HBR ran the image import, and there were 27 errors. On investigation, I discovered the problem. I had a script analyzing the Barkerville database and writing another script to curl jpegs from where the Barkerville db says they should be on their site; then I assumed that if I got something, it was an image. I was forgetting that when the db was wrong and the image wasn't where it was supposed to be, curl would just pull the 404 page, which of course results in an error when AtoM tries to import it as a JPEG. No harm done, though; I've cleaned up the resulting empty records. I think we're now done with Barkerville data.