Numerus- who was counted?

[Jugibar]

Polybius says that a new legion consists of at least 4200 pezoi “foot soldiers” and exactly 300 hippeis “horse soldiers” enrolled in specific types of fighting unit. He ignores noncombatants but implies that they exist by mentioning a market, by mentioning slaves in the camp who have to take an oath, and by not remarking “unlike our own, Roman armies are not accompanied by servants and merchants.” So these would be additional to the given figure.

Oh, so you are referring to Polybius 6.26 (he mentions an oath) and 6.31 (talks about the market)?

Not everything is Maurice also divides his army into infantry and cavalry soldiers, to which are attached a certain ratio of grooms and servants (I think there is an exception in his description of the infantry file, but I am away from my books). To these would be added merchants, engineers, and the retinues of senior officers.

He recommended one slave for each three or four soldiers, which is based off of personal experience. (I know this because I reread Roth last night) Strategikon 1.2.70. In more unusual cases, Tacitus gives the ratio as one to two: 2.87.

That document shows that this particular unit produced counts of the number of milites and centuriones enrolled in a cohort and available for duty which did not also track servants, slaves, merchants, hangers-on, etc. in its camp. You can search the documents cited in the commentary for any reference to a count of noncombatants, but I have not heard of one.

I am not sure that Livy 21.8.3 is a good support for your position, because Livy tells readers what his number represents. The PHI text has "Abundabat multitudine hominum Poenus – ad centum enim quinquaginta milia habuisse in armis satis creditur" that is "The Carthaginian abounded in number of men- it is reliably believed that he had a hundred and fifty thousand in arms." In Greek and Latin, men in arms/in arma/en oplois are normally trained, equipped, and arrayed soldiers.

Ok, bad example on my part. A better reference would be Hyginus 1.4, 5.2-3 where the milites in a century would be 80 but there is 600 homines in a cohort. Correct me if I'm wrong, but since we're dealing with cohorts then there is no confusion between infantry and cavalry.

Another piece of evidence Roth offers is when discussing the size of Vespian's army (BJ 3.69), Josephus says the total strength of the army is sixty thousand without servants. The distinction between slaves and fighting men may have been necessary because Josephus' sources has the two counted together, not separately.