I like the tile limit, if anything it is too high. One problem though is that includes resources. This means that if you merge with the resource-squares your city suddenly use 4 tiles extra. This means you have to be "gamey" and avoid merging into resources unless you need collective defense (like on a frontier).

In Elemental, magic plays tricks on you!

I think Founding new citys should be cheap initially and become ridiculously expensive as you build more of them.   AND that the tile limit should be something that starts at a fixed value but can be increased by the construction of specific stuctures(actually you could just tie it into prestige we already have structures that can increase prestige and they are appropriatley limited) and by research.

Example:

1st city=FREE

2nd city=50 gildar (early game this is a signifigant enough  cost to make the choice important)

3rd City=200 gildar (Not too expensive really, but costly enough to deter just plopping it down for 1 more library)

4th city=700 Gildar (this is expensive PERIOD)

Each additional City = 1000 Gildar (A dragon is 1000 gildar A dragon can easily crush a starting city, and 1000 gildar is NEVER cheap, Late game its completely doable, but still costly enough to make you think twice.)  

Base Tile Size = 50 (seems good... may need to scale down though)

+1 Prestige = +10 Tiles (Flat values get a bad rap nowadays, but in this case I think it works since prestige is pretty hard to come by)

Then add a couple of techs somewhere (Anywhere but civics.... that tree is soooo bloated) that Either give you +x tiles OR +%20 tiles from prestige, so that each point of prestige grants you +12 tiles.

Unlimited citys SOUND like fun, but realistically its problematic middle to late game. And your going to have at least one gigantic city (thanks to palaces) Which makes sense, there should be a capital somewhere right?

Nitty Gritty Mechanics:

Im a former gamedev, so if anyone at stardock / Enterprising-modders  is reading this, heres some implementation details that may or may not help:

-Pioneer cost doesnt change.

-Count CURRENT-PLAYER-BUILT Cities. This means that if you build three citys, and your fourth city gets destroyed, the next one you build will still only cost 700 gildar. This means that conquered citys dont count towards the current count if you build three cities then conquer a fourth one, your next city will only cost 700 gildar. This encourages aggressive tactics, it means that asidfe from the cost of maintaining and feeding that city its free. This also makes remote undefended citys and frontline/influence citys more desirable targets, as new city tile mechanics make it easier for you to join up these citys to your Zone of influence, allowing you to expand by conquest while still protecting yourself from monsters.  Also means its better to try and fix a mismanaged city than to raze it. and start your own city in its place. 

-When you click the found city action, a box pops up informing you that it will cost X Gildar to transport the materials/people etc to the site of your new township. If you so wish this is where you put your snippet about how with safer and more comfortable places to live in more developed citys, it takes a good deal more money to convince people to travel to the untamed wilderness and risk life and limb.

-I think (and I propably did the math wrong) that this would roughly triple the maximum city size. Right now your only real options are bananatowns, or lots of towns. This would change that, and might even encourage citys to grow more naturally into the terrain. (around mountains and towards resources) Thats not really an implementation detail but I like to see more organic play than static and right now the optimal play pattern is fairly static in regards to city building.

-This game needs a high end gold sink pretty bad, right now there is RARELY any real reason to save up gold in large quantity. Making gold stockpiles useful for something creates more interesting pacing scenarios for the player. They will never appreciate it from that perspective, but it WILL influence play in fun and exciting ways.  It allow for periods of sustained growth followed by sudden expansion late game, those shifts are what made galciv2 interesting late game. Early Game it creates intresting decisions for the player about whether to use his gold now, or whether to save it up and try to expand before his opponents, and the ever challenging "How quickly is my opponent planning to expand?"  That leads to a whole nother set of interesting games that always interact nicely with diplomacy/politics metagames. Whats more once the AI is in some kind of shape, it will all work well enough in single player.

I do like the tile limit, because it forces you to expand and specialize, rather than building a giant city with every possible building (imagine what that would do to frame rates...).  So I think it makes it more interesting to play, even though it might not be "realistic".  Also, it wouldn't make sense to be limited to one building of each type if the city size wasn't limited.

If your AI actually evaluates its resources and compares them to a pre planned timescale, its not hard to make the AI expand at a specific pace. The problem is that then you wind up with a very predictable AI, it always plays the exact same way under the same circumstances. 

Also if your AI doesnt just cheat like crazy, it simply wont be capable of the sort of map swallowing expansion you often see in 4x AI's. it will get the second town out fast, and then it will be stuck for a while.

Granted this is all pointless right now, since the AI never really TRIES to win the game.  It just sort of paces back and forth randomly declaring war on people. ( note this doesnt mean it will actually attack you, just that it will SAY its going to attack you).

Also, If your game is big enough (Like the 64 bit maps they keep talking about) It would NOT be difficult to get 50 cities.

Time consuming? Certainly,

Chalenging? You bet!

Epic? BEYOND ANY POSSIBLE DOUBT!

Instead of a hard 50 space limit, the limit should be dynamically based on population, and somhow change the population system so that growth is limited by certain factors such as food supply, wealth, building bonus etc.

This is more realistic and give player something to work for by bring in population management and the benefit it brigns being city titles and perhaps taxation etc.

I am not a Fan of the 50 tile rule. 

I find it a crude tool to achieve the worthy goals of specialization and different paths to optimize city growth.  A 50 tile cap and the food mechanics, coupled with the troop training based entirely on build times and have nothing to do with the size of the city, really push the player to ignore large late game cities and instead focus on alot of small little ciites covering the map.

Having ONE best strategy is a flaw.