this has been one of several vexing beetle mysteries ever since it was posted.

Yes

Now to see what Tom says about his.

FYI, the larvae in this photo are from Floridian stock and were raised exclusively on a diet of Ganoderma lucidum. I have also had specimens shipped to me from Pennsylvania, where they were raised on Ganoderma applanatum. The Pennsylvanian larvae looked like Tom's photo. Both the Pennsylvania and Florida specimens have been raised to adulthood in a lab, so the identity of both larvae has been confirmed in a controlled environment. I can't say why there might be larvae with two different forms. It might be the difference in diet, or it could be an ongoing divergence between the populations. B. cornutus isn't exactly a long-distance migrator, and according to what I've read about them from previous studies (Whitlock 1992, e.g.), it wouldn't be surprising to find distant populations to have large genetic variations. As fascinating as that it, my own study is restricted to the insect's relationship with G. lucidum and G. tsugae, so I won't be the person to confirm whether this is happening. There's a good opportunity for publication for somebody else here, though, I'm sure.

How on earth could these larvae be so different?

Tom and I have radically different larvae yielding the same adult species for another tenebrionid, Cent*ronopus calc*aratus, but they are geographically close in our case. I suspect mine is erroneous -- a mixup of larvae -- but what if both are correct? Wow!