I think it's the solarisation. The thing is, of the three I did, the one that came out positive was the least exposed.
I uploaded my hasty scan. I used a positive transparency (been doing anthotypes, so had a positive laying around), so the one on the left is a positive-of-a-positive, while the one on the right is the expected negative-of-a-positive. The left hand one is twenty minutes, the right hand one is thirty.
(I foolishly followed the exposure time of someone in Europe, forgetting that I'm in Melbourne, in summer, and we long since donated our ozone layer so others could have deodorant and refrigerators, so that "twenty minutes or more" exposure time should have been "five minutes, maybe less", which is what I use now. So these are burned to a crisp. Live and learn.)