Polybius' Levy of Four Legions

[Steven James]

In book six of his Histories, Polybius describes the Roman military system relating to the levy of four Roman legions. Polybius writes that the military tribunes:
 
“Selected four males of more or less the same age and physique. When these are brought forward the officers of the first legion have first choice, those of the second choice, those of the third, third, and those of the fourth last. Another batch of four is now brought forward, and this time the officers of the second legion have first choice and so on, those of the first choosing last. A third batch having been brought forward the tribunes of the third legion choose first, and those of the second last. By thus continuing to give each legion first choice in turn, each gets men of the same standard.”
 
Following Varro’s that the smallest unit in the Roman army was the ‘manus’ (5 men), Polybius’ batches of four men is a red flag. And most importantly, batches of four men does not divide by the number of campaign divisions the men in a legion are undertaking for that year.
 
Firstly, to arrive at batches of four men, this indicates that Polybius has done his own levy calculations.
 
Secondly, Polybius has based his calculations on the correct size of a Roman legion and not his mistaken 4,200 infantry per legion. The legions Polybius has based his calculations are based on the military doctrine of a legion before 211 BC.
 
Thirdly, Polybius has used the correct number of tribes for the levying of four legions.
 
The mistake made by Polybius is that he has divided the number of men that were already levied in a tribe by the maximum number of junior centuries in a tribe. At this point, Polybius has taken the number of men based on the correct number of centuries in a tribe being levied, and then divided this number of men by the maximum number of centuries in a tribe. The result is the number of men already levied is drastically reduced.
 
So, by following Polybius, I get batches of four men, which when distributed means each legion selected one man from each batch of four. Polybius’ batch of four men would be extremely time consuming, especially in light of Livy’s comment that in 446 BC, the levy was completed by 10 o’clock in the morning. Polybius’ levy mistake would involve the 2,400 triarii of the four legions, when divided into batches of four triarii, the levy process would be repeated 600 times (150 times per legion).
 
The levying of four legions involves a proportion of the 35 tribes being selected and a small proportion of the centuries in each tribe being levied. Long gone were the days when the Romans had to levy every tribe and every century in a tribe.
 
I consider Polybius’ mistake to be God sent. It confirms once again I have the correct number of centuries of juniors in a tribe, and the correct number of tribes used for the levy, which is the standard practice the Romans adhere to for all their levies, which occurred after a particular date.
 
The question this leaves is, if Polybius knew the correct size of a Roman legion, why did he claim a legion had 4,200 infantry? This has convinced me that the machinations of the various Roman legion’s organisations and how they are employed had him baffled.

[Sean Manning]

Hi Steven,

I am not an expert in Polybius, but it looks to me like he is describing how men are assigned to a legion one man at a time, not one group at a time. One man can fit into any size of unit.

Its surprising that Polybius' Romans don't have a more sophisticated way of spreading the burden than "call everyone eligible to Rome where 'they' pick recruits from the same tribe four at a time." That is like Athens during the Peloponnesian War or England under Elizabeth I. But its always possible that he misunderstood something or his source was outdated (or that the Loeb translation is misleading).

Sean

[Steven James]

Hello Sean,
 
You wrote:
I am not an expert in Polybius, but it looks to me like he is describing how men are assigned to a legion one man at a time, not one group at a time. One man can fit into any size of unit…. Its surprising that Polybius' Romans don't have a more sophisticated way of spreading the burden than "call everyone eligible to Rome where 'they' pick recruits from the same tribe four at a time."
 
Polybius total of four men from each tribe means his legion of 4,200 men divided by four men per tribe, requires 1,050 tribes from a maximum of 35 tribes. Something is amiss.

[Sean Manning]

Dear Steven, it looks to me like Polybius says that one of those four men is assigned to each of the four legions. Recruiting 120 groups of 4 men from each of 35 tribes would levy 16,800 men, divided into four legions of 4,200 men each (120 men per tribe × 35 tribes).

[Steven James]

Hello Sean’
 
I’m following up with a lot of questions for you. This is not done to put you on the spot or prove wrong from right, I just want to provide food for thought.
 
You wrote:
it looks to me like Polybius says that one of those four men is assigned to each of the four legions. Recruiting 120 groups of 4 men from each of 35 tribes would levy 16,800 men, divided into four legions of 4,200 men each (120 men per tribe × 35 tribes).
 
Yes, but is that how it was actually done?
 
Dionysius claims there are six property classes, Livy five property classes. Your 480 men per tribe divided by Livy’s five property classes equals 96 men per property class. 480 divided by Dionysius’ six property classes equals 80 men per property class. Who is right?
 
How many centuries does each of the 35 tribes allocate to the levy?
 
Polybius states the men selected are of the same age range. Are the 480 men of the same age range? If different age ranges, then what age ranges are there and how many men allocated to each age rage?
 
Are the 480 men sorted into groupings representing their military service? Polybius does mention military service so it must be also be incorporated into the levy system so there is an equal number of raw recruits and veterans. At Cannae, mention is made of the troops being made up of a lot of raw recruits.
 
There is more to the levy than just dividing a set of numbers. Your 16,800 can be divided by 35 tribes, 32 tribes, 30 tribes, 28 tribes, 25 tribes, 24 tribes, 21 tribes, 20 tribes, 16 tribes, 15 tribes, 14 tribes, 12 tribes, 10 tribes, 8 tribes, 6 tribes, 5 tribes, 4 tribes each of 4,200 men, 3 tribes and 2 tribes.
 
Dividing the total number of infantry by the number of tribes will result in some numbers working. It’s a fact of mathematics. With a legion allocated 300 cavalry, four legions gives a total of 1,200 cavalry, which when divided by 35 tribes as per your example, produces 34 point 28 cavalry per tribe. Following Polybius a legion had 200 cavalry, the total of 800 cavalry by 35 tribes equals 22 point 85 cavalry per tribe. Your levy example is in trouble.
 
Returning to your 16,800 infantry and 35 tribes, with a legion having 60 centuries, four legions produce 240 centuries, and when divided by 35 tribes means each tribe allocates to the levy six point 85 centuries. Again, 35 tribes is not working.
 
The levy is done by centuries because a tribe is organised by centuries. Also during a levy, each legion has to have an equal number of centurions and optiones of the same ranking. To this must be added the musicians and artificers.
 
Also which military tribune in a legion, in which Polybius informs us there are six, what does each of the six military tribunes do during the levy. Do they each select 80 men from each property class and if so why?
 
Steven

[Sean Manning]

Steven, its you who claims to have a giant theory for how the Romans conscripted troops (which I don't think you have shared in complete form after talking about it for more than ten years!) I am just paraphrasing a translation of Polybius. If you think he is wrong or does not mean what his language normally means that is OK but you need to argue for that if you want to convince people. When scholars read a theory, they check it against sources, and if it seems to contradict the sources cited (or sources which an expert should be familiar with) and there is no explanation they start to frown. If that happens too often or too blatantly they put the book down and go read something else.

Its usually thought that voting centuries have nothing to do with centuries in the Roman army. Polybius seems to say that preexisting groups within each tribe are broken up when men are conscripted, since men are assigned to each legion one at a time.

As far as I can tell, in this passage Polybius says that "they" (the tribunes?) pick four men from a tribe however they like. He does not mention property classes at all except that the cavalry are "selected according to wealth" and that infantry from some property classes need to bring mail. He does not AFAIK mention any way to prefer men who have served less years. But he does seem pretty clear that four men are chosen from the same tribe and divided among four legions, so there is not as much room to speculate about that.

I am not making any claim about how many tribes Rome had in 220 BCE or 146 BCE. I am giving an example of how a number you gave me can be divided into a legion of 4200 men.

[Steven James]

Hello Sean,
 
You wrote:
Its you who claims to have a giant theory for how the Romans conscripted troops (which I don't think you have shared in complete form after talking about it for more than ten years!) 
 
I am reaping the reward of having revealed the structure of a Roman tribe, as per my academia paper. I simply noticed in the Servian constitution that there were 200 musicians and Dionysius claim that the Roman army had 20,000 men or 200 centuries. After two years of comparing the tribal structure with the army, it was fine tuned, and has found over and over again to work and match the data in the primary sources. With the Roman tribal system exposed, learning how the levy worked was just a matter of time.
 
You wrote:
I am just paraphrasing a translation of Polybius.
 
Sean, you are not under attack. I was pointing out the many other variables that are part of the levy, and which are found in Polybius’ chapter on the Roman army.
 
You wrote:
If you think he is wrong or does not mean what his language normally means that is OK but you need to argue for that if you want to convince people.
 
Actually, many of my posts are updates on progress for those who are interested in what I am doing. We do this via RAT, and hopefully to try and keep RAT alive. I truly have no increase in convincing people of anything. It would like trying to change someone’s religion…pointless.
 
You wrote:
Its usually thought that voting centuries have nothing to do with centuries in the Roman army.
 
No, that is not so. Livy (1 42), when explaining the Servian constitution states: “from it he (Servius Tullius) drew up the classes and centuries and the following distribution of them, adapted for either peace of war.”
 
You wrote:
As far as I can tell, in this passage Polybius says that "they" (the tribunes?) pick four men from a tribe however they like. He does not mention property classes at all except that the cavalry are "selected according to wealth" and that infantry from some property classes need to bring mail. He does not AFAIK mention any way to prefer men who have served less years.
 
Because Polybius does not specifically detail all the information he writes about in his description of the Roman army in book six, does not mean it does not apply to the levy. And Polybius is not our only source of information. The Servian constitution allocates property class a fiscal value and the weapons associated with each property class. This cannot be ignored.
 
You wrote:
But he does seem pretty clear that four men are chosen from the same tribe and divided among four legions, so there is not as much room to speculate about that.
 
Why isn’t there much room to speculate? Polybius is not the only source we have to examine.
 
You wrote:
I am not making any claim about how many tribes Rome had in 220 BCE or 146 BCE. I am giving an example of how a number you gave me can be divided into a legion of 4200 men.
 
Yes, I understand that completely, but I also add other variables that must be taken into account and added in the levy of the 300 cavalry to show a problem has arisen with the levying of 35 tribes. You pointed out that 16,800 infantry were divisible by 35 tribes and I pointed out that 1,200 cavalry for the four legions and 16,800 infantry was not divisible by 35 tribes. Therefore, it is a dead end. Sean, you are not under attack.
 
From what I have leant from the levy, which unfortunately has taken years of research to understand, is the following:
 
For this example, the princeps from the Fabia tribe will be used, as the triarii have been processed. The number of princeps in the Fabia tribe is divisible by the number of legions being raised. The princeps belong to a particular age range. Livy claims the princeps are in the prime of life, so they have been allocated an age range.
 
Before being presented to the military tribunes, the princeps of the Fabia tribe were organised into groups of princeps representing an equal number of princeps relating to their military experience. We cannot ignore that Polybius in his description of the Roman army mentions years of military service. Also let’s not forget that Livy claims most of the men at Cannae were raw recruits, so how did they know this if they didn’t have a system of selection based on military experience. If they did not group the princeps by years of military experience, then this would violate Polybius’ quote that each legion gets men of the same standard because when the levy started, the firs legion given first selection would grab all the veterans.
 
Returning to the levy, as Polybius tells us a legion had six military tribunes, for the levy, I have found the following:
 
1 military tribune from Class I selected the triarii
2 military tribunes from Class I selected the princeps
3 military tribunes from Classes II to IV selected the hastati
 
The first legion selected the same number of princeps from each batch of military experience. The second legion then did the same. Next the third legion made their selection, leaving the fourth legion with the residue. Last legion in each selection process gets the leftover.
 
Next the cavalry from the Fabia tribe are presented. The three highest ranking military tribunes of property Class I select the cavalry.
 
Next the hastati of Class II from the Fabia tribe were presented and selected in the same manner as the princeps.
 
Next the hastati of Class III from the Fabia tribe were presented and selected in the same manner as the Class II hastati.
 
Next the hastati of Class IV from the Fabia tribe were presented and selected in the same manner as the Class III hastati.
 
Next the velites from the Fabia tribe are presented, with the first legion having first selection. All six military tribunes in the first legion selected the velites.
 
Next the musicians from the Fabia tribe were presented. All six military tribunes selected the their maximum number of musicians.
 
Next the artificers from the Fabia tribe were presented. All six military tribunes selected their maximum number of velites.
 
I have excluded the levying of the centurions and optiones. Polybius claims that “from each of the classes except the youngest, they elect 10 centurions according to merit. Polybius’ reference to 10 centurions per class I have found to be accurate, and in relation to merit, the numbers tell me this has been based on military experience, and each property class ends up with an equal distribution first time serving centurions to veteran centurions. Polybius’ 10 centurions per property class can still be found in Caesar’s description of the promotion of a centurion: the Primi ordinum, Superiores ordines, Inferiores ordines and Infirmi ordines.
 
After the artificers from the Fabia tribe were processed, the next tribe determined by lot was presented to the military tribunes. This meant the second legion now had first selection, the third legion had second selection, the fourth legion had third selection, leaving the fourth legion with the leftovers.
 
The number of tribes presented and the number of centuries each tribe contributed to the levy had been calculated that every legion had an equal number of first selections, second selections, third selections, and the fourth selection gets the residue.