I have a SWAG for the higher frequency vibrations on the Peugeot that went away once the dropouts were straightened.

All physical vibrations can be modeled as a spring-mass-damper system. You mentioned torsional vibration, which could meant that one or both of the dropouts were twisted. Once the wheel is clamped in, there's an induced torque on one or both of the fork blades. As you bend the fork vertically, the torque probably increases. When it vibrates, the torque increases and decreases as a function of the vertical fork flex. The torsional harmonic frequency is probably different than the vertical. Once Peter Mooney straightened those up the induced torque went away. At least I think my model matches the data, which is the most important thing when making up a good story...

At any rate, that's some great data, and gives some good insight as to why good frame alignment helps optimize ride characteristics.