From Drogin, page 64.

I made a list of all the letters with bows to the left and/or right, and then put them together in the possible combinations.

The ones with a "x" are combinations that don't occur in English (or Latin, as far as I know).
There are repeats in the list.

And here's the final list, with the repeats removed, and also the other relevant rules associated with letter forms (half r and long s).

I've struck out the ligatures to do with version 2 of 'a' just lightly - because I think when I'm using a source other than the Bedford P, which I'm copying faithfully, I'll just use version 1 of a - I like it better.

Oh - that last line should have 'o' as the row header.


Later : I found this comment in "An Introduction to Palaeography for Scribes", Migistra Nicolaa de Bracton of Leicester
at http://members.tripod.com/nicolaa5/articles/pal.html

"They also began a practice known as "biting", in which adjoining letters with rounded parts (bowls) would be shoved together so that the bowls actually touched".

Drogin refers to the ligatures of the bows as "biting' somewhere as well.

She also says

"The use of ligatures is part of this. In some hands (such as half-uncial), ligatures--one or more letters within a word joined together--were mostly used as a space-saving measure towards the ends of lines. In other hands (Beneventan, for example), certain letters were always joined in ligature, no matter where they were located in a line. Still other hands (Luxeuil, for one) feature a dazzling variety of possible ligatures--some used more often than others."