: : Treadgold has worked out that the total number of troops in the
: : Byzantine state varied between 1.2 to 2.4% of the total
: : population in the period from 284 to 1081 AD. At around 850 AD
: : the state had about 170,000 troops and a population a little in
: : excess of 12 million.

: : Estimation of the size of Gondorian armed forces is difficult
: : but I would imagine it was between 40,000 and 50,000.

: : If the parallel holds then this would suggest, for a
: : conservative percentage of 1.5% in the armed forces, a figure
: : of around 3.5 million inhabitants for Gondor, and for Minas
: : Tirith (compared to 250,000 people for Constantinople) a
: : population of 73,500.

: Martin, unless I'm doing the math wrong, your 73,500 population for Minas Tirith allows only for 1102.5 men. Even rounding up, that wouldn't begin to approach the numbers we've discussed for Minas Tirith in the other thread.

: 1.5% of the population does sound about right, however. Assuming Gondor at the end of the Third Age had about 35,000 soldiers, the population would be around 2,333,333.333333. For Minas Tirith, if there were 11,000 soldiers stationed there, the supporting population would in the neighborhood of 733,333.3333333. This includes not just the city but the supportive lands as well.

: If we assume Gondor had 45,000 soldiers, the overall population goes up to about 3,000,000.

I didn't mark the transition in method of estimation from the whole population to that of MT very well.

I think the ratio of army size to total population size is fairly valid but I was not using that for the size of the capital city - the connection between this size and the local forces is less valid.

What I did is, once having produced a figure for the total population from the ratio of soldiers to overall population, I compared the overall population of the Byzantine Empire to that predicted for Gondor then simply applied that ratio to the size of Constantinople's population to predict that of Minas T.

If there is any validity in the ratio of troop numbers to size of capital population it would be between those troops normally stationed around the capital.
In the case of Byzantium this would be the numbers of the elite units of the Tagmata which were normally stationed in and around Constantinople. In the 9th century this numbered around 24,000, which with a population of around 250,000 for the city is around 10%. For Minas T (at around 73,000) it would predict a locally stationed force of roughly 7,000 men - who, one might suppose, would represent the elite units of the Gondorian army.

This ratio also works fairly well for an earlier era. In the Late Roman Empire the troops normally stationed close to Constantinople consisted of the Imperial Body Guard - the Scholae of around 3,500, and the two Praesental Armies (commanded by the two "Masters of Soldiers in the Presence [of the Emperor]) which were each of some 21,000 men. A total of around 45,500 troops, compared to a population of c. half a million for Constantinople at the time.