During orthorectification, we discovered a problem in line 4, where
the info given for frame 34 was spurious, and all subsequent frames
were "off" by one (e.g., data for frame 35 belonged to frame 34,
etc.), and the mysterious "extra" frame at the end of the line (69)
was actually frame 68.

All of this was painstakingly corrected in the index we were using for
ortho work.  Rather than duplicate those efforts here, I've simply
replaced the original index with the ortho copy.  Amounts to
essentially the same thing, with a few extra fields (angles, fiducial
orientation and whatnot).