I may have discovered a bug, but it may be in LaTeX itself. Steps to reproduce:
1. Create a figure float 2. create a sub-float within it. 3. In the subfloat caption, enter (in math mode): $[A]$, where $ indicates math mode. 4. Compile the file. This results in a LaTeX error that has to due with the right-bracket. If you remove it, things compile just fine. I have attached a very brief LyX file (2.0beta3 format) that illustrates this problem. Is there a known work-around for this problem? Thanks, James
subfloat_bug.lyx
Description: Binary data