I'm not yet convinced, aussietrev, but it is food for thought. I agree that the cracks may be propagating from the clamped joint where the engine mounts to the rails, since we often see cracks there which have not extended to the vertical flanges. Hence that feature would be odd if the problem is vertical vibration, since the top web of the beam should be stressed a lot less than the bottom edges of the flanges. However if the problem were horizontal vibration as you suggest, those holes are right on the neutral axis of the beam and therefore should not be stressed much - a little shear stress, but no tensile or compressive component at all.
The whole issue is contained in your first paragraph. If we can get to the bottom of what is causing the stress, the rest is easy.
Your proposed cure has the effect of greatly reinforcing the structure right at the point where the cracks begin, and that is very often an effective solution. I'm never happy though until I have an analysis that begins with a convincing explanation of how the cracks are being caused.
