Strange high-voltage remote effect

I still on the twin tower Relay testing  tinkering around: to supply the part with the correct voltage, something had to be fiddled. With approximately 16 kV that fits well, energy source is a 12V lead acid battery.
To determine the minimum distance of the spark gaps, I tested :), the stuff on the "keyboard drawer", so the part where the keyboard resides and is extendable. Here a picture of the building:
The transformer is supplied with the "standard - ZVS', but irfp260n. Idle, takes little power the circuit, also not very much burdened with approximately 4A.
But the real question:
This "keyboard drawer" runs on a rail. And I throw the ZT with the ZVS, there is spark discharge between the guide rails, although nowhere is a conductive connection between the Earth and the structure. Make this track a capacity which must compensate by induction loads?
Would glad if someone told you that.
Could be that the aperture sends out radio waves and catch the iron Rails then the power off, aka Tesla's basic idea.
My first ZVS with 135Watt was to compete with the multi meters by 5 meters without any antenna alone the emits hung vertically (!)
If you think the wooden plate where the building on it, which can go quite through the wood, because wood contains a residual moisture, even if it seems dry fart.
Put on but time somewhere else the whole, as a control test.
The underground is probably cheaper chipboard cardboard, which is not well insulated. But that doesn't explain why there are sparks between the metal rails of this document and those that are attached to the table.