SteveB wrote:
....
As an aside (something probably totally irrelevant and not to do with this
discussion), I have seen and fixed LOTS of spindles in ornamental metal
railings and fences that were burst from freezing. When constructed, the
ends were not welded totally shut, and water infiltrated from somewhere. It
is amazing to look at these burst tubes, as they are burst in all places.
At the top, the bottom, in the middle, everywhere. One would think that the
water would migrate to the low point, and freeze there. Yet there are
obvious bursts in the middle of six foot sections. Maybe that's just the
point at which it started freezing after filling the tube half full. And
this was in the temperate climate of Las Vegas, and it's not that cold
there.

Point is, freeze bursting may be totally predictable and scientific and
explainable, but from my observations, it burst in some pretty weird places.

If one had the local tube thickness, weld strength/lack, etc., etc.,
etc., ... as well as the water infill pattern it would undoubtedly make
more sense. Simply looking w/o analyzing all the factors would make the
cause/effect relationship difficult to pick out, no question...

--