Epirus. When no name is given to a place, you can bet it is a fabricated event. The 6,000 represents a legion of 5,000 infantry and the 1,000 garrison troops have been given new nationalities (500 Chaonians and 500 Thesprotians). The garrison troops are also organised into their own fleet, with their own commanders.
Following Livy's given figures (not rounded), after changing the 10,400 Roman infantry to 10,000 Roman infantry, Livy's 50,400 is reduced to 50,000 infantry. Next by removing the 20,000 infantry (16,000 sent to Macedonia and 4,000 still in Greece), this leaves 30,000 infantry. The 16,000 allied infantry allocated to Macedonia, does not belong here and must be reduced, thereby now leaving a residue of 14,000 infantry, which consists of the 8,000 infantry replacements for the consul in Italy, which leaves a residue of 4,000-garrison infantry for Greece, which allocates each of the four legions each of 5,000 infantry, a further 1,000 garrison infantry, giving a splendid total of 6,000 infantry, which Polybius has confused as being a legion of 6,000 infantry. Therefore, 20,000 infantry were sent to Macedonia, and 4,000 infantry remained in Greece.
Now a more serious problem arises, and it has to do with those 50 old quinqueremes stored in the ship-yards at Rome, the 25 allied quinqueremes being raised, the 38 quinqueremes launched and the 12 quinqueremes sent from Sicily near the end of 172 BC. The number of quinqueremes that remained in Greece, amount to 38 quinqueremes. Roman doctrine has 38 quinqueremes return to Italy, not 50 quinqueremes. In 201 BC, after returning from Africa, Livy (31 3) writes the fleet of the propraetor Gnaeus Octavius was handed over to the propraetor M. Valerius Laevinus. The propraetor, M. Valerius Laevinus then selected 38 ships from Gnaeus Octavius' praetorian fleet and with these ships sailed to Macedonia. On his arrival in Macedonia, M. Valerius Laevinus met M. Aurelius, who provided details about the strength of the Macedonian army and navy. Later Livy mentions the information provided by M. Valerius Laevinus and M. Aurelius was read in the senate.
The standard practice for a praetorian fleet is 38 quinqueremes being discharged, and 50 remain, which indicates someone possibly has gotten all their ship data upside down. I've now decided to include a version following the standard practice.
For the military levy of 168 BC, Livy has 5,000 seamen conscripted for the fleet. Since the First Punic War I have found the rowers follow the same number of campaigns they must undertake, which means they have the same replacement system. Livy's 5,000 seamen have been rounded from 5,040 rowers, which then can be used to determine the size of the fleet. However, the mistake made is Livy's source has arrived at 5,040 rowers by giving all the horse transport rowers, and when correct, the number of replacement rowers should be just over 4,000 rowers. This mistake of allocating rowers to the horse transports is a common problem throughout the sources.
In 168 BC, Livy (44 26) also has Perseus army reinforced with 1,000 Gallic infantry and 1,000 Gallic cavalry. Livy claims the Gallic infantry would run alongside the Gallic horses, and when the Gallic rider fell, a Gallic infantryman would seize the riderless horse and ride it into battle. Plutarch mentions that Perseus had 10,000 Basternae horsemen and 10,000 infantry to ride at their side. Plutarch's figures, which are 10 times that of Livy, highlights that Plutarch has taken this information from a source that has provided the number of infantry centuries and cavalry centuries, and being Greek, and not understanding how the Roman military system works, has taken a century as having 100 men.
In conclusion, the military data for the whole of the Third Macedonian War, like the rest of Roman military history, has been constructed on the Roman army and fleet. There size of the Macedonian army remains unknown. Those figures of Perseus running around with an army of 10,000 infantry, of which 4,000 were phalangites is just nothing more than a Roman praetorian army.