Thank you for your reply. I think the salient comment was "does not support continuous conduction" -- this is helpful. I'm guessing, then, that the voltage multiplier spark consists of continuous conduction for as long as the capacitor stack can sustain it before needing time to recharge. Also...
I have noted three different phenomena due to high electric fields in air.
1. When I run a ten-stage voltage multiplier I get sparking and crackling.
2. When I run the multiplier but current limit the output with ten 720k resistors in series, I get a violet arc that is sometimes smooth and...