The size 110 city I posted above was definitely NOT in the course of a normal Civ game. This game was specifically geared toward getting the largest city I could possibly get, and nothing I did here should be used in any other type of game. Settings first, then my reasoning:
- Huge
- Settler
- Lakes - small lakes
- Wet, 5bn years
- Marathon speed
- 28 CSs (max, and 4 more than normal for Huge map)
- 1 AI - Arabia - (deleted all others during setup)
- No barbs
- OCC
- Policy Saving
- Legendary Start
As I said, the idea was max growth from turn 1 on. Lake map with small lakes seemed the best choice for lots of flat farmland with some rivers and no wasted tiles from water or mountains. Huge map was to give as many ruins as possible. The decision to only have 1 AI was influenced by that too. Each ruin has a (1/8?) chance to give additional population, so more ruins = more free citizens. Wet, 5bn years should yield more flat grassland (after chops if necessary). Max CSs means all Seafaring buddies will be there, so more food for me. OCC to ensure ruins don't give settlers. I'm not sure if that means every ruin that would have been a settler ruin gives a worker instead, or if it distributes that probability to all the other ruin options, but either way I didn't want settlers. Extra workers were used to explore and find more ruins, or contact the CSs for their gold. Both times I tried this, all the CSs ended up lined up at the north and south edges in two rings. That doesn't always happen in Lakes maps, but maybe it gets set because of either the difficulty setting or the map size or the fact that there were only 2 civs.
I'm not sure if difficulty affects city growth or not, but if it does, I figured Settler had to be easiest. That also meant that the AI wouldn't be competing for Wonders. I had them all, but the question is what order to build them. This game I teched to Civil Service for the food bonus, then switched straight to Engineering for aqueducts. The Great Library helped speed up Civ. Service, and obviously then made the run to Engineering faster. The detour up to Education wouldn't have helped speed things up, because my population was high enough already that the science was pouring in.
Policy Saving is actually important. In a game like this, you want a lot of culture quick for all of the Tradition policies and the 1/2 food for specialists policy, but you don't want too much. After you max out seven policy branches, you'd have to start switching between either Freedom/Order/Autocracy or Piety/Rationalism. Any switch means turns of anarchy, where you don't get the benefit of any policy (I don't think) and you lose all growth.
My last (and most annoying) question was game speed. A trial on Quick only got me up to size 96 (agonizingly close to 100, which had been my goal). Since extra food doesn't carry over after growth, it seemed like I'd waste proportionally much less food on Marathon than on Quick, thus giving faster expansion on slower speeds. No idea if that's really how it worked, or if I just executed better here.
Arabia did expand quite a bit in the mid-game. They were up to ten cities or so at one point. They had the only accessible gems, so I had to get those from them for WLTKD growth bonus. The first time I could just buy them, but they wanted some of my luxes in trade. Since deals last 90 turns but WLTKD only lasts 25, I needed my traded luxes back long before the deal was up and had to war to break the deal. I figured I might as well destroy some cities while I was at it.
One thing that did occur to me was that it would be much much better to get all the free population from ruins as late as possible since late citizens are much more expensive, food-wise. I think my last couple of citizens cost >5,000 food before city growth. It'd be much harder to guarantee getting all the ruins if you don't get them fast, but maybe I could camp a guy next to the last 5 or 8 and only get them when an opposing scout was coming.
I am anxious to hear how Joshua got up to 142, especially if it wasn't on a hand-built map!
per specialist, and I reached size 70 with it once, giving 500 million citizens. Imagine what 140 citizens would give... 
