If you look at the edges of frame 13A, you can see that the exposure patterns curve into the centres of the sprocket holes and that there is no pattern to the right of each sprocket hole.
That makes me believe that it is either stress-exposure or triboluminescence (tape-ripping glow) caused by the film sticking together and then being pulled apart as it comes out of the cassette. Note how the no-pattern part is upstream of the sprockets (comes off first); this is consistent with the way the film comes off the cassette; once the trailing edge of film in a hole (unexposed) is lifted, the leading edge catches. It pulls away along the edges first, then the corners, working its way towards the centre of the sprocket hole. You can see each curved line is a "click" where a mm of film pulled away across the width of the roll.
The sticking could by caused by moisture getting in on a refreeze cycle. The pitch of the exposure pattern also makes it look like the sticking was mostly on one side of the spool and running freely on the other side.
Further evidence of sticking is in comparing 13A with 14. 13A has clearly stuck to 14 at the sprocket holes, which stressed 13A and exposed it. However, you can see that the tension of 13A pulling away has loosened the sprockets of 14 so that when it was then pulled from 15, the ripping happened in the middle with no exposure around the sprockets because they'd already been loosened.
Edit: looking at the high-res scans, the lateral lines really do look like stress marks from sticking - you can see a dense bit where it catches, then it fades away to the next click. I don't know what would cause the longitudinal lines, unless it's some regular pattern in the stickiness of the film (but why!?).