If the failure had been a discrete blow (e.g. a big pothole) the failure would have occurred contemporaneously with the blow. Further, if it were a discrete stress, you'd have the rim fail at the same point pretty much, and the rim would likely be very out of round. That there are many different points of failure suggests to me at least, that the failure mode is fatigue. Other's have suggested that a white powder (which would be aluminum oxide) may be present at the cracks. That would be another hint that fatigue cracks had formed, the bare aluminum exposed had oxidized, and that the continued stress cycling continuously wore off old oxide.

JohnDThompson's post, with the Jobst quote, was interesting. I learned something. Thanks, John.