It would be better to do the switching at the grid of the last stage. Otherwise the difference signal sees an unnaturally low load impedance, causing distortion in the tubes even though that difference signal is cancelled in the net output. The same goes for the Seduction/Reduction.
Such a switch is at a high impedance node and thus in danger of picking up hum and noise - keep the leads as short as possible. That may make positioning the switch difficult! Be prepared to move the switch r wires to find a suitable location - you are not guaranteed to find the best position the first try!
Edited 2018 May 26:
It's now 5 years later, and I looked again at the circuit. I do NOT, at this point, recommend ANY mono switch in the Eros.
Back then I must have missed the problem that would occur when the EF86 plates are connected together. The bias servos would interfere with each other.
The intermediate connection (grids of the second stage) should work with Seduction/Reduction since the stages are cap-coupled and the RIAA equalization has a large series resistance at the first stage outputs, which isolates them from each other.
In some cases, it can be done with little or no sonic damage in a preamp, but that depends on the particular circuit.
A general-purpose circuit to work with any preamp/amp would necessarily involve some loss of gain, complicating the design and system.
I'll go edit my old post.
