Further thought, trying as I might to connect dots and sort through who struck John first and then what happened next, etc, that led to this fine exercise. So consider this: Fact 1, the CV drive axel I removed (I determined that it is indeed an EMPI part) had the spline cuts at the very end of the drive shaft, outside of the circlip, mushroomed out a little such that the CV bearing doo-dad would not just slip off the shaft after the clip was removed. I had to file the nubs down enough to get the bearing cage to come off the shaft. Fact 2: when the suspension flexes, the CV drive axel must stretch to accommodate, and the clever design is such that the shaft assembly indeed can extend in or out through a limited range. If one tries too much travel of the suspension wrt the body the limit of axial motion of the CVs will be reached and the circlips will press hard against the shoulder of the groove they are seated in. There is precious little material left at the end of the shaft outboard of the circlip groove. If the suspension travel is sudden (think pothole) there will be a sizeable impact on that clip and the material that it retains. Some deformation of that material could easily happen, esp'y if repeated enough times and the material is, um, not as fine a piece of steel as it might be. Do the following dots connect?: no bump stops means the suspension is free to move through an arc larger than possible with a bump stops present, and the CV drive axel can not properly accommodate the asked for linear extension. Just wondering how this all might play. If we have an actual extreme event here, the jolt would be transmitted through the entire assembly: bearings, stub axel, spacers and that ever present castle nut. It might even put some kind of undue force on some of the transmission innards. BTW: I know I have a broken gear tooth in my tranny. The broken tooth is on my dining room table.