Chooper stabilized OPamps like TLC2652 for instance are very nice, do offer only 1µV offset voltage (max) and 100pA input bias current (max). But only using them, without such a chopping scheme (modulation methode) will not remove the thermoelectric potentials and contact potentials. But this is needed, for at least three reasons:

1. The DUT (device under test) will not be soldered to the input, but clamped by clips supporting four wire technique. After some clampings the surface of clips can erode, leading to unpredictable contact potentials.

2. He might touch the clips and/or terminals of reed relay and warm them up unevenly, leading to unpredictable thermoelectric potentials.

3. The reed relay itself suffers from thermoelectric potentials of about 40µV/°C, when the contacts are warmed up unevenly. This is not only the case, when you touch the relay terminals by hand, but also when current is (or was!) flowing through the contacts.

Using a modulation scheme allows him to use cheap contacts for the clips and to touch them and the reed relay by hands. Also, he will not need to wait for "cooling" of contacts, if a higher current was just flowing through the contacts from another test proceeded earlier.

There are much better modulation schemes, by the way: Using a sine in combination with a Lock-in nethode (AD630, for instance), like it can be seen in sophisticated bridge configurations, would allow him to suppress hum and noise even much higher. But then he would need costly parts and parts, which might be unavailable where he lives.
So, I have proposed a modulation scheme, which uses a square wave modulation in combination with simple, cheap and available LM324, powered by +5V. I think the efficiency versus cost factor is very high with this scheme.

Kai