Yesterday my grad class met in the library, and looked at facsimiles of three Canterbury Tales manuscripts and Caxton's edition.  I'd made up a worksheet for them to fill out, things to look for in each book, and sometimes hints about where to find them.  They worked in groups, 3 students on each book at one time, and every 15 minutes moved on to another one.  I wandered around the table, taking questions, dropping hints, listening to discussion.  They were enthusiastic, interested, intent.
And then we met as a big group to discuss findings, and discussion suddenly flagged.  I wasn't sure if it was just the late hour, or what.  After class I asked one student, someone I've taught before, what she thought.  She thought a lot of the problem was that they're not used to looking at manuscripts and it's hard to remember what you saw in each one; it's clear when the page is in front of you, but not later on.
That made perfect sense to me.  I should have remembered what that stage is like.  Now I have a better idea of how to guide students in taking notes on manuscripts that will help them re-visualize what they saw.
I don't think of myself as having a good visual memory.  It's certainly not photographic.  Sometimes I retain information via placement (top of the left-hand page), but that seems to be spatial perception rather than visual re-creation; and I don't always file information that way; and sometimes I think I have but I'm wrong.  One of my brothers is red-green colorblind, and the other sees those colors but has trouble processing information involving them.  Though I'm certainly not colorblind, I don't have accurate color memory.  In my graduate paleography course, I always focused on the wrong details, the ones that don't tell you much, that appear in half-a-dozen different scripts or are a standard feature of a given script rather than the tell-tale identifier of a particular scribe.
But apparently, over time, through sheer persistence, I have trained myself to have a better memory for manuscript pages than I thought.  There are a handful of manuscripts (or facsimiles thereof) whose general "look" I can summon up fairly accurately, and a few more whose pages I recognize when I see reproductions.  I'm happy to be able to see how I've improved, because I've been struggling with this graduate-school-era sense of my abilities for a long time.  And I'm really happy to feel I have some idea of how to teach this skill, that it is a skill that can be learned and not simply a talent that one either has or doesn't.