As flatulent points out, you have to use a local ground reference plane, and everything must be grounded to it. Also the safety ground must be connected to it, in exactly one point. I would try to install all the power outlets on that plane in one group, with preferably only one feed from the fuse box. That is then the point where the incoming safety ground is connected to the plane, and the plane is then connected to the safety ground pins of the outlets, to essentially form a one-point-ground for all instruments. This should reduce the ground voltages greatly.
On the other hand, with uV/fA signal levels, you have to consider many other sources of disturbing stray signals as well: Thermoelectric voltages when joining wires/components of different metals, induced low and high frequency signals on the cables (not least the power cords of the instruments, and secondary ground loops formed when connecting multiple instruments to the device-under-test).
The nature of problem depends quite much on the frequency band you measure, for instance thermoelectrics voltages are DC, while mains leakage currents and magnetic inductions are 50 or 60 Hz (+ often plenty harmonics!), depending where on earth you are (where is that quantum well, anyway?
). The local AM or FM stations, cellphones etc. may also add their signals to the mess, if your bandwith is high enough. Sometimes even a component or bad connection is forming an unintended rectifier/detector for those RF signals, and transposing them to lower frequencies, like audio from an AM transmitter.
You should carefully study the nature of disturbances in your lab, and then try to figure out where they come from, and how. Then it is easier to medicate the problem. But, once more, a good ground reference point (and -plane) is always necessary--use of several wall outlets and messy extension cords is a sure way to trouble.
Good luck,
Ted