Holes are normally used to stop the propagation of cracking. But in the Bond case, it was where the crack initiated, apparently caused by trapped moisture between the plug and the shamfer surface around the edge of the hole.
My problem is believing that the moisture had enough time to cause corrosion in the Bond gearbox after only 167 hours. Since the root causes seem to be twofold, corrosion AND residual stress, one wonders what proportion each root cause played in each shaft failure.
If EC explanations are to be believed, both failures have the same root cause(s) but originating in two separate areas of the shaft culminating in almost the same failure mode.