Late Roman Legion size based on the Perge Inscription

[Steven James]

FlaviusB wrote:
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.

Polybius’ largest legion size is 5,200 infantry and 300 cavalry, not 5,600 men in total. Livy’s largest legion size is 6,200 infantry (mistake) and 300 cavalry for Scipio’s African campaign of 204 BC. His second largest 5,400 (5,000 infantry and 400 cavalry). Polybius’ 5,200 infantry is his 4,000 infantry plus 1,200 velites. Polybius’ other figure of “about” 4,000 infantry is the correct figure he should have used to add the 1,200 velites and not the rounded figure of 4,000 infantry. I have also found that Polybius’ infantry figures also include the cavalry, due to Polybius not understanding that the cavalry were part of the same century as the infantry. His figures for the infantry levy of 225 BC, also includes the cavalry, so they get double counted.

https://www.academia.edu/15833849/225_BC...n_Campaign

FlaviusB wrote:
the Legions varied in size based on available manpower and campaign demands.

I disagree. The size of the legion was fixed, and the number of legions and vexillations in a campaign varied. This informs me that you have not done an intensive investigation of the data in the primary sources for the principate, but are quoting the general view of academia.

FlaviusB wrote:
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.

The property class system for the levy of the legions did not become a requisite in the late republic, but did the Romans actually abolish the property classes in the tribal system that dictated the size of the legion? That is a question that has not been asked by academia. So, how was the size of the legion determined? Under the principate, a legion has 4,800 infantry and the cavalry has 16 cavalry squadrons. Later, a legion has 6,000 infantry and 600 cavalry. Following the premise that the 600 cavalry (minus officers), were organised into squadrons of 30 cavalrymen, this produces 20 cavalry squadrons. This shows a growth of 1,200 infantry and 120 cavalry (4 squadrons), since the principate, and indicates that behind the scenes, a system, which I have discovered, is generating the size of a legion for a specific period.

https://www.academia.edu/12646553/The_Ro...erspective

Having the whole working tribal system at my disposal to the year 297 AD, has given my greater insights into the Roman army, that is unprecedented. And the designer of the Roman tribal system, was Pythagoras, of which the evidence is overwhelming, and has been vetted by two universities. By applying Pythagorean mathematical methods, I can give the dates on when every legion reform was undertaking, and it was connected to some very strong religious practices, that the Romans adhered to, of which much has been written about by the ancient sources. One huge reform, occurred in 387 BC. Now depending on the calibration date employed, if it gets rounded by one year to an even number (the calibration point), you arrive at 388 BC.

FlaviusB wrote:
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.

Is the figure of 5,600 a rounded number?

FlaviusB wrote:
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.

Generally, the Romans do their reforms in June, so whether it is 387 or 388 BC, is inconsequential. Lydus and Servius must have had a source that stated 60 centuries, and believing a century had 100 men, arrived at 6,000 infantry. Actually, the number of cavalry give the game away, but alas, as I tried to explain about the cavalry on this forum, and got shouted down I was wrong, I have no interest in explaining it again. And yet, understanding it is so obvious and simple, that is remains hidden due to people’s perspectives getting in the way. 20 years ago, I decided that everything I knew about the Roman legion was bs, so I decided to flush my brain out and start with a blank sheet, and most importantly, went with the data in the primary sources without any prejudice.

Hyginus (3) also writes that “if the legions received are odd [in number], that is three, two First Cohorts will have to camp on the sides of the praetorium.” “Whenever five or six cohorts are received, two First Cohorts will have to camp on the sides of the praetorium. Hyginus (4) “No more should be subtracted from them than the 500-man infantry cohorts next to the First Cohort.” Hyginus (40) “Those who should camp near to the First Cohorts are the banner carriers. Hyginus (24)

Hyginus’ mention of two First Cohorts can imply that Hyginus has confused two First Cohorts camped together as one First Cohort. As mentioned in those reference above, Hyginus is explaining the camp layout from three to six legions. Confusion will arise and the numbers get muddled. Where does other data, including army numbers as found for the principate, support a double first cohort on the battlefield?

Hyginus’ claims a century has 80 infantry, which means his figure of 760 infantry being part of a 1,000-man mounted cohort does not work as 760 infantry divided by 80 infantry equals nine point five centuries. Hyginus gives 1,600 banner carriers for three legions, which would allocate each legion 533-point 333 banner carriers. Hyginus also gives the 500-horse wing 16 squadrons, which would allocate each squadron 31-point 25 cavalry.

Hyginus (27) writes that “the 1,000-man mounted cohort has 10 centuries of soldiers.” Here, Hyginus has allocated each century 100 men, as 1,000 divided by 10 = 100. So, if he believes a century had 80 infantry and 20 cavalry, there should be 800 infantry and 200 cavalry, not as he claims 760 soldiers and 240 cavalry. As is obvious, all of Hyginus’ figures are rounded numbers. His 760 infantry has been arrived at by deducting his 240 cavalry from his 1,000-man mounted cohort. Here Hyginus has made a catastrophic mathematical miscalculation, which I have covered in my research. I believe Hyginus has gotten himself confused due to his usage of the word “soldier” instead of using the terms infantry or cavalry.

The writings of Dio, Paterculus and Tacitus provide references to veteran soldiers and veteran cohorts accompanying the army or acting independently of the army. When in camp, where were the veterans, who were organised into cohort of 500 men, billeted?

[FlaviusB]

FlaviusB wrote:
the Legions varied in size based on available manpower and campaign demands.

I disagree. The size of the legion was fixed, and the number of legions and vexillations in a campaign varied. This informs me that you have not done an intensive investigation of the data in the primary sources for the principate, but are quoting the general view of academia.

FlaviusB wrote:
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.

The property class system for the levy of the legions did not become a requisite in the late republic, but did the Romans actually abolish the property classes in the tribal system that dictated the size of the legion? That is a question that has not been asked by academia. So, how was the size of the legion determined? Under the principate, a legion has 4,800 infantry and the cavalry has 16 cavalry squadrons. Later, a legion has 6,000 infantry and 600 cavalry. Following the premise that the 600 cavalry (minus officers), were organised into squadrons of 30 cavalrymen, this produces 20 cavalry squadrons. This shows a growth of 1,200 infantry and 120 cavalry (4 squadrons), since the principate, and indicates that behind the scenes, a system, which I have discovered, is generating the size of a legion for a specific period.

https://www.academia.edu/12646553/The_Ro...erspective

Having the whole working tribal system at my disposal to the year 297 AD, has given my greater insights into the Roman army, that is unprecedented. And the designer of the Roman tribal system, was Pythagoras, of which the evidence is overwhelming, and has been vetted by two universities. By applying Pythagorean mathematical methods, I can give the dates on when every legion reform was undertaking, and it was connected to some very strong religious practices, that the Romans adhered to, of which much has been written about by the ancient sources. One huge reform, occurred in 387 BC. Now depending on the calibration date employed, if it gets rounded by one year to an even number (the calibration point), you arrive at 388 BC.

FlaviusB wrote:
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.

Is the figure of 5,600 a rounded number?

FlaviusB wrote:
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.

Generally, the Romans do their reforms in June, so whether it is 387 or 388 BC, is inconsequential. Lydus and Servius must have had a source that stated 60 centuries, and believing a century had 100 men, arrived at 6,000 infantry. Actually, the number of cavalry give the game away, but alas, as I tried to explain about the cavalry on this forum, and got shouted down I was wrong, I have no interest in explaining it again. And yet, understanding it is so obvious and simple, that is remains hidden due to people’s perspectives getting in the way. 20 years ago, I decided that everything I knew about the Roman legion was bs, so I decided to flush my brain out and start with a blank sheet, and most importantly, went with the data in the primary sources without any prejudice.

Hyginus (3) also writes that “if the legions received are odd [in number], that is three, two First Cohorts will have to camp on the sides of the praetorium.” “Whenever five or six cohorts are received, two First Cohorts will have to camp on the sides of the praetorium. Hyginus (4) “No more should be subtracted from them than the 500-man infantry cohorts next to the First Cohort.” Hyginus (40) “Those who should camp near to the First Cohorts are the banner carriers. Hyginus (24)

Hyginus’ mention of two First Cohorts can imply that Hyginus has confused two First Cohorts camped together as one First Cohort. As mentioned in those reference above, Hyginus is explaining the camp layout from three to six legions. Confusion will arise and the numbers get muddled. Where does other data, including army numbers as found for the principate, support a double first cohort on the battlefield?

Hyginus’ claims a century has 80 infantry, which means his figure of 760 infantry being part of a 1,000-man mounted cohort does not work as 760 infantry divided by 80 infantry equals nine point five centuries. Hyginus gives 1,600 banner carriers for three legions, which would allocate each legion 533-point 333 banner carriers. Hyginus also gives the 500-horse wing 16 squadrons, which would allocate each squadron 31-point 25 cavalry.

Hyginus (27) writes that “the 1,000-man mounted cohort has 10 centuries of soldiers.” Here, Hyginus has allocated each century 100 men, as 1,000 divided by 10 = 100. So, if he believes a century had 80 infantry and 20 cavalry, there should be 800 infantry and 200 cavalry, not as he claims 760 soldiers and 240 cavalry. As is obvious, all of Hyginus’ figures are rounded numbers. His 760 infantry has been arrived at by deducting his 240 cavalry from his 1,000-man mounted cohort. Here Hyginus has made a catastrophic mathematical miscalculation, which I have covered in my research. I believe Hyginus has gotten himself confused due to his usage of the word “soldier” instead of using the terms infantry or cavalry.

The writings of Dio, Paterculus and Tacitus provide references to veteran soldiers and veteran cohorts accompanying the army or acting independently of the army. When in camp, where were the veterans, who were organised into cohort of 500 men, billeted?

I never claimed the size of the Legion wasn't fixed... in the Principate. As my full quote says, "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." I was speaking of the early Republic, not the Principate, but thank you for misquoting me!

Whether or not 5600 is a rounded number is inconsequential when we have Tacitus stating that there were sixty centuries in a Legion, which alongside Hyginus' claim of 80 milites to a century gives us 60x80 = 4800. We also have the five ranks of centurion (Primus Pilus, Princeps Prior and Posterior, Hastatus Prior and Posterior) from Vegetius, with a sixth being known (Pilus Posterior) thanks to inscriptions uncovered from Legio II Parthica:  https://www.livius.org/pictures/germany/...-parthica/

It doesn't require a huge logical leap to conclude that if there are 60 centuries in a legion and 6 centurions are grouped together in a cohort, then 60/6 gives us 10 cohorts.

[Steven James]

FabiusB wrote:
I never claimed the size of the Legion wasn't fixed... in the Principate. I was speaking of the early Republic, not the Principate, but thank you for misquoting me!

You did write that 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." That is what I was referring to. However, my next line was a change of thought, or a jump in time to the principate, so I do apologize for the confusion I caused.

FabiusB wrote:
Whether or not 5600 is a rounded number is inconsequential…

Well, I strongly disagree, as knowing whether it is rounded makes a big difference. The problem I have with these discussions, and academia in general, is no one seems interested in gathering up all the army and unit numbers for the principate, so as to better make conclusions. The most common method is cherry picking the data to suite one’s sensibilities. If all the data from the principate was correlated, Seutonius’ claim of 5,600 men is not alone.

So, does Seutonius’ figure of 5,600 include cavalry? In 67 AD, to suppress an uprising from spreading in the district of Samaria, Vespasian sent Cerealius, the legate of the fifth legion with a force of 3,000 infantry and 600 cavalry. In 70 AD, the Roman commander Titus, when surveying the fortifications of Jerusalem was accompanied by 600 selected cavalry.

FabiusB wrote:
It doesn't require a huge logical leap to conclude that if there are 60 centuries in a legion and 6 centurions are grouped together in a cohort, then 60/6 gives us 10 cohorts.

I’m completely at a loss here. Have I disputed a legion having 10 cohorts, or are you disputing Isidores’ comment that a legion had 12 cohorts?

[FlaviusB]

FabiusB wrote:
I never claimed the size of the Legion wasn't fixed... in the Principate. I was speaking of the early Republic, not the Principate, but thank you for misquoting me!

You did write that 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." That is what I was referring to. However, my next line was a change of thought, or a jump in time to the principate, so I do apologize for the confusion I caused.

FabiusB wrote:
Whether or not 5600 is a rounded number is inconsequential…

Well, I strongly disagree, as knowing whether it is rounded makes a big difference. The problem I have with these discussions, and academia in general, is no one seems interested in gathering up all the army and unit numbers for the principate, so as to better make conclusions. The most common method is cherry picking the data to suite one’s sensibilities. If all the data from the principate was correlated, Seutonius’ claim of 5,600 men is not alone.

So, does Seutonius’ figure of 5,600 include cavalry? In 67 AD, to suppress an uprising from spreading in the district of Samaria, Vespasian sent Cerealius, the legate of the fifth legion with a force of 3,000 infantry and 600 cavalry. In 70 AD, the Roman commander Titus, when surveying the fortifications of Jerusalem was accompanied by 600 selected cavalry.

FabiusB wrote:
It doesn't require a huge logical leap to conclude that if there are 60 centuries in a legion and 6 centurions are grouped together in a cohort, then 60/6 gives us 10 cohorts.

I’m completely at a loss here. Have I disputed a legion having 10 cohorts, or are you disputing Isidores’ comment that a legion had 12 cohorts?

This was your quote from a previous post in this thread:

"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."

I could easily have misinterpreted your comment about a Legion having 12 cohorts by distributing the 11th and 12th ones among the rest of the cohorts while in camp, and I apologize if that's the case. Other wise I dislike being led in circles by disingenuous arguments, so I'll ask you this: regarding my original post that began this thread, what issue do you have with the idea that, under Constantine, the Roman Legion was reconfigured into smaller Legions about 1/3rd the size of their Principate counterparts, that is from 4800 to 1600 rank-and-file soldiers, excluding officers, slaves, non-legionary cavalry, etc.?

[Steven James]

Flavius B wrote:
Otherwise, I dislike being led in circles by disingenuous arguments…

No, I don’t belong in that category.

Flavius B wrote:
so I'll ask you this: regarding my original post that began this thread, what issue do you have with the idea that, under Constantine, the Roman Legion was reconfigured into smaller Legions about 1/3rd the size of their Principate counterparts, that is from 4800 to 1600 rank-and-file soldiers, excluding officers, slaves, non-legionary cavalry, etc.?

Partially yes and the majority no. I posted some months back my research on the Late Roman army, and as no one was interested, I deleted it.

Following my research, Diocletian took the 6,000 infantry and 600 cavalry legion and divided it by three. My research also provides the religious reasons as to why Diocletian did this, and it involves more than just having more legions.

[Nathan Ross]

Anyone willing to provide evidence that Vegetius’ legion of 6,000 infantry is not historical?

I think you've addressed that yourself! :

Vegetius’ mistakes... Vegetius getting confused about... Vegetius has incorrectly added... Vegetius has also made a mistake... Vegetius has erroneously allocated...

Seriously though... Vegetius is describing the 'ancient legion', so it's supposed to be 'historical' and not the organisation of his own time (whenever that was!). It's possible that what he describes is close to the genuine legion organisation of some point in the second or early third century, maybe. But, as you've confirmed yourself, there do seem to be mistakes in the picture he presents. 

It looks very much like he's reorganised the 6-century cohort into a 5-century cohort, on the model of the old enlarged first cohort, which for some reason does seem to have had only five double-strength centuries rather than six. As we know, six centurion positions within the normal cohort are known from the early third century, and so this is probably wrong, and so the legion size and structure is distorted accordingly.

As for the 12 cohorts... I'm sure we've discussed this before but I still cannot imagine why any army would restructure itself in order to make camp. Why would they do that? What purpose would it serve? Would it not cause total chaos, every evening and then every morning while reorganising to march out of camp? What madness would have compelled the Romans to do this, and (perhaps more importantly) why did nobody ever notice or report them engaging in this strange activity?

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.

Hmm, maybe - although the evidence that we have (both literary and, for example, fort sizes) suggests that all units were getting smaller, not just the cavalry. 

But I think it's quite likely that the legions of the 2nd/3rd century had more than 120 cavalry. If you take a Flavian-era legion of 9 normal cohorts (each comprising 6 centuries of 80 soldiers, plus optio, signifer, centurion = 83 men) and one 'double sized' first cohort (comprising 5 centuries of 160 soldiers, plus optio, signifer, centurion = 163 men), and add a 'Josephan' 120-man cavalry complement you get 5417 men. But if you increase the cavalry by 2.5 to 300 men, the total is 5597; add the aquilifer and praefectus castrorum, and we're close to Suetonius's 5600. So perhaps even by his day the legions had more cavalry?

(alternatively, of course, Suetonius was just describing a legion with a larger century. Century size is one of the greater imponderables of all these calculations.)

I still don't think we can be certain about identifying the 'veredarii' with cavalry, of course. I'm also perplexed about the apparent clear division of their number into 9 subunits of 25 and one double unit of 50... Surely nothing like a 'Vegetian' milliary first unit could have survived into the army of the late 5th century? Or could it...?

[Steven James]

Nathan wrote: I think you've addressed that yourself! :

No, I have not! I have shown via Isidore how to understand how Vegetius arrived at his numbers, not that Vegetius 6,000-man legion is unhistorical. That would contradict all my research. Wink wink.

Nathan wrote: Seriously though... Vegetius is describing the 'ancient legion', so it's supposed to be 'historical' and not the organisation of his own time (whenever that was!).

Vegetius is describing the legion before the reforms of Dioceltian, which I have termed “the last Pythagorean legion.” This is because a legion of 6,000 infantry and 600 cavalry is the size of a legion at the end of the Pythagorean system. The end of the Pythagorean system came about during the reign of Dioceltian, and what the Romans did was restart the system from the beginning. So, Livy’s books 1 to 5 provide more information on the Late Roman legion because it is this legion organisation that came into play. So, in 464 BC, when Dionysius explains that four cohorts each of 600 men were stationed outside Rome, bingo.

Nathan wrote: It's possible that what he describes is close to the genuine legion organisation of some point in the second or early third century, maybe.

The 6,000-man legion came into being during the reign of Septimius Severus, or in conjunction with the ninth saeculum.

Nathan wrote: But, as you've confirmed yourself, there do seem to be mistakes in the picture he presents.

Mathematical mistakes. Most ancient historians have made them, even Pliny in his belief the Pythagorean cosmos had seven tones, as opposed to Censorinus’ figure of six tone. For some years I followed Pliny until I found it was wrong. Dioceltian’s reforms held the key.

Nathan wrote: As for the 12 cohorts... I'm sure we've discussed this before but I still cannot imagine why any army would restructure itself in order to make camp. Why would they do that?

And yet the maths shows they did. Space would be one reason. However, Hyginus does tell us how it was done. Just another piece of information in Hyginus that has gotten no attention. When I came across the reference to 12 cohorts, I did not immediately reject it, I played with it to see where it went. Then overtime, more data backed it up. When I find a military doctrine, I then explore to see how far back in time it went. Hyginus provided more information that this was occurring during the principate, ok good, keep going back, and eventually, it goes right back to the early republic. Polybius claims a legion had “about” 4,000 infantry. I have no problem with this as it relates to an eight-cohort legion. There is a lot that is not understood about the legions and their military doctrines.

Nathan wrote: What purpose would it serve? Would it not cause total chaos, every evening and then every morning while reorganising to march out of camp?

No, chaos at all. What if the 12 cohorts were termed numeri? Cohort is a very general term and just means several maniples. This is where my research differs from most. Say cohort to most and 10 cohorts come to mind. The Roman legions has more organisations than the cohort. There are horizontal organisations and vertical organisations. However, on many occasions they get called a cohort. For example, take Dionysius’ four cohorts at 600 men.

Nathan wrote: What madness would have compelled the Romans to do this, and (perhaps more importantly) why did nobody ever notice or report them engaging in this strange activity?

It is not a strange activity. And the sources do give the data that supports this. It is all there. The problem has been academia’s lack of seriously going after all the data, and keeping an open mind, and not being fixated with finding only the information that conforms to their theory.

Nathan wrote: Hmm, maybe - although the evidence that we have (both literary and, for example, fort sizes) suggests that all units were getting smaller, not just the cavalry.

My take on the Late Roman legion is, like all legions since the being of the republic gets smaller due to some military doctrine kicking in, like assigning the triarii to be camp guards. For the Late Roman legion, I have the seniores as being in the same century as the juniors. So, removing the seniores reduced the size of the century and also the legion. These numbers exist in the primary sources. Next, and a different doctrine to the rest, is those troops termed the reserve are also in the same centuries, so when they are detached, the century decreases in size. Ammianus and so many provide the data. However, if no one takes the time to understand it, it continues to remain a secret.

Nathan wrote: But I think it's quite likely that the legions of the 2nd/3rd century had more than 120 cavalry.

I’m with you on this.

[FlaviusB]

Hmm, maybe - although the evidence that we have (both literary and, for example, fort sizes) suggests that all units were getting smaller, not just the cavalry. 

But I think it's quite likely that the legions of the 2nd/3rd century had more than 120 cavalry. If you take a Flavian-era legion of 9 normal cohorts (each comprising 6 centuries of 80 soldiers, plus optio, signifer, centurion = 83 men) and one 'double sized' first cohort (comprising 5 centuries of 160 soldiers, plus optio, signifer, centurion = 163 men), and add a 'Josephan' 120-man cavalry complement you get 5417 men. But if you increase the cavalry by 2.5 to 300 men, the total is 5597; add the aquilifer and praefectus castrorum, and we're close to Suetonius's 5600. So perhaps even by his day the legions had more cavalry?

(alternatively, of course, Suetonius was just describing a legion with a larger century. Century size is one of the greater imponderables of all these calculations.)

I still don't think we can be certain about identifying the 'veredarii' with cavalry, of course. I'm also perplexed about the apparent clear division of their number into 9 subunits of 25 and one double unit of 50... Surely nothing like a 'Vegetian' milliary first unit could have survived into the army of the late 5th century? Or could it...?

My point about the Cavalry unit sizes is that if the Romans were still raising new units en masse, then raising smaller Cavalry regiments would provide some room for increasing the cavalry element of the newly reduced legions, assuming of course that they didn't already possess more cavalry than the notional 120 Equites Legionis. 

If my count is correct, the Notitia Dignitatum lists 99 Equites, 14 Cuneus and 10 Ala regiments in the Western Roman Army, and 117 Equites, 30 Cuneus and 73 Ala in the East; assuming an Ala of 512 and a Equites of 360, this means Cavalry regiment size was reduced by a third and the Legion size by two-thirds, so there would be some room to allocate more Cavalry to the smaller legions.

[Nathan Ross]

some room for increasing the cavalry element of the newly reduced legions

I tend to think that losing the cavalry complement - detached to form new 'Equites Promoti' units - was probably the first stage of the shrinking of the legions, so adding more cavalry to them again would seem a backwards step.

Then again, the Perge inscription seems to show a cavalry complement still in existence, or restored again, in c.498, so who knows!

the Notitia Dignitatum lists 99 Equites, 14 Cuneus and 10 Ala regiments in the Western Roman Army, and 117 Equites, 30 Cuneus and 73 Ala in the East; assuming an Ala of 512 and a Equites of 360

Using the ND to calculate army sizes is very tricky, I would say - it's a mishmash of fragmentary info from various periods. Parts of the western list might (I believe) date back to Valentinian and Valens, while other parts have been updated to the 420s. The eastern list may date to c.392, or be an entirely new construction of c.450 (if we follow Anthony Kaldellis's recent work), or be another mix of various periods. Either way, the units listed almost certainly did not exist at the same time.

Units sizes are another unknown, of course. I've written in another thread about the difficulties of attempting to judge the size of late units by the size of later forts. Quite possibly these traditional unit titles no longer connoted a set number of troops by this date.

[FlaviusB]

some room for increasing the cavalry element of the newly reduced legions

I tend to think that losing the cavalry complement - detached to form new 'Equites Promoti' units - was probably the first stage of the shrinking of the legions, so adding more cavalry to them again would seem a backwards step.

Then again, the Perge inscription seems to show a cavalry complement still in existence, or restored again, in c.498, so who knows!

the Notitia Dignitatum lists 99 Equites, 14 Cuneus and 10 Ala regiments in the Western Roman Army, and 117 Equites, 30 Cuneus and 73 Ala in the East; assuming an Ala of 512 and a Equites of 360

Using the ND to calculate army sizes is very tricky, I would say - it's a mishmash of fragmentary info from various periods. Parts of the western list might (I believe) date back to Valentinian and Valens, while other parts have been updated to the 420s. The eastern list may date to c.392, or be an entirely new construction of c.450 (if we follow Anthony Kaldellis's recent work), or be another mix of various periods. Either way, the units listed almost certainly did not exist at the same time.

Units sizes are another unknown, of course. I've written in another thread about the difficulties of attempting to judge the size of late units by the size of later forts. Quite possibly these traditional unit titles no longer connoted a set number of troops by this date.

Given that the likely reason for the Promoti being created was Gallienus' requiring more cavalry units (and more units in general) to make up for those lost to Postumus' army and the 'special relationship' with Odenathus in the east, I don't think we can say for certain that all the Legions lost their mounted complements.

As for the Notitia, I think a date in the 420s for the West and the 440s for the East (going off of Kaldellis) is a reasonable estimate; my point was that the numbers show a substantial amount of cavalry regiments bearing the new designations. Based on this, it seems like the Roman Army was able to mobilize a significant amount of mounted troops, which would allow for some to be assigned to the newer, smaller legions.

[FlaviusB]

the Notitia Dignitatum lists 99 Equites, 14 Cuneus and 10 Ala regiments in the Western Roman Army, and 117 Equites, 30 Cuneus and 73 Ala in the East; assuming an Ala of 512 and a Equites of 360

Using the ND to calculate army sizes is very tricky, I would say - it's a mishmash of fragmentary info from various periods. Parts of the western list might (I believe) date back to Valentinian and Valens, while other parts have been updated to the 420s. The eastern list may date to c.392, or be an entirely new construction of c.450 (if we follow Anthony Kaldellis's recent work), or be another mix of various periods. Either way, the units listed almost certainly did not exist at the same time.

Units sizes are another unknown, of course. I've written in another thread about the difficulties of attempting to judge the size of late units by the size of later forts. Quite possibly these traditional unit titles no longer connoted a set number of troops by this date.

While we're on the subject of the Notitia, the idea of the Legions being divided into thirds seems to fit with the numbers we're given, at least by my calculations. In DB Campbell's article "Did Diocletion overhaul the Roman Army," he mentions that Emil Ritterling identified 56 Legions that were in existence by the end of Diocletian's reign. If we exclude 9 which are only evidenced much later according to Campbell, 47x3 gives us 141 reduced Legions; by my count, the ND lists 103 Legions of all kinds (Palatinae, Comitatenses, Limitanei, Pseudocomitatenses) in the East and 90 in the West, which means there should have been 193 in either half by the 440s, going by Kaldellis' arguments. 52 1600-man Legions raised between, say, AD 324 when Constantine became sole ruler of the Empire and c. 440, is a rate of c.2 per year. That seems quite feasible alongside the raising of other units and incorporating new barbarian numeri.

[Nathan Ross]

I don't think we can say for certain that all the Legions lost their mounted complements.

Yes, perhaps not - there aren't enough Promoti units anyway!

But I think a legion of a the comitatensis and a legion of the ripariensis or limitanei could have been very different by the later 4th century. Perhaps some of the frontier formations still included cavalry, but the field army units are unlikely to have done so.

As for the Notitia, I think a date in the 420s for the West and the 440s for the East (going off of Kaldellis) is a reasonable estimate

Some of the western list seems to be c.AD420 - but the British section, particularly the wall garrison, and I think a lot of the units listed in Spain too, would be anachronistic after c.410. Most of the Danube garrison appears to be from an earlier era too.

While Kaldellis makes a good case for the eastern praesental field armies being c.440s, I don't think his reasoning extends to the regional field armies or the limitanei. With occasional exceptions (the 'Equites Arcadi' in Scythia, for example) most of the Danube and eastern frontier units look very 'mid 4th century'!

the idea of the Legions being divided into thirds seems to fit with the numbers we're given

But when would this have happened, and how? We have at least a working theory for a smaller division - the two-cohort vexillations that appear in inscriptions and other texts of the late 3rd and early 4th fourth century, detached for field army expeditions, and later forming 'mini-legions' of their own. This rather ad-hoc but firmly evidenced division, perhaps continuing over several decades, would reduce the size of the frontier legions and create multiple smaller field legions without the need for one single officlally-organised breakup of all the legions in the entire army into equal-sized bits.

Under this hypothesis, two cohorts of an old-style legion - totalling 12 centuries, or about 960 men - would be combined into one 'mini-legion' under a new name. The organisation would then perhaps be decimalised into ten 'mini cohorts' of c.96-100 men, each under an ordinarius. I believe this system is the root of what we see in the Perge inscription.

In time, and perhaps rather piecemeal, the remaining legions on the frontiers may have been reorganised similarly, with ten or more 'mini cohorts'. This might explain, for example, the listings in the ND for Dacia, where we see the legion tertiaedecima gemina dividided into five parts, and legio quinta Macedonica into four parts (with other parts presumably having gone to field army units elsewhere). In Scythae we have five cohorts of legio secunda Herculia at one place, five more cohorts somewhere else, a 'headquarters' at a third location and another cohort of muscularii elsewhere. The same pattern for legio prima Iovia, and for other legions along this frontier. Apparently these legions now contain more than ten cohorts (and some of those cohorts, based on the forts they occupy, are very small)!

I think what we are seeing here is a 'cohort' much reduced in size, perhaps c.100-120 men on the pattern of the enlarged 'century', with 'legions' of varying sizes and perhaps varying internal organisation. I do not think we see evidence here of one single uniform reorganisation of the entire army on a new model.

[FlaviusB]

the idea of the Legions being divided into thirds seems to fit with the numbers we're given

But when would this have happened, and how? We have at least a working theory for a smaller division - the two-cohort vexillations that appear in inscriptions and other texts of the late 3rd and early 4th fourth century, detached for field army expeditions, and later forming 'mini-legions' of their own. This rather ad-hoc but firmly evidenced division, perhaps continuing over several decades, would reduce the size of the frontier legions and create multiple smaller field legions without the need for one single officlally-organised breakup of all the legions in the entire army into equal-sized bits.

Under this hypothesis, two cohorts of an old-style legion - totalling 12 centuries, or about 960 men - would be combined into one 'mini-legion' under a new name. The organisation would then perhaps be decimalised into ten 'mini cohorts' of c.96-100 men, each under an ordinarius. I believe this system is the root of what we see in the Perge inscription.

In time, and perhaps rather piecemeal, the remaining legions on the frontiers may have been reorganised similarly, with ten or more 'mini cohorts'. This might explain, for example, the listings in the ND for Dacia, where we see the legion tertiaedecima gemina dividided into five parts, and legio quinta Macedonica into four parts (with other parts presumably having gone to field army units elsewhere). In Scythae we have five cohorts of legio secunda Herculia at one place, five more cohorts somewhere else, a 'headquarters' at a third location and another cohort of muscularii elsewhere. The same pattern for legio prima Iovia, and for other legions along this frontier. Apparently these legions now contain more than ten cohorts (and some of those cohorts, based on the forts they occupy, are very small)!

I think what we are seeing here is a 'cohort' much reduced in size, perhaps c.100-120 men on the pattern of the enlarged 'century', with 'legions' of varying sizes and perhaps varying internal organisation. I do not think we see evidence here of one single uniform reorganisation of the entire army on a new model.

I'm in agreement with you that the Vexillations are the key to how this reorganization was carried out. The Beatty Papyri suggest that the Legions were still of a similar size to the Principate prior to Diocletian's abdicating, so my guess is that smaller Vexillations began to be amalgamated, and those legions large enough to be reduced in size were so reduced, some point just before or after AD 305 when Diocletian and Maximian retired. The process of creating the new 1600-milites legions from the components of the older 4800-milites Legions would seem to have been completed by the reign of Constantius II, which is when we have access to Marcellinus' narrative, so between AD 305 and 353. That's almost half a century, which should have been enough time for the reorganization to take place, facilitated by many Legions already being split in vexillations.

As to the irregular numbers of cohorts, the organization I suggested for the Perge legion allows for this. It can be divided into 20 80-man centuries, 10 160-man double-ceturies, 5 320-man units and 2 800-man units, and 6 80-man or 3 160-man units can form a 480-man unit. I'm sure sub-units of various sizes would have been created to occupy forts and other such positions along or just behind the frontiers.

[Steven James]

FlaviusB wrote:
The process of creating the new 1600-milites legions from the components of the older 4800-milites Legions would seem to have been completed by the reign of Constantius II.

How do you factor in the seniores?