Why were the ingots mostly round?

[Jeff Pringle]

what is this, is it a crack or is it a scarf weld?

That looks like a weld flaw - when doing layered damascus, it's easy to get small areas (or large areas ) that don't join up, that's what they look like in the finished product. If it was a crack, the 'texture' would be different, if a scarf weld, it would be going across the layers at an angle.

A higher temperature would make it easier, and also possible to remove the crack, but then the wootz pattern would have gone. I have a feeling that I am missing something – but what?

The missing bit of info might be that attempting to work wootz at welding temp results in more cracks, turning the ingot into something that reminds smiths of cottage cheese. It's not that the pattern goes away, the whole buiscut crumbles.
Which is interesting, because there are obvious examples of two ingots welded & turned into a sword and scarf welds joining two halves of a wootz blade as B.I. points out - this has mystified western smiths for hundreds of years. Obviously, they had a different way of welding than modern smiths are used to.
So the ways of getting non-patterned areas on wootz are -
- weld wootz to non-patterned steel, this should show a seam down the back/near the edge
- some methods of working wootz give it a decarburized outer layer which needs to be sanded off before the metal will show it's pattern - this usually shows up in patches, though - you'd have to be a very forgetful smith to leave an entire side of a blade in this condition
- obviously, sanding can remove the pattern, but re-etching should bring it back - that'd be the first thing I'd look at if a blade had one sided wootz and no visible seam on the back.
Here is a small test I did - wootz sides welded to 1070 commercial steel - the lower line is the weld zone, the upper line is the transition between hardended and unhardened metal - note the wootz pattern dissappeared in the hardened area. this method of pattern removal would be impossible to do on the side-to-side of a blade, but could happen on a top-to-bottom direction.