[ducked]

Or could it be spalling due to thermal shock, from water droplets hitting the cylinder wall?

IIve done water decokes a few times, using a plant sprayer, and for a while I rigged an IV line to suck water in at relatively high revs using the negative pressure generated inside the air cleaner. It seems rather more linear than I would expect, but I dont really know.

This cylinder seemed to have less carbon than the others, consistent with it having got more water

In future (if there is one) ill maybe restrict water decokes to using steam rather than liquid water.