Late Roman Unit Sizes

[Nathan Ross]

the Roman Fortresses chnged their size and organization, here an example

Thanks! That's an interesting contrast. However, aside from the external defences, it looks like the main change between the Antonine and Diocletianic era forts was in the number of barrack blocks, and the size and number of the rooms within blocks: the later 3rd century fort appears to have fewer but much larger rooms on the same 'footprint' as the earlier- style fort. This might indeed have something to tell us about changes in unit size or internal structure (larger contubernia? smaller centuries?) - but unfortunately interpreting the changes is none too easy!

It is not really easy to estimate the number of men of a fortress of the late period, comparing it to to the number of men of a fortress of the second century.

I'm not so sure. The example given for the late fourth century is indeed very irregular, but we have examples from the Diocletianic era that also appear like this.

However, there are also fortifications from the 4th century onward that are quite regular quadrilaterals (Boppard, Deutz, Udrah, etc), and where we have remains of barracks blocks they seem to follow a similar sort of plan to the earlier forts and fortresses. So unless the later Romans were using a totally different method of accommodating troops in fortresses (two-storey blocks? Massive out-stationing or local billeting? Some sort of 'hot bunking'?) we ought to be able to use the numbers from earlier eras to estimate later numbers too. Although, as I've come to see from my various rough attempts at calculations, this isn't that simple either...