Directly comparing a coefficient of variation (CV) derived from a homogeneous (i.e., dried, ground, and sieved) soil sample using a chemical extraction to CVs using PRS™-probes is inappropriate. The PRS™-probes are an in situ measurement, and therefore, are influenced by certain edaphic effects (i.e., micro-scale variability, biological factors, and environmental influences such as moisture and temperature), which do not affect a chemical extraction in the lab. Listed below are selected mean (n = 68) nutrient supply rate values (µg 10 cm^-2 24 h^-1) measured in a standard soil, along with the corresponding CV (%). The standard soil used is aeolian topsoil, sampled from along an old fenceline and thoroughly homogenized.

Historically, the repeated measurements of nutrient supply rates from this standard soil in the lab have yielded CVs ranging from 9 to 27 %, which are exceptional considering the degree of variability existing within biological systems. Naturally, these CVs become larger when the PRS™-probes are placed in situ due to micro-scale variability in edaphic properties. Herein lies the rationale for spreading a number of PRS™-probes throughout the experimental unit and then combining them for a composite analysis, which accounts for this variability. The number of PRS™-probes required per plot depends on the heterogeneity expected.

It is important also to remember that the magnitude of mean nutrient supply rate values has a significant influence on the resultant CV. For example, a variation in nutrient supply rate of 2µg/10cm²/24hr for NO₃^- (i.e., mean of 12) would have a much larger CV compared to the same variance in a nutrient with a larger mean (i.e., Ca2+; mean of 1624). Therefore, in such situations it is more meaningful to talk about the reproducibility of the PRS™-probes in terms of the relative percent difference among replications.