The slitting saw was running clockwise, as seen from the arbor support, and traversing right to left, so it was conventional milling. The table does have a climbing milling knob, which is basically a backlash adjuster. I'm happy climb milling on the vertical and CNC mills, but I haven't had the nerve to try it on the horizontal yet.
I think that Martin has come closest to identifying the issue. The casting is partially hollow underneath as can be seen by the body of the casting sitting above the table, even although the casting is in fact bolted directly to the table. The cut was carefully lined up with the central T-slot. The intention was to cut most of the way through, leaving a couple of 1/4" sections to be cut through with a hacksaw. Unfortunately I was a bit ambitious on the depth of cut and the two parts of the casting separated. So the slitting saw was left with a fair depth of cut, but suddenly unsupported on one side. So it tried to modify its form into a cone, and shattered. I picked up pieces of the cutter from both sides of the workshop, a good 16 feet apart. That's why I always wear proper safety googles when machining. Had I clamped both halves of the casting it probably wouldn't have happened.
Regards,
Andrew