Late Roman Legion size based on the Perge Inscription

[FlaviusB]

I know this subject has been covered before and I'm wading into 'well churned' waters, but I've recently begun reading up on the Roman Army and Late Antiquity and wanted to see if anyone had thoughts on the following reconstruction of the Perge legion. I'll post links to the previous RAT thread on this subject and to Dr. Fatih Onur's publication on the translated inscription, for anyone reading this that hasn't encountered the subject before:

https://www.romanarmytalk.com/rat/showth...?tid=19361
https://www.academia.edu/49012668/The_An...nd_Edition

My reconstruction is based on the assumption that this legion, and possibly Late Roman Legions in general from Constantine onwards, were a third the size of the Principate Legions. This is based on the fact that the latter Legions had 6 Tribunes serving under the Legatus alongside 60 centuries commanded by a Centurion (I'm assuming 80 men to a Century, leaving aside the so-called milliary cohorts). The Perge inscription lists 2 Tribunes (Tribunus Numeri and Minor) and 20 Ordinarii or Centurions by comparison, which suggests that these later Legions had numbered 1600 men excluding supernumeraries. 

Using 1600 as a baseline, I was stuck on how to divide up the troops listed on the inscription and especially the 136 Torquati Semissales and 256 Bracchiati Semissales. In RAT thread posted above, Julian de Vries pointed out that 256 is 16 x 16, which can be split into 32 x 8 just as 136 is divisible by 8 into 17 x 8; I split the Torquati and Bracchiati into 8-man Contubernia on this basis, assuming an 80-man Century comprised of 10 Contubernia. The same can be done with the various grades of Augustales and Flaviales, whose combined totals of 120 and 200 respectively can be divided into groups of 6 and 10 for each century, ensuring each century has 2 Contubernia consisting of 3 Augustales and 5 Flaviales. 

The previous groups plus the 59 Munifices listed partially on the inscription add up to 771, which subtracted from 1600 gives us a remainder of 829. The number we're looking for is the first of the three digit amount of Munifices ending in 59, and this can be found by including the 20 Armaturae Semissales in the ranks of the centuries. I accounted for the remaining 9 by adding the Librarii, Mensores and Beneficarii from among the 'double annonae men' to the centuries, my reasoning being that the musicians, standard bearers and praeco would be necessary for signalling and communicating in battle, whereas the other nine seem to fulfill non-combat administrative roles. We thus arrive at 800 Munifices or 859 including those already accounted for.

The resulting make-up of the Centuries is as follows:

Centuries 1-9: 16 Bracchiati, 8 Torquati, 6 Augustales, 10 Flaviales, 1 Armarturae Semissales, 1 of the Librarii, Mensores or Beneficarii, and 38 Munifices.

Centuries 10-12 have the same structure as 1-9, but with a 39th Munifice instead of one of the 'double annonae men.'

Centuries 13-17 only have 8 Bracchiati Semissales, and thus have 47 Munifices instead of 39.

Finally, Centuries 18-20 have no Torquati and have 55 Munifices instead of 47. 

800 Munifices plus the 1245 officers and men listed on the inscription adds up to 2045; tripling this number gets us to 6135, which is slightly higher than the 6100 figure given for the 'ancient legion' by Vegetius in De Re Militari, excluding the 726 cavalry. As for the rest of the Legion's structure, I split the 275 Veredari into 5 Turmae of 55 with 10 Veredari and 45 Veredari Alii each, while each 'maniple' of two centuries has an Optio, Vexillarius, Imaginifer and Signifer. 

This leaves a Legion staff of 2 Tribunes, 8 Cornicines, 4 Tubicines, 2 Buccinatores, 1 Praeco and 73 Clerici et Deputati, the latter of which can be distributed in a way that makes sense with this structure; by assigning 1 to each of the 5 Veredari Turmae, and distributing the rest among the Legion. Five less than 73 is 68, assigning two to each Ordinarius reduces this to 28 and assigning 2 more to each Optio reduces the number to 8, of which 4 could then be assigned to each Tribune. This gives us 1765 officers and men for a legion without cavalry, which can be tripled to 5295 and isn't far off from the 5550, the latter being the size Vegetius attributes to non-milliary centuries multiplied by ten (excluding cavalry).

[Nathan Ross]

800 Munifices plus the 1245 officers and men listed on the inscription adds up to 2045; tripling this number gets us to 6135... I split the 275 Veredari into 5 Turmae of 55 with 10 Veredari and 45 Veredari Alii each, while each 'maniple' of two centuries has an Optio, Vexillarius, Imaginifer and Signifer. 

73 Clerici et Deputati, the latter of which can be distributed in a way that makes sense with this structure...

Well done for engaging with this question, and for making the numbers work (as far as I can see, anyhow!) - it does become a little maddening after a while... I'm impressed that you figured in the clerici et deputati too; I confess I'd discounted them, as they're supernumeraries with no annonae, and besides I'm not sure if the number 73 is accurate (the inscription itself is so fragmentary at that point I cannot make out what Prof Onur was drawing on!)

Getting the 'veredarii' cavalry (if that's what they were) into a balanced structure is one of the main difficulties with the Perge list. Dividing them into five turmae might work. The problem would be that we do not have five decurion-type officers to command them, or five standard bearers either. And 55 is a very large turma.

All the officer, nco and standard-bearer ranks on the list, meanwhile, are in 10s. That suggests to me that the complete organisation was decimal. However, as you point out, many of the internal numbers break down by 8s and 16s.

My own theory, outlined on the years-old thread you linked, would be a paired 'numerus' of ten infantry and ten cavalry subunits, each commanded by an ordinarius. It's not a perfect system by any means. But I still like it for several reasons. Firstly as both the infantry and cavalry complement appear close to the estimates for late Roman units of their type (c.1200 and c.300 respectively). Secondly because we can see the core of the old principiate 80-man century still preserved in the enlarged 110-man 'mini cohort'. Thirdly because Vegetius says that a century (in his day, perhaps?) was 110 men - an otherwise-inexplicable comment!

So, while I still don't think after all this time that this is the only possible solution to the Perge enigma, it does at least seem a close-ish fit to the limited evidence we have from elsewhere:

   

   

[FlaviusB]

800 Munifices plus the 1245 officers and men listed on the inscription adds up to 2045; tripling this number gets us to 6135... I split the 275 Veredari into 5 Turmae of 55 with 10 Veredari and 45 Veredari Alii each, while each 'maniple' of two centuries has an Optio, Vexillarius, Imaginifer and Signifer. 

73 Clerici et Deputati, the latter of which can be distributed in a way that makes sense with this structure...

Well done for engaging with this question, and for making the numbers work (as far as I can see, anyhow!) - it does become a little maddening after a while... I'm impressed that you figured in the clerici et deputati too; I confess I'd discounted them, as they're supernumeraries with no annonae, and besides I'm not sure if the number 73 is accurate (the inscription itself is so fragmentary at that point I cannot make out what Prof Onur was drawing on!)

Getting the 'veredarii' cavalry (if that's what they were) into a balanced structure is one of the main difficulties with the Perge list. Dividing them into five turmae might work. The problem would be that we do not have five decurion-type officers to command them, or five standard bearers either. And 55 is a very large turma.

All the officer, nco and standard-bearer ranks on the list, meanwhile, are in 10s. That suggests to me that the complete organisation was decimal. However, as you point out, many of the internal numbers break down by 8s and 16s.

My own theory, outlined on the years-old thread you linked, would be a paired 'numerus' of ten infantry and ten cavalry subunits, each commanded by an ordinarius. It's not a perfect system by any means. But I still like it for several reasons. Firstly as both the infantry and cavalry complement appear close to the estimates for late Roman units of their type (c.1200 and c.300 respectively). Secondly because we can see the core of the old principiate 80-man century still preserved in the enlarged 110-man 'mini cohort'. Thirdly because Vegetius says that a century (in his day, perhaps?) was 110 men - an otherwise-inexplicable comment!

So, while I still don't think after all this time that this is the only possible solution to the Perge enigma, it does at least seem a close-ish fit to the limited evidence we have from elsewhere:

My thinking, which you may have read already in the older thread, is that the 45 Veredarii Alii in each Turmae would be divided into 5 files of 9, while the 10 Veredarii could be divided into 5 'file leaders' giving each file 10 men in total and the other 5 are a headquarters. I should also note that 6 of these 55-man Turmae plus a larger headquarters staff would get us close to the 360 number mentioned by Julian in his letter to the Athenians, and the 350 figure implied by Ammianus for Late Roman Cavalry regiments.

[Nathan Ross]

the 10 Veredarii could be divided into 5 'file leaders' giving each file 10 men in total and the other 5 are a headquarters.

Yes, that might work if we assume that these turmae are integral to the command of the two-century 'maniples', each under two ordinarii. But it seems an overly complex structure. 

Usually we would expect each subunit to have its own dedicated commander, and that commander to be dignified by a particular rank title and paid more in annonae than the men he was leading. Do you see each individual file operating as its own subunit, perhaps, under the Veredarii?

[FlaviusB]

the 10 Veredarii could be divided into 5 'file leaders' giving each file 10 men in total and the other 5 are a headquarters.

Yes, that might work if we assume that these turmae are integral to the command of the two-century 'maniples', each under two ordinarii. But it seems an overly complex structure. 

Usually we would expect each subunit to have its own dedicated commander, and that commander to be dignified by a particular rank title and paid more in annonae than the men he was leading. Do you see each individual file operating as its own subunit, perhaps, under the Veredarii?

That would be my guess since the 275 Veredarii appear to be the Legion's cavalry; being able to operate as turmae as well as in their files would allow some flexibility in creating detachments for scouting, screening and delivering messages.

[FlaviusB]

800 Munifices plus the 1245 officers and men listed on the inscription adds up to 2045; tripling this number gets us to 6135... I split the 275 Veredari into 5 Turmae of 55 with 10 Veredari and 45 Veredari Alii each, while each 'maniple' of two centuries has an Optio, Vexillarius, Imaginifer and Signifer. 

73 Clerici et Deputati, the latter of which can be distributed in a way that makes sense with this structure...

Well done for engaging with this question, and for making the numbers work (as far as I can see, anyhow!) - it does become a little maddening after a while... I'm impressed that you figured in the clerici et deputati too; I confess I'd discounted them, as they're supernumeraries with no annonae, and besides I'm not sure if the number 73 is accurate (the inscription itself is so fragmentary at that point I cannot make out what Prof Onur was drawing on!)

Getting the 'veredarii' cavalry (if that's what they were) into a balanced structure is one of the main difficulties with the Perge list. Dividing them into five turmae might work. The problem would be that we do not have five decurion-type officers to command them, or five standard bearers either. And 55 is a very large turma.

All the officer, nco and standard-bearer ranks on the list, meanwhile, are in 10s. That suggests to me that the complete organisation was decimal. However, as you point out, many of the internal numbers break down by 8s and 16s.

My own theory, outlined on the years-old thread you linked, would be a paired 'numerus' of ten infantry and ten cavalry subunits, each commanded by an ordinarius. It's not a perfect system by any means. But I still like it for several reasons. Firstly as both the infantry and cavalry complement appear close to the estimates for late Roman units of their type (c.1200 and c.300 respectively). Secondly because we can see the core of the old principiate 80-man century still preserved in the enlarged 110-man 'mini cohort'. Thirdly because Vegetius says that a century (in his day, perhaps?) was 110 men - an otherwise-inexplicable comment!

So, while I still don't think after all this time that this is the only possible solution to the Perge enigma, it does at least seem a close-ish fit to the limited evidence we have from elsewhere:

I should also add that a 1600-man legion could make sense of a lot of the figures that Marcellinus cites in his Res Gestae. He mentions units of 300 four times (XVIII, 2, 11; XX, 4, 2; XXXI, 11, 2; XXXI, 15, 4), two times each for units of 500, 800 and 1000 (500: XXV, 7, 3; XXXI, 11, 2; 800: XVII, 1, 4; XXIV, 6, 4; 1000: XXIV, 1, 6; XXX, 1, 11). Four centuries/Two maniples of 320 men could be rounded down to 300, while 6 centuries/3 maniples of 480 men could be rounded up to 500. 800 is the easiest since 10 centuries or 5 maniples gets you exactly 800 men, while 12 centuries/6 maniples of 960 could be rounded up to 1000.

[Nathan Ross]

the 275 Veredarii appear to be the Legion's cavalry; being able to operate as turmae as well as in their files would allow some flexibility in creating detachments for scouting, screening and delivering messages.

I suppose we could have the veredarii integrated into the centuries, as the old equites legionis seem to have been during the principiate. But again we have the problem that the full number does not divide by the number of centuries, which the other estimates from earlier history do - Josephus's 120 legion cavalry give us 2 horsemen per century in the legion. Vegetius (II.6) gives 66 horsemen per cohort, and 132 in the first cohort. But 275 Veredarii cannot be divided by either ten (my version) or twenty (your version) centuries.

Adding an extra five men (3 mensores, 2 bucinators?) to the number of the veredarii alii gives us 230 of them, which together with the 50 veredarii allows a far neater spit into ten groups of 28 men, plus two or three 'officers' each giving us ten turmae of 30 or 31 each.

Milner's note on p.36 has Vegetius 'artificially conceive' the size of the cavalry component by adding two 32-man turmae and two decurions. So the original turma would presumably be 33 men, including decurion. Whether this has any grounding in fact is another matter!

Another (perhaps more 'Vegetian'?) way of doing it might be to divide the 225 veredarii alii into nine turmae of 25 men each, then to have the veredarii all in one 'double strength' turma of 50 men. Again, that would be for a 10-way split.

EDIT - looking more into the mess of the Vegetian 'antique legion' system, it appears that he is dividing each of his cohorts into 5 centuries - giving 111 men per century in a normal cohort and 221 in the first cohort; both figures include the centurion. And that is presumably the origin of his idea that a century contains 110 men. However, his cavalry components do not match up: neither 66 troopers for the normal cohort or 132 for the first cohort are divisible by the number of centuries!

So there might be some support there for a cavalry component that does not divide equally between the number of centuries. But I'm inclined to think it is just shoddy calculation by Vegetius (apparently some manuscripts have tried to 'correct' his figures?). Where that leaves my 110-man 'century' I don't know...

[Steven James]

Nathan wrote:
But again we have the problem that the full number does not divide by the number of centuries, which the other estimates from earlier history do - Josephus's 120 legion cavalry give us 2 horsemen per century in the legion. Vegetius (II.6) gives 66 horsemen per cohort, and 132 in the first cohort. But 275 Veredarii cannot be divided by either ten (my version) or twenty (your version) centuries.
 
First mistake is not to count the number of officers for the cavalry when dividing by the number of centuries in a legion.
 
Nathan wrote:
Milner's note on p.36 has Vegetius 'artificially conceive' the size of the cavalry component by adding two 32-man turmae and two decurions. So the original turma would presumably be 33 men, including decurion. Whether this has any grounding in fact is another matter!
 
A 33-man cavalry squadron works for me.
 
Nathan wrote:
EDIT - looking more into the mess of the Vegetian 'antique legion' system, it appears that he is dividing each of his cohorts into 5 centuries - giving 111 men per century in a normal cohort and 221 in the first cohort; both figures include the centurion.
 
I love Vegetius. His 6,000-infantry legion is historical. Academia has failed to understand the many military doctrines of the Roman army. Vegetius’ mistakes have come about due to Vegetius getting confused about the billeting arrangement of the legion, when in camp. And this is quite easy to do. The fact that no academic has recognised these military doctrines for the last 400 years proves this. And to prove my point, I will share some of my research, which will most likely be plagiarised in one form or another.
 
When in camp, the 10 cohorts of 600 infantry in a 6,000-infantry legion are rearranged into 12 cohorts each of 500 infantry. Isidore states the 6,000-man legion has 60 centuries, 30 maniples and 12 cohorts. Oh, cried academia, Isidore does not know what he is talking about, a legion does not have 12 cohorts.
 
Oh yes it does! When in camp, cohort 11 is distributed amongst the 10 cohorts, thereby allocating each cohort 50 infantry, bringing the total of cohorts 1to 10 to 550 infantry per cohort, organised into five centuries each of 110 infantry. Interestingly, in his account of the siege of Hatra in 199 AD, Cassius Dio mentions a body of 550 European men. That would make a legion of 5,500 infantry.
 
Returning to Vegetius, cohort 12 of 500 infantry, when in camp is allocated to the first cohort, thereby increasing cohort one to 1,050 infantry. So, to arrive at 1,105 men in cohort one, Vegetius has incorrectly added the 500 infantry of cohort 12 to the original 600 infantry of cohort one in the 10-cohort system, giving Vegetius a total of 1,100 infantry and five officers (centurions). Here Vegetius has also made a mistake about the distribution of the centurions.
 
Nathan wrote:
However, his cavalry components do not match up: neither 66 troopers for the normal cohort or 132 for the first cohort are divisible by the number of centuries!
 
To arrive at the figure of 132 cavalry for the first cohort, Vegetius has erroneously allocated an additional 66 cavalrymen to the 12th cohort, thereby increasing the first cohort from 66 cavalry to 132 cavalry. This means the 660-cavalry increased to 726 cavalry. In actuality, the number of cavalry with officers was 660 men (600 cavalrymen and 60 decurions), organised into 20 squadrons each of 33 men (30 cavalrymen and three decurions).
 
Lydus also mentions that a 6,000-man legion also had the same number of horsemen (6,000 cavalry). Isidore claims a 6,000-man legion had 200 squadrons of cavalry. Well, you should be able to work out how Isidore arrived at 200 cavalry squadrons, and it has nothing to do with a copyist error, it is a mathematical mistake.

[Steven James]

And the world held its breath waiting for the Perge document to finally reveal all about the Late Roman army. Well, I guess that did not eventuate. Army rosters are useless if we do not have the exact organisation of a Roman legion or any unit for that matter to compare with.
 
And yet, a lot of the data we do have on the Late Roman army is brushed aside by academia as being either unreliable, exaggerated, or, insert the numerous reasons why. Below is some of the data that has been dismissed by the higher and better informed:
 
The Passion of Saint George reports that Diocletian allocated the general Euchios, command of 3,000 soldiers to overthrow the Christian churches in Egypt. The Passion of Saint George also reports that Anastasius, the governor of Palestine being allocated 3,000 soldiers by the emperor Diocletian. Zosimus has the emperor Julian sail to Sirmium with 3,000 troops.
 
The figure of 3,000 soldiers is a rounded number, but 3,000 is where one should start, as it includes juniors and seniores, both infantry and cavalry, plus officers and supernumeraries. It is the whole package. Unfortunately, I have been told on many occasion that anything written in the Ante Nicene Fathers has been dismissed.
 
Well then, try and explain this.
 
In the Roman Martyrology, Saint Meletius with 252 of his soldiers was put to death for being avowed Christians. The historian Macarius, mentions that 1,104 soldiers were stationed at Melitene.
 
From “The Acts of the Disputation of Archelaus.” Archelaus was offered for a sum of money, a large number of Christian prisoners (men, women and children), by the soldiers of the camp. The tally of Christians consisted of ‘some’ 7,700 prisoners, of which 500 were wounded, and ‘about’ 1,300 Christians were killed. This gives a grand total of ‘about’ 9,000 Christians. The Second Book of Chronicles also uses the number 7,700 for the size of a flock of rams and goats. It could be that the original author recognised the similarity to the biblical number of 7,700 rams and goats, and therefore, partition those taken prisoners and wounded to closely recreate that figure.
 
The 9,000 Christians divided by Macarius’ 1,104 soldiers equals 8 point 15, so leave at 8 units x 1,104 soldiers = 8,832 soldiers. Now if I doubled the eight units to 16 units and then divided the 8,832 soldiers by 16 units, this produces 16 units each of 552 soldiers. But there’s more! At this stage, if I double the 552 soldiers this makes 1,104 soldiers, which brings us right back to Macarius’ figure of 1,104 soldiers stationed at Melitene. Can someone tell me where the maths is wrong? But there’s more. What if I take the figure of 552 and minus the 252 soldiers that died with Saint Meletius, I get a residue of 300 soldiers, so minus 252 again, and this leaves 48 soldiers. Well, bless my soul, I’m not scratching my head to the identity of those 48 soldiers but I guess you are.
 
The Passion of Saint George mentions on the twenty-fifth of the month of Epiphi, 2,408 Christian soldiers were executed by the order of Dadianus the governor of Bithynia. From this, one can extract the legion’s wedge formation.
 
Ammianus mentions the death of 50 soldiers. The Passion of Saint George also mentions two officers each commanding 50 soldiers. Coincidence? Hardly.
 
Ammianus mentions on three occasions that the Late Roman army was organised into cohorts, centuries and maniples. Like Ammianus, for the year 398 AD, Claudian also attests to the existence of maniples during the reign of the emperor Theodosius I (379 AD to 395 AD). When writing about the army of emperor Theodosius I, Pacatus writes that “the plain was bristling with troops: cavalry sent out to the wings, light troops placed in front of the standards, cohorts arranged by maniples, legions arrayed in squares, moving their columns forward at a quick pace, occupied the whole field as far as the eye could see.”
 
Wait up. Claudian claims the Late Roman army was made up of cohort arranged by maniples. Well, this goes against academia’s claim that the cohort legion replaced the maniple legion. So, who is right and who is wrong. Well, Claudian is wrong because, academia can never be wrong, nor can every amateur historian or wargamer.
 
The above references to the Roman organisation show the Late Roman army was organised into cohorts, maniples and centuries, and had been organised this way since 513 BC. Nothing has changed, just the size for the legion. The rule of thumb is centuries make maniples and maniple make cohorts.
 
According to Varro, a cohort consisted of several maniples. Notice how Varro is not specific in exactly how many maniples make a cohort. Throughout the primary sources for the early, mid and late republic, a vexillation or numerus and some other organisations have simply been called a “cohort,’ which accords with Varro’s comment that a cohort had several maniples. Later, during the principate, terms such as vexillation, numerus etc pop up, which shows the precise terminology is coming into play. However, such organisations have been around since the beginning of the republic, but all given the name “cohort.”
 
So, how did I manage to get Vegetius and the 1,104 soldiers to work? Well, first, I never listened to the nay sayers on this forum, and I have never taken on board this ridiculous concept of which ancient historians is more reliable than other historians. Numbers are numbers to me, and all are invaluable. This favouritism methodology amongst the ancient historian is a stupid concept, that as time has proven, has stifled any advancement in our study of the Roman army since Lipsus published the first book on the Roman army in 1596. The same stupid arguments found in Lipsus are still being debated. Professor Ridley once told me to never throw out evidence, and yet, this is exactly what academia does, it continuously throws the baby out with the bathwater.
 
I am still searching for that extensive study that proves that according to academics, Polybius is the most reliable historian. Why does academia believe it can make claims and not back them up?

[FlaviusB]

Milner's note on p.36 has Vegetius 'artificially conceive' the size of the cavalry component by adding two 32-man turmae and two decurions. So the original turma would presumably be 33 men, including decurion. Whether this has any grounding in fact is another matter!

Another (perhaps more 'Vegetian'?) way of doing it might be to divide the 225 veredarii alii into nine turmae of 25 men each, then to have the veredarii all in one 'double strength' turma of 50 men. Again, that would be for a 10-way split.

EDIT - looking more into the mess of the Vegetian 'antique legion' system, it appears that he is dividing each of his cohorts into 5 centuries - giving 111 men per century in a normal cohort and 221 in the first cohort; both figures include the centurion. And that is presumably the origin of his idea that a century contains 110 men. However, his cavalry components do not match up: neither 66 troopers for the normal cohort or 132 for the first cohort are divisible by the number of centuries!

So there might be some support there for a cavalry component that does not divide equally between the number of centuries. But I'm inclined to think it is just shoddy calculation by Vegetius (apparently some manuscripts have tried to 'correct' his figures?). Where that leaves my 110-man 'century' I don't know...

Jonathan Roth's 1994 article in Historia, "The Size and Organization of the Imperial Roman Legion," argues that Vegetius was likely playing fast-and-loose with the numbers for the Legion, along with perhaps confusing some early Republic, late Republic and Principate figures, to make his point: that the Legions had been allowed to decline in size and strength and that this trend should be reversed rather than raising new Auxilia regiments and employing more barbarians. 

Roth also argues that the discrepancy between the 4800 and 6000 figures given for the Legion can be resolved by looking at the latter figure as including Calones (ie military slaves). Based on Pseudo-Hyginus' description of the Century containing 80 milites and the Cohort containing 600 homines, alongside Maurice's suggestion of having 1 slave per 3-4 soldiers, he assigns 1 to each Contubernia, 1 to each of the 120 Equites Legionis, and 1 to each Centurion, with an additional 420 to fulfill other duties throughout the legion. 780 plus 420 is 1200, which gets us from 4800 to 6000; it also equates to 2 calones per contubernium, which gives each century 80 milites and 20 calones for 100 men in a century excluding supernumeraries like the Centurion and Optio. Factoring in a double-strength First Cohort results in a legion of 5280 milites and 1320 calones, or 6600 homines in total, which accounts for the 6600 figure mentioned by Lydus. 

As for the Veredarii not evenly dividing among the centuries, the 280-strong regiment I suggested (5 clerici et deputati, 50 Veredarii, 225 Veredarii Alii) can still be divided into 5 56-man Turmae, which is enough for 1 to support each 320-man cohort. Assuming that I'm correct about the Legions being split into thirds under Constantine, possibly along their Manipular lines, it could be that a cavalry unit was added to at least some of the Legions (perhaps those whose Equites had not been used by Gallienus to create new Promoti units) to give them some added striking power and improved communications and intelligence-gathering capabilities to offset their smaller size?

[Steven James]

FlaviusB wrote:
Jonathan Roth's 1994 article "The Size and Organization of the Imperial Roman Legion," argues that Vegetius was likely playing fast-and-loose with the numbers for the Legion.”

I find Jonathan Ross guilty of doing exactly the same, loose with the numbers. Like many academics, anything that jeopardises his theory is dismissed. For example, Suetonius gives a legion at 5,600 men, and because this does not sit well with his theory of a legion having 4,800 infantry and 1,200 slaves, he writes “Suetonius is almost certainly referring to a republican legion consisting of 5,200 infantry and 400 cavalry.”

So, Roth, prove it. Give me a troop breakdown of the 5,200 infantry plus their organisation, and the same goes for the cavalry. 5,200 infantry divided by 60 centuries per legion equals 86 point 66. So how does that work? And exactly how are the 400-cavalry organised? Give us details. Surprising to many, Tacitus, Josephus and other support Suetonius’ 5,600 figure. I am seriously convinced that academics must be the most mathematically challenged people on this planet.

Roth explains that Lydus assigns 6,000 men to a legion and Lydus gives the date for this legion to the early republic. Roth throws in the usual put down that “Lydus’ numbers are not dependable.” Again, Roth makes no effort to try and understand Lydus. Roth tells us that Lydus gives the date of the 6,000 legion to the early republic. Actually, Lydus gives the year as being 388 BC. Now anyone who knows their Roman history, will be aware that in 387 BC, the Romans added four new tribes, bringing the total to 25 tribes. Has anyone study whether there is a connection? Answer…no.

Let me share some of my research. Lydus, writing in the fourth century AD, has interpreted the 60-centuries as having 100 men, so the legion blows out to 6,000 men. Lydus is not the only ancient historian who has confused a century as having 100 men. There are a few more examples. Take Servius for example, he claims a legion had 6,000 infantry and 300 cavalry. Servius has the right cavalry numbers, but like Lydus has multiplied the 60 centuries in a legion by 100 men.

In relation to Hyginus, Roth writes: “If all personnel in a legion were assigned to the ranks of the centuries, regardless of their status, then the total number of soldiers authorized for a standard imperial legion would be exactly 4,800. Accepting this hypothesis, however, leaves unresolved the question of Pseudo-Hyginus' "missing" 1,200 men. They can be found by carefully noting Pseudo-Hyginus' terminology: he gives the number of "soldiers" (milites) in a century as 80, but says that the cohort has 600 "men" (homines). Indeed, there is a military category which belonged to the legion, but would be homines and not milites: non-combatant slaves. It is confusion about the role and number of military slaves which has led to both ancient and modern perplexity over the legion's size.”

And this is the crutch of Roth’s problem. Roth believes the discrepancy between a legion of 4,800 and 6,000 is due to 1,200 slaves being added to the legion. I can see why many have rejected his theory. It is baseless and preposterous. Like the example I earlier provided in the Vegetius posting above, Roth has no idea of how the legion was organised when in camp, and Hyginus is explaining the Roman army when in a camp. It’s the camp layout. Taking Roth’s 4,800 infantry for a legion, which would be organised into 10 cohorts each of 480 infantry, and following the camp doctrine for the principate, two cohorts amounting to 960 infantry are distributed amongst the remaining eight cohorts, so, 960 infantry divided by eight cohorts equals 120 infantry. Therefore, each of the remaining eight cohorts increased from 480 infantry to 600 infantry as per Hyginus. Someone, tell me where the maths is wrong.

Sadly, there is a lot about Hyginus that academics have failed to understand, and I personally like Hyginus, in fact I adore him. He has been the greatest gateway to understand the legion of the princepate.

And what does Roth say about Dio Cassius claim of 550 men. He claims “the number given in the text almost certainly does not refer to a cohort, but rather a vexillation or perhaps a numerus.” Well, besides his dumb arse opinion, does Roth provide any proof…no. He feels there is no need to for him to provide proof of what is the size of a vexillation or even a numerus, so how can he make such a claim.

And what does Roth have to say about Isidore: “Isidore's description of the legion's organization, given in the same definition, is odd - it begins with the accurate statement that a legion had sixty centuries and thirty maniples, but then adds that there were twelve cohorts and 200 (sic) turmae in a legion. The number of turmae is incomprehensible, and shows confusion and possibly corruption, but the misunderstanding of the number of cohorts may be enlightening. Under the heading of cohors, and probably using a different source, Isidore says that a cohort contains five hundred men. It is possible that, in an attempt to reconcile this (imperial) cohort with his (republican) legion of 6,000, Isidore concluded that a legion had twelve cohorts.

As I showed in the Vegetius example, the 6,000-infantry legion did, when in camp have 12 cohorts each of 500 men. But, the sad majority of people cannot accept that for a period of time, until Diocletian’s reform, there was a legion operating of 6,000 infantry and 600 cavalry. And I can tell you, even if the sceptics were given a time machine, or more irrefutable evidence was found, they would still ignore it, because, they just don’t like being wrong, and are more than happy to protect a falsehood.

Roth writes: “A careful re-evaluation of the literary and documentary sources, however, might well solve the issue.” What Roth means is my paper has resolved the issue” on what the size of an imperial Roman legion was. Roth is not alone in putting a theory first, and then dismissing anything that threatens that theory. Another paper that reaches the heights of absurdity is “Tactical Organisation of the Early Imperial Legion” by David Hoyt, which can be found online in pdf format. Much of Hoyt’s paper is a cut and paste of Roth. Hoyt follows the opinion of most academics, and believes the maniple legion was replaced by the cohort legion. This totally ignores evidence of maniples and cohorts going back to the reign of king Tarquinius Superbus (534 BC to 509 BC. To protect his theory, Hoyt writes:

“But what of Metellus' use of both cohorts and maniples in Africa, as described by Sallust? The legions by this time were undoubtedly trained in both systems as expedients for differing circumstances (cohortal vs. cavalry or undisciplined mobs and manipular vs. disciplined opponents); however, Metellus' use of maniples against the Numidian cavalry and infantry ultimately proved unsuccessful. By Marius' time the cohort had become the dominant tactical unit after a long period of transition. By the time of Sulla's campaigns in the east the cohort was the standard. There is no direct evidence to support the theory that Marius introduced it himself.”

Hoyt provides no direct evidence to support his claim. Opinions now reign supreme, and as most of the public only read secondary sources, such opinions gain traction and are taken as fact because, well, everyone is saying the same thing, it must be right.

FlaviusB wrote:
As for the Veredarii not evenly dividing among the centuries, the 280-strong regiment I suggested can still be divided into 5 56-man Turmae, which is enough for 1 to support each 320-man cohort.

You need to take all the data concerning the Late Roman legion to substantiate this, otherwise it is just a mathematical exercise. The Perge document is not going to provide all the answers.

FlaviusB wrote:
Assuming that I'm correct about the Legions being split into thirds under Constantine, possibly along their Manipular lines, it could be that a cavalry unit was added to at least some of the Legions.

I have a mass of data from the primary sources that it was Diocletian’s reforms that created the Late Roman legion, not Constantine. I give Constantine credit for abolishing the 35 tribes of Rome.

FlaviusB wrote:
(perhaps those whose Equites had not been used by Gallienus to create new Promoti units) to give them some added striking power and improved communications and intelligence-gathering capabilities to offset their smaller size?

Perhaps they did this, perhaps they did, maybe they didn’t. Perhaps is not evidence, just more opinions. Roth’s paper is full of obscurities, such as:

almost certainly referring to
Both are doubtless alluding to republican legions.
it may be a mistake or the number might possibly be lifted from Vegetius.
It is highly probable that after Augustus' reform all Roman legions had a standard nominal size
Upon careful analysis, most of these differences result from the misinterpretation of the evidence. (But not you Roth hey).
but rather to a vexillatio or perhaps a numerus
The first is probably not a legion at all
Vegetius might well be using "legion" loosely for barbarian numeri
Vegetius' numbers represent proposals, rather than descriptions
Vegetius might well have extracted accurate details from them, but only when they were useful to his scheme.” Well, isn’t that what you are doing Roth!

Roth hasn’t proven any of his claims. The fact that you employed Roth’s paper to show that a 6,000-man legion is plausible due to additional slaves tells me you believe the Vegetius legion to be fiction. If so, like many, your prejudices are your barriers to further understanding.

So, why is Vegetius’ legion not historical? Will someone care to answer the question?

[Nathan Ross]

Army rosters are useless if we do not have the exact organisation of a Roman legion or any unit for that matter to compare with.

The Perge tablets are not an army roster in the sense of a temporary roll-call, like the Dura reports or the ones from Vindolanda. They show us the official total numbers of a late 5th-century legion (or 'numerus legionem') at full strength. This is the hardest of hard evidence, literally carved into stone!

The trouble is, of course, that the inscription is fragmentary, hard to decipher, and there's a lot that we don't know about it. Is the unit described a normal one, or exceptional? For how long did this organisation persist? (for the numbers to need recording in such detail, we might assume there was something unusual or novel about them...)

This underlines the difficulties of using ancient texts, or ancient evidence of any sort. None of it can be regarded as a clear window into the past, all of it requires interpretation, and none of those interpretations are going to be exact or perfect.

 8 units x 1,104 soldiers = 8,832 soldiers. Now if I doubled the eight units to 16 units and then divided the 8,832 soldiers by 16 units, this produces 16 units each of 552 soldiers. But there’s more! At this stage, if I double the 552 soldiers this makes 1,104 soldiers, which brings us right back to Macarius’ figure of 1,104 soldiers stationed at Melitene. Can someone tell me where the maths is wrong?

You have shown that 1104 can be divided by 8 and 16. But 9000 cannot, and nor can 7700, 1300. You can assume that these numbers are 'rounded' if you like, but if your thesis depends on exact number concordances you need to show that working, rather than rounding numbers up or down to make them fit.

Pacatus writes that “the plain was bristling with troops: cavalry sent out to the wings, light troops placed in front of the standards, cohorts arranged by maniples, legions arrayed in squares, moving their columns forward at a quick pace, occupied the whole field as far as the eye could see.”

Pacatus has earlier told us that Theodosius's army at the Frigidus was recruited from large numbers of Goths, Huns and Alans - but then has them formed into cohorts and maniples, throwing pila and fighting with the gladius! His battle description is a classicising set-piece, written 'in the grand style', as Ammianus recommends, and like both Ammianus and Claudian he uses traditional language drawn from the great writers of the past. This is totally normal and accepted practice for poets and panegryrists of the period, and we should not be bamboozled by it!

[Edit - however, I don't think this is always the case; when Claudian writes of Stilicho ordering 'a legion' to support the routed Alanic cavalry on his flank at Pollentia, ('instructa Stilicho legione secutus subsidiis peditum pugnam instaurasset equestrem'), I think he was referring to a specific event; in this case the last known reference to Roman regular troops fighting in a field battle!]

So, how did I manage to get Vegetius and the 1,104 soldiers to work? Well, first, I never listened to the nay sayers on this forum

I'm sure that I am one of those nay-sayers, so pay no attention if you like! But to be exact, Vegetius has 1105 infantry in his first cohort, not 1104. The similarity is interesting nonetheless.

Also, as I've mentioned before, Macarius was not a historian but one of the martyrs in the story, and the 1104 men were not a complete unit but only those members of the unit at Melitene who confessed Christianity and were executed.

the 280-strong regiment I suggested (5 clerici et deputati, 50 Veredarii, 225 Veredarii Alii) can still be divided into 5 56-man Turmae, which is enough for 1 to support each 320-man cohort.

The problem is that the clerici and deputati are probably not the same - they may have been unit priests and medical orderlies, as I think Onur suggests? - and they were not on the army payroll, as they drew no annonae. Putting one supernumerary 'civilian' in the ranks of each turma would seem a bit superfluous, surely?

it could be that a cavalry unit was added to at least some of the Legions (perhaps those whose Equites had not been used by Gallienus to create new Promoti units) to give them some added striking power and improved communications and intelligence-gathering capabilities to offset their smaller size?

I've wondered elsewhere whether this Perge numerus may have been specially constituted as a combined infantry and (light?) cavalry unit, for a particular purpose related to the Isaurian situation; a sort of counter-insurgency force, maybe. However, carving the unit organisation onto a huge tablet suggests they were intended to remain there, and keep that same formation, for some time!

(incidentally, could you edit down the posts you quote in your replies to the specific bits to which you are replying? These threads get awfully long and repetitive otherwise!)

[Steven James]

Nathan wrote:
This underlines the difficulties of using ancient texts, or ancient evidence of any sort. None of it can be regarded as a clear window into the past, all of it requires interpretation, and none of those interpretations are going to be exact or perfect.

Well, we won’t know until they are thoroughly investigated. Dismissing them to fit a theory is not a thorough investigation.

Nathan wrote:
You have shown that 1104 can be divided by 8 and 16. But 9000 cannot, and nor can 7700, 1300. You can assume that these numbers are 'rounded' if you like, but if your thesis depends on exact number concordances you need to show that working, rather than rounding numbers up or down to make them fit.

My figure of 8,832 is not 9,000. You are correct that 7,700 and 1,300 cannot be divided by 8 and 16. Even I am aware of that. However, Nathan, the text gives the tally of Christians at ‘some’ 7,700 prisoners, of which 500 were wounded, and ‘about’ 1,300 Christians were killed. The reference to ‘some’ 7,200 prisoners means the number is not exactly 7,200. The reference to ‘about’ 1,300 killed is not exactly 1,300.

I’m noticing the usual pattern here. Find something you can try and turn around and say I’m wrong, yet avoid saying anything about my breakdown on how Vegetius and Hyginus arrived at their figures. No rounding there. I guess hell will freeze over first before this forum admits I could be right about something.

Also, I do not round the numbers up or down to make them fit. That is the prerogative of academia. In 209 BC, Livy has 4,334 replacements shipped to Sicily. I have an exact breakdown of who and what those forces are, and it comes to 4,334 men, with no rounding necessary. The 4,344 men are in many of the primary sources, referred to as ‘about’ 4,000 men, or stated as 4,000 men. It is who they omitted from the list that makes the difference. There are a lot of figures in the primary sources that are exact, and many are rounded up or down. With unit numbers, the rounding can be small, but with army numbers the rounding can be as high as 2,000. Something I did not mention before and that is Macarius 1,104 soldiers and Saint Meletius’ 252 soldiers, both omit the same component that Livy omits in his figure of 4,334 soldiers. Seems to be a long-standing trend.

Nathan wrote:
His battle description is a classicising set-piece, written 'in the grand style', as Ammianus recommends, and like both Ammianus and Claudian he uses traditional language drawn from the great writers of the past. This is totally normal and accepted practice for poets and panegryrists of the period, and we should not be bamboozled by it!

Please don’t include me in the “WE” should not be bamboozled by it.” I set my own course, which does not include throwing out the baby with the bathwater. Can you prove as you claim that these writers were using “traditional language drawn from the great writers of the past.” All I deal with on this thread is unsubstantiated conjecture.

Nathan wrote:
I'm sure that I am one of those nay-sayers, so pay no attention if you like! But to be exact, Vegetius has 1105 infantry in his first cohort, not 1104. The similarity is interesting nonetheless.

Sorry, I made a typo mistake and meant 1,105 men. As to a connection between Macarius’ 1,104 soldiers and Vegetius’ 1,105 soldiers, I have never made that connection, never entered my head and after looking at my research, they are light years apart.

So, any comment on my breakdown on Vegetius, and the 12 cohorts etc? Or, is the golden rule on this forum that when I could be right, ignore it. Anyone willing to provide evidence that Vegetius’ legion of 6,000 infantry is not historical?

[FlaviusB]

the 280-strong regiment I suggested (5 clerici et deputati, 50 Veredarii, 225 Veredarii Alii) can still be divided into 5 56-man Turmae, which is enough for 1 to support each 320-man cohort.

The problem is that the clerici and deputati are probably not the same - they may have been unit priests and medical orderlies, as I think Onur suggests? - and they were not on the army payroll, as they drew no annonae. Putting one supernumerary 'civilian' in the ranks of each turma would seem a bit superfluous, surely?

it could be that a cavalry unit was added to at least some of the Legions (perhaps those whose Equites had not been used by Gallienus to create new Promoti units) to give them some added striking power and improved communications and intelligence-gathering capabilities to offset their smaller size?

I've wondered elsewhere whether this Perge numerus may have been specially constituted as a combined infantry and (light?) cavalry unit, for a particular purpose related to the Isaurian situation; a sort of counter-insurgency force, maybe. However, carving the unit organisation onto a huge tablet suggests they were intended to remain there, and keep that same formation, for some time!

(incidentally, could you edit down the posts you quote in your replies to the specific bits to which you are replying? These threads get awfully long and repetitive otherwise!)

Duly noted regarding the quotes! With regards to the Clerici et Deputati, a way to organize them while accounting for their different positions could be to organize the legion around 'milliary' centuries, that is with 160 men each, commanded by 2 Ordinarii and one Optio with the former corresponding to the centurions commanding the 'prior' and 'posterior' centuries in the maniples. This gives us 6 Clerici et Deputati for one of these centuries, which could be split into 1 Cleric and 5 'Deputies.' The two Tribunes still have 4 each, which could then be divided into 1 Cleric and 3 'Deputies,' while 5 'Deputies' can be assigned to the cavalry unit. As for the cavalry unit itself, it may be that the smaller size of the late Roman cavalry regiments (ie Equites, Comites, Cuneus) allowed for the legions themselves to incorporate a greater amount of cavalry than the 120 Equites Legionis previously available to them.

[FlaviusB]

I find Jonathan Ross guilty of doing exactly the same, loose with the numbers. Like many academics, anything that jeopardises his theory is dismissed. For example, Suetonius gives a legion at 5,600 men, and because this does not sit well with his theory of a legion having 4,800 infantry and 1,200 slaves, he writes “Suetonius is almost certainly referring to a republican legion consisting of 5,200 infantry and 400 cavalry.”

So, Roth, prove it. Give me a troop breakdown of the 5,200 infantry plus their organisation, and the same goes for the cavalry. 5,200 infantry divided by 60 centuries per legion equals 86 point 66. So how does that work? And exactly how are the 400-cavalry organised? Give us details. Surprising to many, Tacitus, Josephus and other support Suetonius’ 5,600 figure. I am seriously convinced that academics must be the most mathematically challenged people on this planet.

Roth explains that Lydus assigns 6,000 men to a legion and Lydus gives the date for this legion to the early republic. Roth throws in the usual put down that “Lydus’ numbers are not dependable.” Again, Roth makes no effort to try and understand Lydus. Roth tells us that Lydus gives the date of the 6,000 legion to the early republic. Actually, Lydus gives the year as being 388 BC. Now anyone who knows their Roman history, will be aware that in 387 BC, the Romans added four new tribes, bringing the total to 25 tribes. Has anyone study whether there is a connection? Answer…no.

Let me share some of my research. Lydus, writing in the fourth century AD, has interpreted the 60-centuries as having 100 men, so the legion blows out to 6,000 men. Lydus is not the only ancient historian who has confused a century as having 100 men. There are a few more examples. Take Servius for example, he claims a legion had 6,000 infantry and 300 cavalry. Servius has the right cavalry numbers, but like Lydus has multiplied the 60 centuries in a legion by 100 men.

In relation to Hyginus, Roth writes: “If all personnel in a legion were assigned to the ranks of the centuries, regardless of their status, then the total number of soldiers authorized for a standard imperial legion would be exactly 4,800. Accepting this hypothesis, however, leaves unresolved the question of Pseudo-Hyginus' "missing" 1,200 men. They can be found by carefully noting Pseudo-Hyginus' terminology: he gives the number of "soldiers" (milites) in a century as 80, but says that the cohort has 600 "men" (homines). Indeed, there is a military category which belonged to the legion, but would be homines and not milites: non-combatant slaves. It is confusion about the role and number of military slaves which has led to both ancient and modern perplexity over the legion's size.”

And this is the crutch of Roth’s problem. Roth believes the discrepancy between a legion of 4,800 and 6,000 is due to 1,200 slaves being added to the legion. I can see why many have rejected his theory. It is baseless and preposterous. Like the example I earlier provided in the Vegetius posting above, Roth has no idea of how the legion was organised when in camp, and Hyginus is explaining the Roman army when in a camp. It’s the camp layout. Taking Roth’s 4,800 infantry for a legion, which would be organised into 10 cohorts each of 480 infantry, and following the camp doctrine for the principate, two cohorts amounting to 960 infantry are distributed amongst the remaining eight cohorts, so, 960 infantry divided by eight cohorts equals 120 infantry. Therefore, each of the remaining eight cohorts increased from 480 infantry to 600 infantry as per Hyginus. Someone, tell me where the maths is wrong.

The Suetonius quote comes from a fragmentary source, but the 5200-5600 figure aligns with Polybius' and Livy's statements that Republican Legions between the 3rd and 2nd centuries BC rose from 4-4200 to 52-5600, and that the Legions varied in size based on available manpower and campaign demands. Roth cites Kubitschek's entry on the Republican Legions from the Pauly-Wissowa, Real-Encycloptidie der klassischen Altertumswissenschaf as his source in this instance. 

Since the Legions at this time comprised 4 different classes of troops (Velites, Hastati, Principes and Triarii), it makes no sense to try to apply their figures numbers one-to-one to the Legions of the Late Republic onwards, when we know that those four classes no longer existed by that point. As to your suggestion that Tacitus supports the 5600-man Legion, Roth cites Tacitus' account of a mutiny in AD 14 where a Primus Pilus received 60 lashes for every century in the Legion; your math on 5200/60 is correct, and 5600/60 doesn't work either. 

388 BC is the Early Republic and is a year before 387, so there's no way that adding 4 new tribes the year after could account for a Legion having 6000 infantry and 600 cavalry the year before. Considering his numbers far exceed those of Polybius and Livy which Roth also cites, and who were writing much closer to 388 BC than Lydus was, I would say this makes his numbers unreliable. As for Principate doctrine parcelling out two cohorts among the other camps, this contradicts Hyginus' own statement (per DB Campbell's article "The Problem of the First Cohort), "the first cohort [...] will receive a double plot, seeing that it has double strength." Why partcel out two cohorts among the others when you've made accommodations for the double-strength first cohort to encamp with all of it's men together?