[olddog]

The more I thought about this, it makes sense. Water will only leak into the cylinder when it sits with the piston down. Depending how far down, it may be rare that the engine stops with the piston in a spot where it can leak.

The tests that were done, like magnaflux are good for finding a crack, but they may not do well at finding a tiny pin hole.

Another thought occurred to me, after you said this cylinder has been sleeved. Assuming the sleeve should be dry - no water touching the outside of the sleeve wall. Then there would have to be a hole in the sleeve that mates up to a hole in the blocks cylinder wall. What are the odds that this would happen? Not likely. However, why was it sleeved in the first place. The cylinder wall in the block may have had a big crack.

There could be a water leak between the sleeve and block. If the sleeve is not a tight enough fit, water will leak between the sleeve and cylinder wall of the block. It could go up into the head gasket area and flow over into the combustion chamber. If the sleeve is of a metal that has a different coefecient (SP) of thermal expansion than that of the block, the tightness of this fit will change with temperature.

This does not explain why it would leak at BDC but not TDC. The only thing I could think of is that the force of the rings pushing out on the sleeve was enough to seal it. Seems like a weak theory, but possible.

See if there is a way to seal up the block, yet still be able to see where the sleeve meets the block at the head deck. Pressure the block with air and put soapy water around the sleeve where it meets the block and in the bore. Look for bubbles. May not be possible, but with enough will power it may be done.