@Raskuz:
Aye, history seeds are not compatible between 34.11, 40.24, 42.06, 43.03, 43.05 32-bit and 43.05 64-bit. Though geography might be same between some of them (have had exact same geography in 43.03 and 42.06 at least).
Regarding towers, there's plenty of advice around, but you're a lunatic, so...
Towers are founded by people afraid of death, typically in a setup where they experience a traumatic event, and when next war happens in an area and not site they go pick the corpses and go off to found a tower. Multiple towers, if they've created multiple nomadic groups, until they run out of nomadic groups.
Simplest ways to increase the number of towers, if you don't care about anything else, is to max out secrets (is more of a cap), add sphere:death to mortal civs and place more civs in worldgen so that they'll go to war with each other. (Megabeast - or even regular beast - caused deaths also work, but is slower and there may not be a civ surviving afterwards. Like, with just wild beasts one might look at something like 1800 years for 1-2 towers.)
There is also
mv_site.lua dfhack script to move the necro towers from elsewhere in the world near to embark; gets invasions, though it doesn't seem to fully work for site generation and founding.
My curren main embark (43.03) has 80 towers all created by 1 necromancer, who I coaxed slowly by micro-adjusting worldgen parameters messing in his travel path to travel more and build more towers. (If I just wanted towers, spamming lots of civs would have been simpler, but I also wanted minimum number of civs to get maxed out major sites with less fragmented market base for better trading and non-tower siege options.)
As for the effect in succession...You surmised correctly; though some manage to beat the necromantic armies of doom, kill all the undead, save fps and retake surface.