Perge Fragments translation published

[Dangun]

suggests that the Dura records show a strength of 923 for Coh. XX Palmyrenorum at one point, 781 at another.

Those numbers are just wrong.
But the book you have quoted doesn't give sources, so I can't parse where the confusion has occurred.
I suspect, that the number given by this book of 923 and 781 are only the number of pedites named in the documents.
If you add in all of the other troop types, ranks, unreadable / missing lines, and make a small allowance for attrition etc., you get a far larger total unit size.

For the original document see:
* Roman Military Records on Papyrus, (Fink, 1972); or
* The Excavations at Dura-Europos. Final Report V, Part I, the Parchments and Papyri, (Welles, Fink, Gilliam, 1960)

For analysis see for example:
* Fink, "The Cohors XX Palmyrenorum, a Cohors Equitata Miliaria", Transactions and Proceedings of the American Philological Association, (Vol. 78, 1947) pp. 159–170
* Davies, "Centurions and Decurions of Cohors XX Palmyrenorum", Zeitschrift für Papyrologie und Epigraphik, (Vol. 20, 1976), pp. 253-275;
* Fink, "Centuria Rufi, Centuria Rufiana, and the Ranking of Centuries", Transactions and Proceedings of the American Philological Association, (Vol. 84, 1953), pp. 210-215

Happy to send a PDF copy of any of these to anyone interested.

To quote Fink "The entire personnel (of Cohors XX Palmyrenorm) would consequently have numbered about 1500 in A.D. 222." (p. 163) and that is before any allowance for normal attrition. Fink includes an extensive comparison of the Hyginus numbers with the actual numbers.

[Patrick Waterson]

I think Nicholas and Duncan each have a point.

The Cohors XX Palmyrenorum appears, at least on Fink's analysis, to have a level of strength closer to a 4th century AD legion than to that of a standard auxiliary unit.

However the Cohors XX Palmyrenorum may itself be a non-standard unit for significant local reasons.

I am not sure we can generally condemn period writers' standard organisations on the basis of a single apparently non-standard unit; equally it would seem that the standard organisation is not necessarily universal and we can encounter exceptions.

The essence of the matter is: do we feel the exception proves the rule?

[Duncan Head]

I haven't yet been able to do any ferreting through actual physical books, but the extract here gives a bit more detail on XX Palmyrenorum: the roster of AD 219 lists 850-860 infantry, 335 cavalry, and 20 dromedarii, a total of 1205-1215 men. The text of this papyrus. P. Dura 100, is here. However DuBois in my original link reckoned this same report showed 721 men present for duty; I don't know if this is a seriously different reading of the papyrus, or if he is simply excluding the men out in detached garrisons - the papyrus lists men on detachment at Appadana, Beccufrayn, and other places.

If, as Edwell in the link suggests, the unit has six double-strength centuriae and five double-strength turmae, that would give XX Palmyrenorum in 219 an establishment of about 960 infantry and 320 cavalry, total 1280(*) - assuming a turma is 32 men and so a double-strength turma 64, then the cavalry are over even this strength. And Hyginus reckons a cohors milliaria equitata should have 760 foot and 240 horse, so Coh. XX is a lot bigger than "standard", assuming that Hyginus is right about what is "standard".

(*)The dromedarii are apparently carried on the books of the infantry centuries, so I presume they don't need to be added to this total.

[Jim Webster]

It does rather beg the question, who was in charge of numbers and how fixated were the Romans by 'system'

For example if a cohort was in garrison in a certain location the situation may have changed. Rather than needing '500' men things were a lot quieter and they only needed 250. Or perhaps there was a feeling that 1000 were called for.
Rather than move out a unit and move one or more in, did somebody take the decision to let the numbers of the unit fluctuate?

Certainly there is evidence of units which appear weaker than their official TOE might suggest and here we have one which might well have been stronger

Jim

[Patrick Waterson]

Certainly there is evidence of units which appear weaker than their official TOE might suggest and here we have one which might well have been stronger

From memory, some papyri relating to garrison units in Egypt showed them to be understrength.  In a quiet backwater there was probably not the same urgency to get them up to TO&E strength.  Conversely, it looks as if XX Palmyrenorum was at times overstrength, perhaps because of being in a lively theatre and probably needing to execute special duties (caravan escort, etc.).

The text of this papyrus. P. Dura 100, is here. However DuBois in my original link reckoned this same report showed 721 men present for duty; I don't know if this is a seriously different reading of the papyrus, or if he is simply excluding the men out in detached garrisons - the papyrus lists men on detachment at Appadana, Beccufrayn, and other places.

The papyrus does seem to leave a certain amount of room for interpretation, not least because of the missing lines.  Can anyone incidentally make sense of section 43, lines 15-18?

Auṛ[el(ius) Nis]ạmsus Heliodori
[ -ca.?- ] c̣xxxiiii in dupl(icari ) ị[ii]
ḍ[ro]madaṛị[i -ca.?- ]
(centuria) Daṇ[y]ṃị Teṛṭullo c̣ọ(n)s(ule)

Just wondering where the number 134 fits in.

[Jim Webster]

Certainly there is evidence of units which appear weaker than their official TOE might suggest and here we have one which might well have been stronger

It was the Egyptian stuff that I had in mind

[aligern]

Thank you Adrian for pointing out the paired legions. I am a little surprised that Patrick did not pick up on this as his schema for the pairing of auxilia neatly made two auxilia into an equivalent of a legion. If legions are paired then what does this mean for that theory.

So far no one has come across with a unified theory of Late Roman numbers that fits all the evidence and explains the exceptions. Given the far reaching nature of the proposed Diocletianic change, if it occurred then there ought to be some evidence of his edict Is there? Or is it possible that this all something that developed gradually through piecemeal reforms and catch ups here and there. Interestingly the situation on the Roman wall in Britain appears to reflect a different pattern .
Roy

[Dangun]

From memory, some papyri relating to garrison units in Egypt showed them to be understrength.  In a quiet backwater there was probably not the same urgency to get them up to TO&E strength.  Conversely, it looks as if XX Palmyrenorum was at times overstrength, perhaps because of being in a lively theatre and probably needing to execute special duties (caravan escort, etc.).

This is an interesting question.
Because it can be argued the other way as well - units closer to the front suffer more attrition (casualties, desertions etc) and are further from natural sources of  resupply.

If we accept for a moment that Hyginus was right and 1000 is standard, would a normal unit have the standard complement?
I would guess not.
Military units are often imperfectly staffed and supplied.
Highly speculative, but I would guess that normal strength would be 80 to 90% of notional/standard strength.

The practical upshot of all this guessing is that a comparison of a literary source's "standard" headcount to a primary sources' actual headcount probably understates the differences.

(PDFs sent)

[Jim Webster]

Not only that but we're looking across centuries. When a legion or an auxilia remains in station for a century or two things are going to change. The actual combat arm of the unit could dwindle because supplying engineers or guards to tax collectors was far more important.

[aligern]

There are certain 'laws' of military strength such as that governments are always strapped for cash , that wuiet sectors get less investment, that a unit in the front line when reduced by the falling out of the weak and imvalid and the desertion of the cowardly is probably a more effective team than when it had more men.  Logic suggests that the fullest legions would be those that were on hostile frontiers but where hot war had not started yet.

It would still be good to see evidence of official edicts on strengths, particularly at the tine of big changes. apparently in Britain the occupation record of firts does not tally with the the supposed garrison status in the Notitia. If this is typical it may be that the document is in many places  unreliable , after all we do not know why it was produced.  Also figures of 650,000 or 450,000 might well bear little relation to actual numbers . Maybe no one but a coupke of accountants and the emperor knew the number and just maybe no one actually knew at any one time. Knowing the total number of troops is not particularly useful if all if your provisioning is done piecemeal and your recruitment is done by request which is accepted or denied at say a regional level. 

Roy

[valentinianvictor]

The Anastasian Edict at Perge was raised as a result of the troops of the legion in question complaining about corruption and other related matters which led to them not receiving their correct dues. The Edict and Sermio was raised so that everyone could see quite literally set in stone what everyone in that legion was supposed to receive. There is no other more accurate evidence you can get other than going back in time and enrolling in a legion to find out!

The legion as described by Vegetius had 6100 foot and 720 horse. It also had attached artillery. Vegetius notes that the legion also had a train of workmen without noting the numbers of those workmen. Also he neglects to mention how many of the 'light-armed' troops attached to the legion were in number. If we removed the workmen, artillery crew and also the light troops we would probably be left with around 4200-4800 heavy infantry. We know that the Late Roman's had separate units of both light troops and artillery so this is probably what happened when Diocletian began the reforms.

I've not got access to the Perge papers at the moment so cannot check to see if the Edict mentions pay for ballistarii crew and workmen or does it just relate to the infantry and officers?

[Patrick Waterson]

This is an interesting question.
Because it can be argued the other way as well - units closer to the front suffer more attrition (casualties, desertions etc) and are further from natural sources of  resupply.

If we accept for a moment that Hyginus was right and 1000 is standard, would a normal unit have the standard complement?
I would guess not.
Military units are often imperfectly staffed and supplied.
Highly speculative, but I would guess that normal strength would be 80 to 90% of notional/standard strength.

The practical upshot of all this guessing is that a comparison of a literary source's "standard" headcount to a primary sources' actual headcount probably understates the differences.

With the Roman army, this is balanced by a systematic recruitment system, which reached, or should have reached, into all important parts of the Empire.  While the allocation of fresh men may on occasion have dragged behind losses to sickness, combat and discharge, the army's notaries would nevertheless be aware of the organisation and establishment they were supposed to be maintaining.

If Vegetius can state definite figures for unit establishments, he has to have derived them from somewhere other than his own imagination, especially in a work dedicated to a soldier-emperor.

It would still be good to see evidence of official edicts on strengths, particularly at the tine of big changes. apparently in Britain the occupation record of firts [forts?] does not tally with the the supposed garrison status in the Notitia. If this is typical it may be that the document is in many places  unreliable , after all we do not know why it was produced.  Also figures of 650,000 or 450,000 might well bear little relation to actual numbers . Maybe no one but a couple of accountants and the emperor knew the number and just maybe no one actually knew at any one time. Knowing the total number of troops is not particularly useful if all if your provisioning is done piecemeal and your recruitment is done by request which is accepted or denied at say a regional level. 

However pay does not seem to have been decided at the regional level except in the event of a usurpation, and if one is running an empire then paying the troops is a primary concern - and you definitely want to know the actual number of men you are forking out money for!

[aligern]

That's a line if logic, Patrick, that the centre wants to know what it is paying for, but there are other logics such as that commanders may include dead pays and that this might, as in Elizabethan times  be tolerated . I have some doubt too that Pay returns are going to Rome and then being matched by oay chests going out from Rome. Surely there has to be some intermediate stage? I just worry that acceptance of the efficacy and accuracy of aroman bureaucracy is based upon ignoring the huge inefficiency of administrations from the dawn of history. That is not to say that thevrecords are always wrong, but they may often be. athat of course leaves out the large possibility that others are right and that numbers just run down and head office tolerates this..
That Vegetius has 6000 men in a legion may reflect that there was at one time such an actual organisation and he had direct knowledge of it or it may mean that he has a written source that is itself theoretical or an ambition or a document describing an ideal project to an emperor.
Removing a speculative number of 'light troops' may be a correct method to get to 4,200 or 4,800 or it may be baseless. It might ge that core legionaries were trained up to use sling and javelin and bow  which were then carried in he baggage.  Do we know that Diocletian or Constantine abstracted soldiers from the legion to form new units on a systematic gasis? Or is it much more ad hoc? What there does seem some dertainty about is the split of legions into seniores and juniores, I haven't got the article in front of me, but  is this a matter of re equalising the army by splitting or could it be founding based upon cadres being transferred and then relocated and thus justifying the Juniores title?
Might it not be that emperors , especially when there are four in the field, build a mobile army from the best of what is available by abstracting circa 1000 or so good fit, men from a legion, perhaps the double cohort and then not returning it? That would leave rump legions of indeterminate size remaining in garrison, so a unit might lose its double cohort, a garrison for a fort, its lancearii and be left with about 2000 effectives?
Roy

[Jim Webster]

And remember the vexellations drawn from legions that never went home. What strength would they be kept at?
Indeed the Vexellation might be a recognition by the centre that there was only so much of a legion that was made up of front rank fighting men at any one time. The rest would be assisting governors, overseeing quarries, training recruits, or on leave

Jim

[Duncan Head]

... the extract here gives a bit more detail on XX Palmyrenorum: the roster of AD 219 lists 850-860 infantry, 335 cavalry, and 20 dromedarii, a total of 1205-1215 men. The text of this papyrus. P. Dura 100, is here. ...
If, as Edwell in the link suggests, the unit has six double-strength centuriae and five double-strength turmae, that would give XX Palmyrenorum in 219 an establishment of about 960 infantry and 320 cavalry, total 1280(*) - assuming a turma is 32 men and so a double-strength turma 64, then the cavalry are over even this strength.

Having now looked at the articles Nick has kindly sent, it seems:
- The papyrus Fink refers to as DP 12 is the same P. Dura 100 I linked to above, available online;
- Fink 1947 argues that "only five centuries of the cohort are included on either side, while it certainly had at least nine and probably had ten centuries"; and this is why he comes to the conclusion that the total strength of the cohort was as high as 1950, since there would have been 4-5 other centuries not listed in that papyrus.
- However Fink 1953 is clear that there were in fact only six centuries (and Davies 1976 quotes Fink's 1971 book as "Since all the evidence shows that the cohort had only six centuries at this time, three of the nine centurions on the rolls must be supernumeraries") - hence Edwell's six double-strength centuries. I think I much prefer this interpretation, which gives the cohort an establishment closer to 1280 than 1950.

Either version makes XX Palmyrenorum a large unit, larger than the apparent "establishment" strength of its type, perhaps uniquely so - or perhaps not.