Numerus- who was counted?

[Jugibar]

See Polybius 6 (where Polybius ignores servants and civilians but assumes they will be present)

I looked through book 6 (assuming that's what you meant) using the Penelope translation, and couldn't find any reference.

the Strategikon attributed to Maurice

Couldn't find a copy on the internet.

and Tabulae Vindolandenses 154 (available online)

It appears that the strength reported represents ualentes, which would be the total strength without absences (absentes) and those unfit for service (eorum). I can't find the term ualentes used by any classical author, and a net search only comes up with King Alfred the Great fragments and Crusader books.

However I did find this enlightening: http://www.academicroom.com/article/mili...vindolanda. And:

http://vindolanda.csad.ox.ac.uk/tablets/...hand.shtml
http://vindolanda.csad.ox.ac.uk/tablets/...army.shtml

It would be interesting to see what Fink had to say in his book on Roman Military Payri as well as Teitler, since they are referenced by Bowman a lot.

In general, see Roth's book on Roman logistics.

I think I know what passage you are referring to. Roth references his own article on the organization of the Roman Imperial Legion from 1994. In the article, his reference for the claim that slaves aren't part of the numerus was Digestus 49.16.11. The problem with this is that the text in question does not even mention numerus. It simply says that slaves are forbidden to become soldiers in the army, not that they weren't counted in the numerus.

Of course, as Sean correctly pointed out, this has nothing to do with the actual credibility of the source material, so deciding on whether the numbers are realistic or not is another thing altogether.

In the instances where the strength reports do represent troop strength (and not tribal numbers), the figures given are logistically reliable. You and me did a thread on that a while ago and we came to the conclusion that, although the numbers handed down by Herodotus are unreliable, they are not as unreliable as most historians would think.