Looks like the filament jammed after about three hours, although I carried on printing air for another three before thinking something looked odd on the remote monitor.
It takes a bit of scalpel work to remove the support structure, but the end result, although still not perfect, is a lot better than before. It does not have the large hollows leading to a thinner shell that the first attempt had, although a few strands that should be on the bowl are visible still attached to the support. The hemisphere section shown here has been lightly sanded. Unfortunately the printing did not get as far as the horizontal section above the legs before it went wrong.
Now trying the head half - estimated eighteen hours to go.