Please ,we have this paper 

Journal of Colloid and Interface Science 251, 311–317 (2002)

doi:10.1006/jcis.2002.8407

Try looking up for the 4 electrode Wenner method. It is very much applied to estimate soil resistivity

http://www.zaminphp.com/users_ftp/1234567895/pdf/G57%20%E2%80%93%2095a.pdf

http://www.landviser.net/webfm_send/10

The mathematics are very simple.The principle is pretty much the same for any application.

@Xochitl:

thank you very much for the information.

I had already investigated 4E instrumentation for soil/water/biomaterial conductivity and I concluded that the Wenner method is sound, at least for conductivity measurements.

Anyway, when we also try to measure the dielectric constant of a specimen, the placement of the potential electrodes make a huge difference regarding conductivity and dielectric measurements. These are my preliminary empirical conclusions.

Although I still do not have the complete theoretical base for my last statement, I believe that what I observed occur mainly due to the effect of the parasitic/stray capacitance for each kind of measurement. I am inclined to think that these effects are related by impacting the measurements (conductivity vs. dielectric) in different ways. The apparent problem seems to be aggravated at relatively higher frequencies (i.e., >50KHz for soil mixtures). 

I will investigate the first document you sent the link (I appreciated for that) and I will read again the second standard document.

My conclusions are summarized below:

1) Regarding only to the EP effects, there is no difference if the potential probes are located closer or far way to/from the excitation probes. However, the current density passing through the potential probes and the electrical characteristics of the sample can still cause EP effects even in a 4E-based instrumentation. The theoretical principle for the EP "elimination" requires zero current through the potential probes which is actually not feasible in practice, even with very high input impedance instrumentation amplifiers attaches to these probes. Nonetheless, we can still achieve negligible EP effects in a certain 4E configuration.

2) On the other hand, regarding to the uniformity of the electric field lines at the dielectric cell, the regions close to the excitation electrodes are different compared to the center of the dielectric cell. This is true even if the stray capacitance is completely avoided, which is another cause of measurement distortion. The region  nearby the (excitation) electrode / specimen interface may presents a strong non-linear impedance gradient. Therefore, in practice, it is better not to place the potential probes very close to the excitation probes to avoid such region. Remember that when we convert the potential measurements into capacitance values (using the so called "cell constant"), it is assumed a uniform electric field at the "capacitor".

3) I cannot find a mathematical model which describes the above statements and I believe that this is the case because it must be preceded by a full mathematical formalization of the EP effects which is still not available, at least for generic specimen materials and geometries for the dielectric cell.

4) However, in practice, it is possible to verify if the placement of the potential probes is distorting the measurements: the dielectric values must vary linearly for different distances between the potential probes starting with a minimum one at center of the dielectric cell. If we increase this distance and move the potential probes toward their relative excitation probes, the mentioned linearity must always hold. When we find a point where this is not the case, then this is the beginning of the "forbidden" region regarding the placement of the potential probes. Such test must be repeated for the extreme dielectric cases under investigation in order to find an optimum placement that can be properly adopted.