Greg Mayman wrote in a message to Jasen Betts:

 GM> I wonder... was it conduction in the body of the board or only on
 GM> the surface? And if it was on the surface, why wasn't it over the
 GM> surface of the components when you did the free-standing version?
 GM> Better ventilation perhaps?

 GM> I'm also wondering what sort of board was it - fibreglass,
 GM> phenolic resin, or what? And did you try wiping it dry before
 GM> taking the readings?

I remember one maker of synthesizers that had a problem in this regard with
a sample-and-hold circuit.  That had a *very* high (FET-input) impedance, 
as I recall.  They originally made the board with solder mask,  and had to
change their processing as the mask was causing moisture to become trapped
in there and there would be pitch drift...

---