Ran Lanzet has pointed out a significant lacuna in the proof of the main result in my paper "Is Frege's Definition of the Ancestral Adequate?" This has been repaired in the 'pre-publication' version of the paper, which can be downloaded here. See p.21 of that document.
I had certainly thought of the missing case, and seem to recall that at some point I'd introduced a 'simplifying assumption' that allowed me to ignore it. But that assumption is not mentioned in the published version of the paper, and it isn't nearly as easy as I'd supposed to see that it's permissible (which is perhaps why I removed it, but without fixing the affected part of the proof).