For all but three specimens, the values σRCT and ωIS for an individual specimen differed by less than 5 %. In addition, the conductivities measured here are in qualitative agreement with values cited elsewhere [3, 23, 17, 22]. Since IS gives the most accurate estimate of conductivity, and since the conductivities of replicate specimens of different length are not equal, the difference in σIS between these specimens of different length must be due to specimen-specimen variability. Since the values of σRCT and σIS for each specimen are nearly equal, the initial current measurements yield a direct and accurate measure of specimen conductivity. Therefore, the ratio of Io values for the two specimens is directly proportional to the ratio of specimen conductances (inverse of resistance) However, the ratio of QT values does not equal the ratio of Io values. Therefore, QT is not a direct measure of specimen conductivity.
With few exceptions, the resistance measurements at 20 kHz R20 are within a few percent of RB. This suggests that there may exist a constant intermediate frequency one could use with the RCT cell to determine sample conductivity to within an acceptable level of accuracy. However, a suitable frequency should be chosen with care. Using a frequency of 100 Hz, as was used elsewhere [22], would not be advisable due to the substantial contribution by the imaginary portion of the impedance. A fixed frequency test would reduce the cost of an AC test that incorporated the RCT cell. However, with few exceptions, the initial DC resistances are a more accurate estimate than the 20 kHz resistances R20 measured here.