- These are mutually contradicting claims, don't you think? Of course that true arc discharges have negative differential resistance, which is why ballast reactance is needed for their stabilization.

Let me make this as clear as possible.

There is no such thing as a true arc discharge until point J on the above graph.

Once a "discharge" (notice not arc discharge) arrives at point I and starts moving towards point J, it has negative resistance UNTIL it fully arrives at point J.

So a "discharge" has a negative resistance on the way to point J at which time it becomes a "true arc discharge" with POSITIVE resistance.

So an electrical discharge on the way to becoming a true arc discharge will go through a period of negative resistance, but does not become a "true arc discharge" until the resistance goes completely positive.

Ballasts are needed because if you arrive at point I, due to negative resistance, a "discharge" with negative resistance can slide quickly towards a "true arc discharge" with positive but VERY LOW resistance. It's the sharp curve of the negative resistance segment that makes a discharge go from point I to J very quickly. The ballast will prevent the circuit from having enough current to start sliding down the curve towards point J and establishing a true arc discharge with positive resistance.