Did another test with the second ranked (total score) capturing the third rank civ capital. No split.
Also tested with third rank civ capturing second rank civ capital. surprise, surprise, civilization split.
Seems that to split civ, conquering civ must be SMALLER

(total score, based on ranking on retire[sp?] screen)
I am not concerned about what happens afterword, just how to get civ to split.
I am surprised that it seems that none of us are hackers. Should be easy to crack the program and find out EXACTLY what the program uses to determine when a civ splits. inspecting registers, tacing program. What the heck, cannot even determine exact formula for predicting trade route benefits.
Sorry for the trolling.
Again, anybody play D&D?
PS. seems that after the split, max city distance /2 determines which cities split providing total population remains equal. second city 8, other six cities 1 or two, would mean that all but the city8 would become new civ. Somebody needs to test this.
At one point I thought it was corruption but since corruption is based on distance; I prefer to start formulas as simple as possible.