The study of soil microarthropod biodiversity is illustrative of problems that are related to other soil organisms (fungi, for instance) or that can be found in other environments (canopy, oceanic sediments, hosts accommodating parasites, etc.). Indeed, the contribution of the soil fauna to global biodiversity remains an enigma even though, in recent years, it has received considerable attention. Our contention is that the debate on soil biodiversity will remain open – and even sterile – as long as adequate sampling methodologies are not set up, critically evaluated and largely used.
First, a critical review of the sampling strategies used for soil microarthropods is presented. In addition to an extensive compilation of publications on extraction method efficiency, articles from two journals devoted to soil biology are compared for two five-year periods (before and after Erwin's papers and before and after Rio). The most frequently used extraction methods (over 90% of studies) have a poor numerical efficiency (e.g. 7–26% for the Berlese-Tullgren funnels) and also are selective with respect to their efficiency for certain taxa (variable taxonomic and functional efficiency), 75% of studies are restricted to the upper 10 cm of soil and therefore overlook largely the microarthropod populations, some groups are often neglected, however diversified they are, and the taxonomic resolution tends to become impoverished in recent years.
In the second part of our study, the importance of bias induced by inadequate or restricted sampling strategies on biodiversity estimates is evaluated: densities are dramatically underestimated (down to 14 times less); conversely species aggregation, a factor advocated to explain the existence of numerous soil species, is overestimated; some functional groups may be quite overlooked; the species distribution along a gradient deduced from the sampling may be rather different from that really existing in the soil and interfere with the evaluation of β-diversity; species richness is often crudely underestimated (down to 50%).
Overall, at most 10% of soil microarthropod populations have been explored and 10% of species described. Obviously, much has still to be done to evaluate soil microarthropod biodiversity and afortiori understand the mechanisms underlying it. Improving and renewing the soil sampling strategy is thus a prerequisite to any real advance in our knowledge of this fascinating and obscure domain.