That's the shared 50v. The grounds are individually controlled by driver transistors. This is why you can manually ground out a specific coil to test it.

For coils, it's the +50v (or +20v for low power coils and flashers) - they daisy-chain the power wire (if you bothered to track the wires through that rats nest of wire, you'd find that it went from the backbox to one coil, to the next, and the next, and so on), and turn the coil/flasher on by grounding the other wire*

For the switches, there will usually be two wires going to BOTH sides of the switch - one side gets a row wire, the other side gets a column wire, and both row and column wires are daisychained in the same way. On WMS machines, at least, they're a bit careful with the row wires, covering it up with an insulated connector when possible (because if it comes into contact with power, even the +6.3v for the GI lighting, it'll nuke the column drive IC).

*Why do they supply power all the time, and fire things by connecting them to ground? NPN transistors (which you need to ground something at a positive potential) are cheaper than PNP transistors (which you would need to turn on the power to something) - this was likely a bigger deal when this convention was set by the first SS machines. It also makes the design simpler - instead of a drive for the +20v and the +50v, all the drive transistors just connect to ground. I think they even use the same part for driving the two in most machines. Simplicity and maintainability is of course a big design criteria for pinball machines.

Big thumbs up for this post good info!